Induction of the metabolic regulator Txnip in fasting-induced and natural torpor

Abstract

Torpor is a physiological state characterized by controlled lowering of metabolic rate and core body temperature, allowing substantial energy savings during periods of reduced food availability or harsh environmental conditions. The hypothalamus coordinates energy homeostasis and thermoregulation and plays a key role in directing torpor. We recently showed that mice lacking the orphan G protein-coupled receptor Gpr50 readily enter torpor in response to fasting and have now used these mice to conduct a microarray analysis of hypothalamic gene expression changes related to the torpor state. This revealed a strong induction of thioredoxin-interacting protein (Txnip) in the hypothalamus of torpid mice, which was confirmed by quantitative RT-PCR and Western blot analyses. In situ hybridization identified the ependyma lining the third ventricle as the principal site of torpor-related expression of Txnip. To characterize further the relationship between Txnip and torpor, we profiled Txnip expression in mice during prolonged fasting, cold exposure, and 2-deoxyglucose-induced hypometabolism, as well as in naturally occurring torpor bouts in the Siberian hamster. Strikingly, pronounced up-regulation of Txnip expression was only observed in wild-type mice when driven into torpor and during torpor in the Siberian hamster. Increase of Txnip was not limited to the hypothalamus, with exaggerated expression in white adipose tissue, brown adipose tissue, and liver also demonstrated in torpid mice. Given the recent identification of Txnip as a molecular nutrient sensor important in the regulation of energy metabolism, our data suggest that elevated Txnip expression is critical to regulating energy expenditure and fuel use during the extreme hypometabolic state of torpor.

Figure 1.

Txnip expression is increased in the hypothalamus of Gpr50^−/− mice during torpor. (A) Representative recordings of oxygen consumption in Gpr50^−/− (black line) and WT (gray line) mice subjected to a 24-hour fast, during which Gpr50^−/− mice entered a state of deep torpor. (B) qRT-PCR analysis of Txnip expression in the hypothalamus of ad libitum-fed and fasted WT and Gpr50^−/− mice (n = 5/group; Gpr50^−/− mice n = 8) demonstrated significantly increased Txnip expression in fasted Gpr50^−/− mice. (C) In situ hybridization studies on mouse brain demonstrated Txnip expression within the ependyma of the lateral ventricles and the choroid plexus. Within the hypothalamus, Txnip expression was pronounced in the ependymal cells lining the third ventricle (thirdV). No hybridization signal was observed in Txnip^−/− mice or when sense riboprobe was used. (D) Quantification of the hybridization signal (n = 5/group) showed Txnip expression was similar between fed WT and Gpr50^−/− mice in both the ependyma of the thirdV and in the whole hypothalamic area minus the ependymal area. Fasting caused a significant increase in Txnip expression in both genotypes within the ependyma of the thirdV, but the induction of expression was significantly higher in the fasted Gpr50^−/− compared with WT fasted mice. Txnip levels showed a similar pattern of induction with fasting in the parenchyma of the hypothalamus. (E) Immunoblot analysis of hypothalamic lysates from ad libitum-fed and fasted WT and Gpr50^−/− mice. Densitometric analysis showed significantly increased Txnip expression in fasted Gpr50^−/− mice. Torpor was assessed by T_b at the time of killing. (F) Txnip immunoreactivity in the murine hypothalamus was limited to the ependymal cells lining the third ventricle and was lost in Txnip^−/− mouse tissue and when no primary antibody was included. Blood vessels, including those of the median eminence, were labeled by the secondary antimouse antibody. Data shown are mean ± SEM; *P < .05, **P < .01, ***P < .001 fasted vs fed; #P < .05, ##P < .01, Gpr50^−/− vs WT. Statistical significance was determined using 2-way ANOVA with Bonferroni's post hoc test. Txnip expressions of individual mice are also plotted against T_b at the time of tissue collection. KO, knockout (GPR50^−/−).

Figure 2.

Peripheral expression of Txnip during torpor. qRT-PCR analysis of Txnip expression in peripheral tissues of fed and fasted WT and Gpr50^−/− mice revealed that fasting did not alter Txnip expression in WT mice in liver or WAT but significantly increased expression in BAT. In Gpr50^−/− mice, Txnip expression was significantly increased by fasting in all tissues. qRT-PCR data are normalized to mouse 18S rRNA control and fold change relative to WT fed animals. Data shown are mean ± SEM; **P < .01, ***P < .001 fasted vs fed; ##P < .01, ###P < .001 Gpr50^−/− vs WT. Statistical significance was determined using 2-way ANOVA with Bonferroni's post hoc test. Txnip expressions of individual mice are also plotted against T_b at the time of tissue collection. KO, knockout (GPR50^−/−).

Figure 3.

Tissue-specific regulation of Txnip expression after energetic challenge. qRT-PCR analysis of Txnip expression in the tissues of WT mice treated with 1500-mg/kg 2DG for 1 hour showed significantly increased Txnip expression in the hypothalamus (A), WAT (C), and BAT (D) but no change in the liver (B). qRT-PCR analysis of Txnip expression in the tissues of WT and Gpr50^−/− mice exposed to 4°C for 2 hours or maintained at 20°C–22°C (RT, room temperature) revealed that although cold challenge did not alter the expression of Txnip in the hypothalamus or liver (E and F), expression was significantly increased in WAT (G) and decreased in BAT (H). Data are normalized to mouse 18S rRNA control and fold change relative to vehicle-treated animals (A–D) or RT-housed WT animals (E–H). Data shown are mean ± SEM; *P < .05, **P < .01 Student's t test or 2-way ANOVA with Bonferroni's post hoc test. Txnip expressions of individual mice are also plotted against T_b at the time of tissue collection. KO, knockout (GPR50^−/−).

Figure 4.

Txnip expression is increased in the tissues of WT mice during torpor. (A) Txnip expression in the brain of ad libitum-fed and 48-hour fasted WT female mice housed at 16°C was examined by in situ hybridization (n = 6/group; 2 fasted normothermic [NT] mice). Induction of torpor by fasting caused a significant elevation in Txnip expression within the hypothalamus, compared with fed and NT fasted mice. Txnip expressions of individual mice are plotted against T_b at the time of tissue collection. (B) qRT-PCR analysis of Txnip expression in the peripheral tissues of ad libitum-fed and fasted WT female mice housed at 16°C again showed that induction of torpor by fasting caused a significant elevation of Txnip expression compared with ad libitum-fed and fasted NT mice. qRT-PCR data shown are normalized to mouse 18S rRNA control and fold change relative to WT fed animals. Data shown are mean ± SEM; **P < .01, ***P < .001 fasted vs fed; #P < .05, ##P < .01, ###P < .001 torpid vs normothermic. Statistical significance was determined using 2-way ANOVA with Bonferroni's post hoc test.

Figure 5.

Txnip expression is increased in the tissues of Siberian hamsters during torpor. (A) Representative T_b recordings of Siberian hamsters maintained under a SD photoperiod and temperature, during which some hamsters spontaneously exhibit daily torpor (black line) or remain normothermic (gray line). Dashed line shows ambient temperature. (B) Txnip expression in the brains of Siberian hamsters housed in different photoperiods (LDs vs SDs, during which hamsters entered torpor [T] or remained normothermic [NT]), was examined by in situ hybridization (n = 4/group). Induction of torpor in SD-housed hamsters caused significant elevation in Txnip expression within the ependymal layer of cells lining the third ventricle, compared with LD-housed and NT-SD animals. Txnip expression in liver was also measured by qRT-PCR, and as in the hypothalamus, Txnip was significantly increased in torpid vs LD and NT-SD animals. Data shown are mean ± SEM; **P < .01, ***P < .001 LD vs SD; #P < .05, ##P < .01, ###P < .001 SD torpid vs NT-SD. Statistical significance was determined using 2-way ANOVA with Bonferroni's post hoc test.