Absence of thyroid hormone activation during development underlies a permanent defect in adaptive thermogenesis

Abstract

Type 2 deiodinase (D2), which is highly expressed in brown adipose tissue (BAT), is an enzyme that amplifies thyroid hormone signaling in individual cells. Mice with inactivation of the D2 pathway (D2KO) exhibit dramatically impaired thermogenesis in BAT, leading to hypothermia during cold exposure and a greater susceptibility to diet-induced obesity. This was interpreted as a result of defective acute activation of BAT D2. Here we report that the adult D2KO BAT has a permanent thermogenic defect that stems from impaired embryonic BAT development. D2KO embryos have normal serum T3 but due to lack of D2-generated T3 in BAT, this tissue exhibits decreased expression of genes defining BAT identity [i.e. UCP1, PGC-1alpha and Dio2 (nonfunctional)], which results in impaired differentiation and oxidative capacity. Coinciding with a reduction of these T3-responsive genes, there is oxidative stress that in a cell model of brown adipogenesis can be linked to decreased insulin signaling and decreased adipogenesis. This discovery highlights the importance of deiodinase-controlled thyroid hormone signaling in BAT development, where it has important metabolic repercussions for energy homeostasis in adulthood.

Figure 1

Deiodinase expression during BAT development. A and B, Schematic of D2 and D3 modulation of thyroid hormone signal. D2 converts T4 to T3 (A), increasing nuclear T3 levels, while D3 can inactivate T3 and T4 (B), decreasing thyroid hormone signal. C, Image of H&E section of wild-type mouse embryo at E16.5 (top) and E18.5 (bottom) with arrow indicating interscapular BAT (iBAT). Inset shows enlarged iBAT, where a. is section of BAT dissected for subsequent analyses. Bars, 1 mm. D, Dio2 and Dio3 mRNA levels of embryonic BAT graphed relative to E16.5 expression. E, D2 and D3 activity of BAT sonicates from E16.5, E17.5, and E18.5 embryos. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 by one-way ANOVA with Newman-Keuls Multiple comparison. F, Chromatogram of T4-fate, as resolved by UPLC, when E16.5 (top) and E18.5 (bottom) BAT sonicates are incubated with ¹²⁵I-T4. Deiodination products are labeled by arrow according to retention time. Area depicting T3 peak is colored in red; rT3 peak in blue.

Figure 2

D2-generated T3 contributes to brown fat identity. A, Plasma TSH, T4, and T3 concentrations of E18.5 WT (n = 3), D2Het (n = 8), and D2KO (n = 5) E18.5 embryos from 5 litters. B–D, Expression of selective genes in iBAT from WT, D2Het, and D2KO embryos at embryonic d E16.5, E17.5, and E18.5. mRNA levels were determined by qRT-PCR and are graphed relative to E16.5 WT expression. Genes are grouped into (B) genes common to both white and brown adipogenesis, (C) genes that are specific to BAT, and (D) genes that are involved in thermogenesis. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 vs. WT of respective day by one way ANOVA with Dunnett’s Multiple Comparison test. E, Gene expression in confluent brown preadipocytes after 24 h in stripped serum plus vehicle or 100 nm T3. *, P < 0.05; and **, P < 0.01 by Student’s t test.

Figure 3

Impaired D2KO brown adipocyte differentiation. A, Brown preadipocytes were isolated from iBAT of WT and D2KO mice and differentiated in culture. Percentage of differentiated brown adipocytes determined by immunocytochemistry after staining with BODIPY 493/503. B, Treatment of D2KO preadipocytes cultures with 50 nm T3 during the early stages of differentiation (d 0–4) restores the WT percentage differentiation at d 10. C, Mitochondrial content in WT and D2KO d 10 brown adipocytes by quantification of Cox1/2 and Cox8 gDNA by qRT-PCR, expressed as mitochondrial/genomic DNA ratio. D, O₂ consumption of WT and D2KO d 10 brown adipocyte cultures in response to increasing concentrations of forskolin. A–D, Values are mean ± sem of 3–30 data points unless otherwise indicated. *, P < 0.05; and **, P < 0.01 vs. WT (or as indicated) by Student’s t test.

Figure 4

Oxidative stress in D2KO embryonic BAT. A, Expression of genes related to oxidative stress response processes in iBAT from WT, D2Het, and D2KO embryos at embryonic d E16.5, E17.5, and E18.5. mRNA levels determined by qRT-PCR are graphed relative to E16.5 WT expression. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 vs. WT of respective day by one-way ANOVA with Dunnet’s Multiple correction. B, Lipid peroxidation in iBAT lysates from E18.5 WT and D2KO littermates as indicated by immunoblotting for malonaldehyde (MDA). α-Tubulin shown as loading control. C, Average CM-H₂DCFDA fluorescence in d 0 brown preadipocytes after quantification with flow cytometry. ***, P < 0.001 by Student’s t test.

Figure 5

ROS causes decreased insulin signaling. A and B, WT and D2KO d 0 brown preadipocytes were serum starved for 20 h, treated for 5 min with varying doses of insulin, and levels of pAkt (Ser473), total Akt, and α-tubulin determined by immunoblotting. C and D, Immunoblotting of pAkt (S473) and α-tubulin in vehicle-treated WT and D2KO preadipocytes, as well as WT preadipocytes treated with rT3 since differentiation. Images are from different regions of same gel. E, Immunoblot analysis of d 0 serum starved WT and D2KO brown preadipocytes for phospho-IRS1 (S307) and α-tubulin. F, Immunoblot of phosphorylated IκBα in extracts from d 2 WT and D2KO brown preadipocytes. G and H, Treatment with the antioxidant ascorbic acid restores phosporylation of pAkt (S473) in D2KO brown preadipocytes to WT levels. I, Analysis of WT and D2KO preadipocytes differentiated with adipogenic cocktail, as described in text. Fractional number of brown adipocytes quantified by immunocytochemistry as previously described. Values are mean ± sem of 2–4 data points. B, D, and H, Quantification of Akt (Ser473) phosphorylation by normalization to α-tubulin levels and total signal on each Western blot. Values are mean ± sem of 3–5 data points. *, P < 0.05 by Student’s t test (B and H). **, P < 0.01 by one way ANOVA with Newman-Keuls Multiple Comparison (D).

Figure 6

Proposed model of positive feedback involving Dio2, PGC-1α, and UCP1 expression during BAT development. Schematic representation of the proposed role of D2 and D2-generated T3 in the development of brown adipocytes. D3, which decreases thyroid hormone signaling, is highest in the undeveloped brown preadipocyte. As the brown preadipocyte matures, D2, by enhancing thyroid hormone signaling, increases expression of PGC-1α, which coactivates TR, leading to enhanced UCP1 expression. Notably, Dio2 is also up-regulated by increased T3-signaling. These changes provide the mature brown adipocyte with its thermogenic function and also limit oxidative stress. If oxidative stress goes unchecked, then insulin signaling and adipogenesis may be altered.