Phase-shifting the circadian glucocorticoid profile induces disordered feeding behaviour by dysregulating hypothalamic neuropeptide gene expression

Abstract

Here we demonstrate, in rodents, how the timing of feeding behaviour becomes disordered when circulating glucocorticoid rhythms are dissociated from lighting cues; a phenomenon most commonly associated with shift-work and transmeridian travel 'jetlag'. Adrenalectomized rats are infused with physiological patterns of corticosterone modelled on the endogenous adrenal secretory profile, either in-phase or out-of-phase with lighting cues. For the in-phase group, food intake is significantly greater during the rats' active period compared to their inactive period; a feeding pattern similar to adrenal-intact control rats. In contrast, the feeding pattern of the out-of-phase group is significantly dysregulated. Consistent with a direct hypothalamic modulation of feeding behaviour, this altered timing is accompanied by dysregulated timing of anorexigenic and orexigenic neuropeptide gene expression. For Neuropeptide Y (Npy), we report a glucocorticoid-dependent direct transcriptional regulation mechanism mediated by the glucocorticoid receptor (GR). Taken together, our data highlight the adverse behavioural outcomes that can arise when two circadian systems have anti-phasic cues, in this case impacting on the glucocorticoid-regulation of a process as fundamental to health as feeding behaviour. Our findings further highlight the need for development of rational approaches in the prevention of metabolic dysfunction in circadian-disrupting activities such as transmeridian travel and shift-work.

Fig. 1. Schematic of circadian and ultradian CORT infusion profiles.

Schematic representation of circadian and ultradian CORT infusion pattern in-phase (a) or out-of-phase (b). Dose of CORT (μM/3 h) for in-phase (c) and out-of-phase (d) infusions. The total amount of CORT given per day was equivalent between the two groups.

Fig. 2. Dysregulated feeding pattern during out-of-phase glucocorticoid exposure.

a Total food intake (g/day) over the course of the experiment; Repeated Measures 2-Way ANOVA (p = 0.47, infusion pattern; p < 0.01, day of treatment, p = 0.66, interaction between infusion pattern and day of treatment). b Twelve-hourly food intake was measured until day 4. Significant infusion pattern-dependent effects on timing of food intake (g/12 h); Repeated Measures 2-Way ANOVA (p = 0.47, infusion pattern; p < 0.01, time of day; p < 0.01, interaction between infusion pattern and time of day). c Percentage of food intake during light period (ZT0-ZT12) and dark period (ZT12–ZT24) for in-phase, out-of-phase, and control groups. d Cumulative measurements throughout the experiment showed a distinctive step-wise increment in pattern of food intake for in-phase group but not out-of-phase group. Repeated Measures 2-Way ANOVA (p < 0.01, infusion pattern; p < 0.01, time of day; p < 0.01, interaction between infusion pattern and time of day). Data are presented as mean ± SD, with individual datapoints representing biological repeats shown on each graph. Bonferroni multiple comparison post-test results are shown on the graphs; Significant one-to-one differences between timepoints (within the same treatment) indicated by *p < 0.05, **p < 0.01 and between treatment (at the same timepoint) indicated by ^#p < 0.05, ^##p < 0.01.

Fig. 3. Altered expression of hypothalamic feeding regulating neuropeptide genes during out-of-phase glucocorticoid exposure.

a Schematic showing analysed brain regions of the arcuate nucleus (ARC) and lateral hypothalamic area (LHA) (schematic adapted from image taken from rat brain atlas). b Representative digital images of ISHH of Npy, Agrp, Pmch, Pomc and Cartpt. Low magnification overview image with the area to be analysed indicated by the black dotted line, and higher magnification image with analysed area indicated by red dotted line. Scale bars indicate 500 μm (low magnification overview image) and 200 μm (higher magnification image). c Significant infusion pattern-dependent effects were found for the timing of gene expression for orexigenic peptides Npy, Agrp and Pmch; 2-Way ANOVA (Npy: p < 0.05, infusion pattern; p = 0.64, time of day; p < 0.01, interaction; Agrp: p < 0.01, infusion pattern; p = 0.62, time of day; p < 0.01, interaction; Pmch: p < 0.01, infusion pattern; p = 0.28, time of day; p < 0.01, interaction). d Significant infusion pattern-dependent effects were found for the timing of gene expression for anorexigenic peptides Pomc and Cartpt; 2-Way ANOVA (Pomc: p < 0.05, infusion pattern; p = 0.05, time of day; p < 0.01 interaction; Cartpt: p < 0.05, infusion pattern; p = 0.28, time of day; p < 0.01, interaction). Data are presented as mean ± SD, with individual datapoints representing biological repeats shown on each graph. Bonferroni multiple comparison post-test results are shown on the graphs; Significant one-to-one differences between timepoints (within the same treatment) indicated by *p < 0.05, **p < 0.01 and between treatment (at the same timepoint) indicated by ^#p < 0.05, ^##p < 0.01.

Fig. 4. Differential gene expression pattern of HPA axis regulators Crh and Pomc between in-phase and out-of-phase CORT exposure.

a Schematic showing analysed brain regions of the suprachiasmatic nucleus (SCN), supraoptic nucleus (SON), and paraventricular nucleus (PVN) (schematic adapted from image taken from rat brain atlas). b Representative digital images of ISHH of Crh in the PVN and Pomc in the anterior pituitary (ant PIT). Low magnification overview image with the area to be analysed indicated by the black dotted line, and higher magnification image with analysed area indicated by red dotted line. Scale bars indicate 500 μm (low magnification overview image) and 200 μm (higher magnification image). c Significant infusion pattern-dependent effects were found for the timing of Crh and Pomc gene expression; 2-Way ANOVA (Crh: p < 0.01, infusion pattern; p = 0.21, time of day; p = 0.61; interaction; Pomc: p < 0.01, infusion pattern; p < 0.01, time of day; p = 0.49; interaction). d Plasma adrenocorticotrophic hormone (ACTH, pg/mL) were measured from trunk blood samples at the end of the experiment. Significant infusion pattern-dependent effects were found for plasma ACTH concentration; 2-Way ANOVA (p < 0.01, infusion pattern; p < 0.01, time of day; p < 0.01, interaction). e Plasma corticosterone (CORT, ng/mL) were measured from the trunk blood samples at the end of the experiment. Significant infusion pattern-dependent effects were found for plasma CORT concentration; 2-Way ANOVA (p < 0.01, infusion pattern; p = 0.18, time of day; p < 0.01, interaction). Data are presented as mean ± SD, with individual datapoints representing biological repeats shown on each graph. Bonferroni multiple comparison post-test results are shown on the graphs; Significant one-to-one differences between timepoints (within the same treatment) indicated by *p < 0.05, **p < 0.01 and between treatment (at the same timepoint) indicated by ^#p < 0.05, ^##p < 0.01.

Fig. 5. Differential gene expression pattern of Oxt and Avp between in-phase and out-of-phase CORT exposure.

a Representative digital images of ISHH of oxytocin (Oxt) and arginine vasopressin (Avp) in the supraoptic nucleus (SON) and paraventricular nucleus (PVN). Low magnification overview image with the area to be analysis indicated by the black dotted line, and higher magnification image with analysed area indicated by red dotted line. Scale bars indicate 500 μm (low magnification overview image) and 200 μm (higher magnification image). b Significant infusion pattern-dependent effects were found for the timing of gene expression for Oxt and Avp; 2-Way ANOVA (Oxt (SON): p < 0.01, infusion pattern; p = 0.35, time of day; p < 0.01, interaction; Oxt (PVN): p < 0.05, infusion pattern; p = 0.25, time of day; p < 0.01, interaction; Avp (SON): p = 0.09, infusion pattern; p = 0.79, time of day; p < 0.01, interaction; Avp (PVN): p = 0.35, infusion pattern; p = 0.71, time of day; p < 0.01, interaction). Data are presented as mean ± SD, with individual datapoints representing biological repeats shown on each graph. Bonferroni multiple comparison post-test results are shown on the graphs; Significant one-to-one differences between timepoints (within the same treatment) indicated by *p < 0.05, **p < 0.01 and between treatment (at the same timepoint) indicated by ^#p < 0.05, ^##p < 0.01.

Fig. 6. Effects of dysregulated glucocorticoid exposure on core body temperature, but not locomotor activity.

a Averaged locomotor activity for in-phase and out-of-phase groups over 24 h. b Averaged core body temperature (°C) for in-phase and out-of-phase groups over 24 h. c Total locomotor activity in 6 h epochs were compared for in-phase and out-of-phase groups. Repeated Measures 2-Way ANOVA (p = 0.73, infusion pattern; p < 0.01, time of day; p < 0.05, interaction). d Averaged core body temperature (°C) in 6 h epochs were compared for in-phase and out-of-phase groups. Repeated Measures 2-Way ANOVA p < 0.01, infusion pattern; p < 0.01, time of day; p = 0.07, interaction Data are presented as mean ± SD, with individual datapoints representing biological repeats shown on each graph. Bonferroni multiple comparison post-test results are shown on the graphs; Significant one-to-one differences between timepoints (within the same treatment) indicated by *p < 0.05, **p < 0.01 and between treatment (at the same timepoint) indicated by ^#p < 0.05, ^##p < 0.01.

Fig. 7. Effects of dysregulated glucocorticoid exposure on Clock-regulating genes and Hypocretin neuropeptide precursor Orexin.

a Representative digital images of ISHH of Period 1 (Per1), Period 2 (Per2), and heteronuclear arginine vasopressin (hnAvp) in the suprachiasmatic nucleus (SCN) and Hypocretin neuropeptide precursor (Hcrt; previously known as Orexin) in the lateral hypothalamic area (LHA). Low magnification overview image with the area to be analysis indicated by the black dotted line, and higher magnification image with analysed area indicated by red dotted line. Scale bars indicate 500 μm (low magnification overview image) and 200 μm (higher magnification image). b Regulation of SCN Per1 and hnAvp; 2-Way ANOVA (Per1: p < 0.05, infusion pattern; p < 0.01, time of day; p = 0.90, interaction; hnAvp: p = 0.52, infusion pattern; p < 0.01, time of day; p = 0.26, interaction). c Regulation of SCN Per2; 2-Way ANOVA (p = 0.17, infusion pattern; p < 0.01, time of day; p = 0.36, interaction). d Regulation of Hcrt; 2-Way ANOVA (p = 0.61, infusion pattern; p < 0.01, time of day; p = 0.57, interaction). Data are presented as mean ± SD, with individual datapoints representing biological repeats shown on each graph. Bonferroni multiple comparison post-test results are shown on the graphs; Significant one-to-one differences between timepoints (within the same treatment) indicated by *p < 0.05, **p < 0.01.

Fig. 8. CORT-induced GR binding at GRE sites and pSer5 RNA Pol2 enrichment at TSS of Per1 and Npy genes in the arcuate nucleus.

a Region dissected from hypothalamus (co-ordinates bregma –2.0 mm to –5.0 mm, schematic adapted from image taken from rat brain atlas) used for collection of arcuate nucleus-enriched tissue (photograph by author) for ChIP processing. Scale bar indicates 2 mm. b CORT concentration in trunk blood collected from adrenalectomized rats at exactly 1 h after acute subcutaneous injection of either CORT (3 mg/kg)-HBC-Saline or HBC-Saline (Vehicle control). Significantly higher CORT levels were detected in CORT-injected compared to vehicle-injected controls (**p = 0.0085). c GR binding at the hypersensitive Per1 GRE was significantly induced by the acute CORT injection (**p = 0.0019). d pSer5 RNA Pol2 enrichment at Per1 TSS was significantly increased by the acute CORT injection (**p = 0.0010). e GR binding at the GRE upstream of the Npy gene was significantly induced by the acute CORT injection (**p = 0.0015). f pSer5 RNA Pol2 enrichment at Npy TSS was significantly increased by the acute CORT injection (*p = 0.0106). Data are presented as mean ± SD (n = 4 each).