Differential rescue of light- and food-entrainable circadian rhythms

Abstract

When food is plentiful, circadian rhythms of animals are powerfully entrained by the light-dark cycle. However, if animals have access to food only during their normal sleep cycle, they will shift most of their circadian rhythms to match the food availability. We studied the basis for entrainment of circadian rhythms by food and light in mice with targeted disruption of the clock gene Bmal1, which lack circadian rhythmicity. Injection of a viral vector containing the Bmal1 gene into the suprachiasmatic nuclei of the hypothalamus restored light-entrainable, but not food-entrainable, circadian rhythms. In contrast, restoration of the Bmal1 gene only in the dorsomedial hypothalamic nucleus restored the ability of animals to entrain to food but not to light. These results demonstrate that the dorsomedial hypothalamus contains a Bmal1-based oscillator that can drive food entrainment of circadian rhythms.

Fig. 1

Bmal1 expression in SCN rescues circadian rhythms and light entrainment in Bmal1^−/− mice. Representative body temperature (T_b) records (double-plotted actigrams) and period analyses for (A) Bmal1^+/−, (B) Bmal1^−/−, and (C) Bmal1^−/− mice with AAV-BMAL1 injections into the SCN under both light-dark (LD) and constant darkness (DD) conditions with AL access to food. The power spectrum for period (right panel) shows a dominant frequency at 23.7 hours for both Bmal1 heterozygous littermates and mice with AAV-BMAL1 rescue in the SCN, and the lack of a ~24-hour harmonic in the Bmal1^−/− mice. (D to F) Bmal1 mRNA expression (by in situ hybridization) in the SCN in (D) Bmal1^+/− mice, (E) Bmal1^−/− mice, and (F) after AAV-BMAL1 rescue by injection into the SCN (all ZT18). Scale bar, 100 μm.

Fig. 2

Food entrainment of the T_b rhythm is rescued by AAV-BMAL1 injection into the DMH. Under DD conditions, Bmal1^+/− and Bmal1^−/− mice were placed under RF (food available CT4-8, gray vertical bar). (A) Bmal1^+/− but not (B) Bmal1^−/− mice demonstrated a clear rise in T_b ~2 to 3 hours before food availability under RF [(A) to (C) are waveforms showing mean ± SEM T_b on days 10 to 14 of RF for an individual mouse]. (C) After bilateral injection of AAV-BMAL1 into the DMH of Bmal1^−/− mice, there was a preprandial elevation in T_b, but not a circadian rise in T_b during the presumptive dark cycle (CT12-24) [compare (C) with (A)]. As seen in the summary data plot, during the 3-hour window preceding food availability in RF (D), Bmal1^−/− mice with DMH AAV-BMAL1 injections (n = 4, black trace) showed a comparable preprandial elevation in T_b (mean ± SEM) to that of Bmal1^−/+ mice (red trace), whereas noninjected Bmal1^−/− mice (blue trace) showed no preprandial elevation in T_b; an increase in T_b occurred only after food presentation in the noninjected Bmal1^−/− mice (gray bar in D is first hour of food availability). Under RF, Bmal1^−/− mice demonstrated very low expression levels of Per1 mRNA (E) (CT4 shown) and no expression of Bmal1 mRNA at CT18 (G) (near-peak expression time in the heterozygote littermates) in the DMH. By contrast, after AAV-BMAL1 injection into the DMH (F and H), Bmal1^−/− mice demonstrated robust Per1 expression in the DMH at CT4 (F) and Bmal1 expression at CT18 (H).

Fig. 3

AAV-Bmal1 injection into the DMH rescues food entrainment but not light entrainment in Bmal1^−/− mice. (A) A double-plotted actigram of T_b from a Bmal1^−/− mouse after bilateral injections of AAV-Bmal1 into the DMH. These mice demonstrate the lack of entrainment to the light-dark (LD) cycle and persisting ultradian rhythmicity in constant darkness (DD) under AL feeding conditions (Fig. 1B). These same mice, however, demonstrated anticipation and entrainment (B) to a RF cycle (red line, food presentation) in DD conditions (arrow, last day, no food given). By contrast, noninjected Bmal1^−/− mice (C) did not demonstrate anticipation or entrainment to the RF cycle. (D to G) Per1 and Bmal1 gene expression (optical density, mean ± SEM; n = 3 mice per time point) in WT mice across the circadian day under both AL and RF (all data 12:12 LD) at 2-hour intervals, except for a 3-hour interval between ZT9-12. (D) Per1 expression in the SCN showed a peak at ~ ZT6 under both AL and RF (gray bar, ZT4-8) conditions, demonstrating that the SCN remained phase-locked to the LD cycle during RF. (E) Bmal1 expression in the SCN showed a peak at ~ZT18 under both AL and RF, further indicating that the SCN remained synchronized to the LD cycle during RF. (F) Per1 expression in the DMH was undetectable at all ZT under AL; by contrast, Per1 expression was sharply up-regulated by RF, with a peak at ~ZT7-8. (G) Bmal1 expression in the DMH was also undetectable at all ZT (except very modest expression ~ZT 3-5) under AL, however, and similar to Per1, during RF Bmal1 demonstrated up-regulation with a peak at ~ ZT18.