Type 2 Iodothyronine Deiodinase in Human Skeletal Muscle: New Insights into Its Physiological Role and Regulation

The paper by Heemstra et al. (1) in this issue of JCEM provides some striking and unexpected results that have significant implications for our understanding of thyroid hormone physiology in humans. Their findings help to clarify how the regulation of type 2 iodothyronine deiodinase (D2) in skeletal muscle is affected by thyroid status, fasting, and insulin, and illustrate the importance of its functional activity in that tissue. They also contribute to our understanding of the effects of illness and hypothyroidism on thyroid hormone metabolism and help define the role of skeletal muscle D2 in the production of the active hormone T₃.

The iodothyronine deiodinases are a fascinating group of three selenoenzymes, two of which, the types 1 (D1) and 2 enzymes, activate T₄ by removing an iodine atom from its outer or phenolic ring, converting it from an inactive prohormone to the active hormone T₃. The type 3 deiodinase (D3) inactivates both T₄ and T₃ by removing an iodine atom from the inner (tyrosyl) ring. Unlike the case in rodents, where a large fraction of T₃ is secreted directly by the thyroid, in humans 80 to 90% of the T₃ is produced from T₄ by these deiodinases in various tissues. The first of these enzymes to be recognized, D1, principally found in liver, kidney, and thyroid, was thought for many years to be the only outer ring deiodinase. Curiously, D1 is quite susceptible to inhibition by the antithyroid drug 6-n propylthiouracil (PTU) (2). Subsequent studies of the feedback regulation of pituitary TSH secretion by thyroid hormones revealed a second outer ring deiodinase (D2), which was insensitive to PTU, was expressed in pituitary and cerebrocortical tissue, increased in activity in response to hypothyroidism (instead of decreasing as does D1), and had different kinetic properties from D1 (3,4). The T₃ produced by D2 was especially effective in entering the nucleus and binding to thyroid hormone receptors, a property later explained by its location in the endoplasmic reticulum. D1, on the other hand, is located in the plasma membrane, and the T₃ produced by this enzyme preferentially enters the plasma pool (5). Thus, D2 serves the special purpose of providing nuclear T₃ from intracellular T₄ in a cell-specific fashion. This makes D2 especially important for regulating the hypothalamic-pituitary-thyroid axis, where its activity increases in response to a decrease in serum T₄ concentrations, such as occurs in iodine deficiency or early autoimmune thyroid disease, well before the serum T₃ falls. If the decrease in plasma T₄ is too great to be compensated for by the increase in D2 activity in the hypothalamic-pituitary feedback sensors, an increase in TRH and TSH will occur to stimulate the thyroid.

With respect to skeletal muscle, early studies showed that the expression of D2 mRNA and specific D2 assays confirmed that PTU-resistant deiodinase activity was present (6,7,8,9). In later studies in which we expressed D1 and D2 in intact cells at concentrations similar to those in human liver or skeletal muscle, respectively, and exposed them to physiological concentrations of free T₄, D2 performed with a higher efficiency than D1 in T₃ production (10). This might be expected from the effectiveness of D2 in catalyzing this reaction and its much lower Km for T₄ than that of D1. Extrapolating from the early estimates of D2 activity of about 1 fmol/mg·min in skeletal muscle, the T₃ production by D2 under physiological conditions was much lower per gram of muscle protein than that of liver. When we predicted what the contribution of skeletal muscle to whole body T₃ production might be, the large mass of skeletal muscle protein compensated for the lower activity, leading to the conclusion that skeletal muscle D2-catalyzed T₃ production could exceed that arising from D1 (10).

It has subsequently become clear that there can be an artifactual release of iodide not accompanied by T₃ production during D2 assays of muscle homogenates, leading to an overestimation of D2 activity (Marsili, A., A. L. Maia, and P. R. Larsen, unpublished data). This iodide release is not saturable by high substrate concentrations and can be attenuated, but not eliminated, by high concentrations of reducing agents such as dithiothreitol. To avoid artifactual results using muscle, it is necessary to specifically quantify the labeled T₃ produced during the deiodination reaction by chromatographic methods. This has been done in more recent studies, with estimates of D2 activities in normal skeletal muscle being about two orders of magnitude lower, in the range reported by Heemstra et al. (1) of 0.007 fmol/mg·min. However, even with specific assays, D2 activities in human myoblasts or myocytes in culture are reported to be severalfold higher than this, suggesting that measurements of D2 in intact cells in culture may give more reliable estimates of the in vivo situation (6,11). Nonetheless, as pointed out by these authors, it is unlikely that skeletal muscle D2 contributes significantly to peripheral T₃ production in humans, although D2 in the thyroid, bone, or brain could do so (5). It would also follow that the decrease in serum T₃ that occurs in severe illness is not due to down-regulation of muscle D2, although a reduction of D2 in other tissues could contribute to this effect, in addition to the well-recognized reduction in D1 function.

Although these results argue against a role for muscle D2-mediated T₄ to T₃ conversion in peripheral T₃ production, it does not mean that the low levels of D2 and the T₃ that it may generate are biologically irrelevant. It has been suggested that a potential approach to achieving increased resting energy expenditure via thyroid hormones could be achieved by activating D2 in the huge mass of skeletal muscle. One such approach to achieving this in rodents has been the administration of bile acids (12). These agents interact with the bile acid G protein-coupled receptor TGR5 on brown adipocytes to increase cAMP. Because the Dio2 promoter is cAMP responsive, this increases the level of D2 and subsequently leads to an increase in O₂ consumption. Eliminating D2 by gene targeting eliminates this effect. Skeletal muscle cells also express TGR5 and could be stimulated in the same way (12). An increase in D2 activity in cultured primary murine myotubes has been induced by bile acids by Grozovsky et al. (11). This approach is more feasible if the T₃ produced remains sequestered in the myocyte and does not escape into the general circulation. The latter event could affect the hypothalamic-pituitary-thyroid axis, increasing the ratio of circulating T₃ to T₄. This effect has been documented in some patients who develop metastatic thyroid cancer due to tumors that express high levels of D2 (13). From that point-of-view, the lack of D2-mediated T₃ from muscle could be seen as advantageous.

Heemstra et al. (1) also report the surprising result that D2 activity does not increase in skeletal muscle from hypothyroid adults. As mentioned, it has been known for some years that D2 is negatively regulated by thyroid hormone both at a transcriptional level and a posttranscriptional level. Although the mechanism(s) has not been defined, T₃ reduces the level of D2 mRNA about 50% in the cerebral cortex of the hypothyroid rat (14). However, the major negative regulation of D2 activity appears to be posttranslational. In intact cells expressing human D2, exposure to T₄ induces ubiquitination that programs the enzyme for proteasomal degradation (15). This effect of substrate to accelerate D2 degradation also occurs in cells expressing endogenous D2. The ubiquitination process also interferes with catalysis by D2 (16). The opposite effect occurs when T₄ falls. This can explain the increased fractional conversion of T₄ to T₃ observed when T₄ is reduced, which has been a consistent finding in both in vitro and in vivo studies in humans and rats (5). Of course, the gravimetric quantity of T₃ produced by this D2-mediated increase in fractional T₄ to T₃ conversion will depend on the concentration of plasma T₄ available for deiodination.

To evaluate the response of human skeletal muscle D2 to hypothyroidism, the investigators discontinued levothyroxine therapy in a group of 11 patients who had undergone prior thyroidectomy and radioiodine ablation for thyroid cancer. The quadriceps muscle was biopsied for D2 assay 4 wk after discontinuation of levothyroxine and 8 wk after its reinstitution. Although serum T₄ concentrations fell to 5% of normal and TSH rose to a mean of 142 mU/liter during withdrawal, with a return to normal levels after resumption of T₄, there was no increase in D2 mRNA, nor was the slight increase in D2 activity in the hypothyroid state statistically significant. An effect of the local anesthetic used for the biopsies on D2 expression was ruled out by appropriate control studies. It is difficult to explain this surprising result, unless longer periods are required for the response of human muscle D2 to hypothyroidism or D2 expression is regulated differently in the quadriceps than in other muscles. It should be noted, however, that the response of muscle D2 to hypothyroidism in rodents has not been well studied due to its low expression.

To test other stimuli that might alter D2 activity, the authors evaluated the effects of fasting and a hyperinsulinemic euglycemic clamp. D2 mRNA was unchanged after a 14-h fast but decreased after this was extended to 62 h. D2 activity was unaffected at both timepoints. Hyperinsulinemia increased D2 mRNA in muscle after a 62-h fast but not after a 14-h fast. There was no significant change in muscle D2 activity induced by the insulin infusion. However, the muscle D2 activity was quite low after the 62-h fast and the se values were large, perhaps concealing a significant change.

One caveat that must be raised with respect to studies of D2 in skeletal muscle is the recent report of Kahn and colleagues (17) in which ectopic brown adipocytes were identified interspersed between skeletal muscle cells in mice. The C57BL/6 mice had very few such cells and developed metabolic syndrome rather readily on a high-fat diet. On the other hand, 129S6/SvEvTac mice were quite resistant to a high-fat diet and had a much higher expression of the brown adipocyte marker uncoupling protein-1 in muscle. The results of positron emission tomography scans have made it clear that there are collections of ectopic brown adipose tissue in many humans, and there could even be smaller deposits that might contribute to the levels of muscle D2 (18,19,20).

Taken together, these results provide the first evidence that there may be significant differences in deiodinase physiology between rodents and humans. It is difficult to explain the lack of an increase in D2 activity during hypothyroidism, but perhaps this is a muscle-specific effect. The low concentration of skeletal muscle D2 argues that the effects of the T₃ produced by this enzyme will be local rather than systemic. This would be advantageous should muscle D2 prove to be an appropriate target for reagents that could enhance resting energy expenditure through increasing this T₄-activating enzyme.