The dorsomedial hypothalamic nucleus as a putative food-entrainable circadian pacemaker

Abstract

Temporal restriction of feeding can phase-shift behavioral and physiological circadian rhythms in mammals. These changes in biological rhythms are postulated to be brought about by a food-entrainable oscillator (FEO) that is independent of the suprachiasmatic nucleus. However, the neural substrates of FEO have remained elusive. Here, we carried out an unbiased search for mouse brain region(s) that exhibit a rhythmic expression of the Period genes in a feeding-entrainable manner. We found that the compact part of the dorsomedial hypothalamic nucleus (DMH) demonstrates a robust oscillation of mPer expression only under restricted feeding. The oscillation persisted for at least 2 days even when mice were given no food during the expected feeding period after the establishment of food-entrained behavioral rhythms. Moreover, refeeding after fasting rapidly induced a transient mPer expression in the same area of DMH. Taken in conjunction with recent findings (i) that behavioral expression of food-entrainable circadian rhythms is blocked by cell-specific lesions of DMH in rats and (ii) that DMH neurons directly project to orexin neurons in the lateral hypothalamus, which are essential for proper expression of food-entrained behavioral rhythms, the present study suggests that DMH plays a key role as a central FEO in the feeding-mediated regulation of circadian behaviors.

Fig. 1.

Food-anticipatory locomotor activity under RF. Hourly plots of distance traveled under ALF, RF on day 9, and RF followed by fast (RF-F1, see Results and Methods). The period of food availability is shaded. The dark phase is indicated by a solid horizontal bar. ∗, P < 0.05 ALF vs. RF; +, P < 0.05 ALF vs. RF-F1; #, P < 0.05 RF vs. RF-F1; by one-way repeated-measures ANOVA and Turkey post hoc tests.

Fig. 2.

Circadian expression of mPer2 mRNA in the DMH and NTS/AP of mice under RF. Serial coronal brain sections, prepared from mice under ALF or RF at ZT1, ZT7, ZT13, and ZT19 on day 9, were hybridized in situ to a ³⁵S-labeled mPer2 antisense probe. Representative sections from duplicate mice at the level of the DMH and NTS/DVM/AP are shown, together with coronal diagrams from the mouse brain atlas (39). (Scale bar: 0.5 mm.)

Fig. 3.

Localization of mPer1-expressing neurons in the DMH of mice under RF. Coronal brain sections were hybridized in situ to digoxigenin-labeled mPer1 antisense probe. (a and b) Low-magnification view of coronal sections containing the DMH (a) and the SCN (b) from a mouse under RF followed by fast (RF-F1) at ZT7 on day 7. (c and d) Representative higher-magnification images of mPer1-expressing cells in the DMH (c) and the SCN (d) in the fields shown by rectangles in a and b, respectively. (e–h) Coronal sections, prepared from a mouse under RF at ZT5, were dual-labeled by in situ hybridization for mPer1 (blue staining in perinuclear area) coupled with immunohistochemistry for a neuron-specific antigen NeuN (brown staining in nuclei). Cells with intense mPer1 expression are clustered in the ventromedial region of the DMHc (e–g in successive sections). Designated in lines are the whole DMH and its compact part. Arc, arcuate nucleus; VMH, ventromedial hypothalamic nucleus; 3v, third ventricle. (h) A representative image with higher magnification shows that mPer1-expressing cells in the DMH are neurons. (Scale bars: a and b, 1 mm; c and d, 0.1 mm; e–g, 0.2 mm; h, 0.05 mm.)

Fig. 4.

Detailed temporal patterns of circadian mPer1 expression in the DMH under RF. (a) Coronal brain sections, prepared from mice under ALF, RF, or RF followed by fast (RF-F1) on day 7 or under prolonged fasting on day 8 after RF-F (RF-F2) at time points shown in ZT, were hybridized in situ to a digoxigenin-labeled mPer1 antisense probe. Representative sections from duplicate mice at the level of the DMH and SCN are shown. Note that samples at ZT1 and ZT4 under RF are shared by the RF-F1 condition. (Scale bar: 0.2 mm.) (b and c) Numbers of mPer1-expressing cells in the ventromedial part of the DMHc. Values from two mice per data point (three mice for ZT7 under RF-F1) are dot-plotted; averages of duplicate counts are line-plotted. (b) Comparison of cell numbers under ALF, RF, and RF-F1. The period of food availability is shaded. (c) Food-entrained circadian oscillation of mPer1 expression in the DMHc sustains for at least 2 days without food. The preceding four time points (ZT1, ZT7, ZT13, and ZT19) are the same as those of RF-F1 in b, followed by three time points (ZT1, ZT7, and ZT13) on the second day of prolonged fasting (RF-F2). At the end of the experiments, mice under RF-F2 fasted for 53 h since the last food availability. For comparison, values under ALF, the same as those in b, are double-plotted.

Fig. 5.

Rapid induction of mPer1 in the DMHc by feeding. (a) Coronal brain sections, prepared from mice on the first day under RF (RF1) or fasting (F) at time points shown in ZT, were hybridized in situ to a digoxigenin-labeled mPer1 antisense probe. Representative sections from duplicate mice at the level of the DMH and SCN are shown. Note that samples at ZT1 and ZT4 under RF1 are shared by the fasting condition. (Scale bar: 0.2 mm.) (b) Numbers of mPer1-expressing cells in the ventromedial part of the DMHc. Values from two mice per data point (one mouse for ZT13 under RF1) are dot-plotted; averages of duplicate counts are line-plotted. Food was removed at ZT8 preceding time points shown in the graph, and returned, only under the RF1 condition, during the period shaded.

Fig. 6.

Feeding-dependent up-regulation of mPer1 in the NTS, DMV, and AP. Coronal brain sections, prepared from mice at ZT4 and ZT5 under RF or RF followed by fast (RF-F1) on day 7, or RF (RF1) or fasting (F) on day 1, were hybridized in situ to a digoxigenin-labeled mPer1 antisense probe. Representative sections from duplicate mice are shown, together with a coronal diagram from the mouse brain atlas (39). Note that samples at ZT4 are shared between RF and RF-F1 and between RF1 and F. (Scale bar: 0.2 mm.)

Fig. 7.

The DMH as a food-entrainable circadian oscillator. When food is freely available (ALF), the light-entrainable oscillator in the SCN predominates in orchestrating circadian behavioral and physiological rhythms according to the environmental photic cycle. The DMH mediates circadian information from the SCN and also receives feeding information directly and indirectly via humoral and neural pathways. When food availability is restricted to a period scheduled at a fixed time of the day (RF), a FEO in the DMH entrains to feeding information, uncouples from the SCN, and starts regulating food-entrainable circadian rhythms. Dotted lines from the SCN designate that the predominance of SCN-derived signals is contingent on food availability. LH, lateral hypothalamus; OC, optic chiasm; PBN, parabrachial nucleus; PVH, paraventricular hypothalamic nucleus.