Suppression of Circadian Timing and Its Impact on the Hippocampus

Abstract

In this article, I describe the development of the disruptive phase shift (DPS) protocol and its utility for studying how circadian dysfunction impacts memory processing in the hippocampus. The suprachiasmatic nucleus (SCN) of the Siberian hamster is a labile circadian pacemaker that is easily rendered arrhythmic (ARR) by a simple manipulation of ambient lighting. The DPS protocol uses room lighting to administer a phase-advancing signal followed by a phase-delaying signal within one circadian cycle to suppress clock gene rhythms in the SCN. The main advantage of this model for inducing arrhythmia is that the DPS protocol is non-invasive; circadian rhythms are eliminated while leaving the animals neurologically and genetically intact. In the area of learning and memory, DPS arrhythmia produces much different results than arrhythmia by surgical ablation of the SCN. As I show, SCN ablation has little to no effect on memory. By contrast, DPS hamsters have an intact, but arrhythmic, SCN which produces severe deficits in memory tasks that are accompanied by fragmentation of electroencephalographic theta oscillations, increased synaptic inhibition in hippocampal circuits, and diminished responsiveness to cholinergic signaling in the dentate gyrus of the hippocampus. The studies reviewed here show that DPS hamsters are a promising model for translational studies of adult onset circadian dysfunction in humans.

Keywords: Siberian; acetylcholine; dentate gyrus; hamster; memory; theta.

FIGURE 1

Development of the DPS protocol. (A) Light-dark (LD) cycles given by white and black rectangles, respectively. A phase advance or delay of the LD cycle is indicated by a plus (+) or min (–) sign, respectively, followed by phase shift duration in hours; L or D to indicate whether the shift was accomplished via a change in the light or dark phase, respectively (Aschoff et al., 1975). For example, a 5-h phase delay made by extending the light phase is indicated by –5L. ZT (zeitgeber time), ZT0 = baseline (BL) time of lights-on. (B) Coronal brain sections of mRNA hybridization for c-fos and per1 from hamsters that were entrained (ENT), free ran in the LD cycle (FR), or were circadian-arrhythmic (ARR). (C) The percent of animals that reentrained (left panel) and the duration of their active phases (α, right panel) after given a 2-h light pulse given early (E), middle (M), or late (L) in the night, and followed by a -3L shift on the following day. (D) Representative actograms with consecutive days double plotted from top to bottom of hamsters that reentrained (D1) or free-ran after -5L (D2,D3); arrows indicate day of -5L, gray shading indicates nighttime to visualize phase shift. (E) Representative actograms of animals that became arrhythmic after the DPS protocol (E1–E3, indicated by arrows). (F) Mean mRNA values of clock genes (per1, per2, bmal1, and cry1) quantified by RT-PCR from the SCN of ENT (red) and ARR (green) hamsters. Error bars removed for clarity. ZT0 = light onset in 16:8 LD cycle. Figures modified from Ruby et al., 1996, , , Barakat et al. (2004, and Grone et al. (2011).

FIGURE 2

SCN lesions appear to improve memory. (A) Rats were trained in a discrimination task that required them to obtain a food reward from one feeder in the morning and another feeder in the afternoon. Higher discrimination scores indicate better memory retention. Circles represent mean scores for each group by session. SCNx rats (blue) appeared to acquire the discrimination in fewer sessions than did sham-operated rats (gray). Figure modified from Mistlberger et al. (1996). (B) Rats in a passive-avoidance task learned to avoid entering a dark box where they previously received a foot shock. Longer latencies to reenter the dark box signify better memory of the shock. SCN lesions resulted in longer latencies and appeared to improve memory of the shock at 18 or 30 h after training (C, control; S, sham-lesioned; and L, SCN-lesioned). Figure modified from Stephan and Kovacevic (1978). (C) Rats were trained to replace a conditioned reflex to sound with one conditioned to a fixed time interval of 20 s. Training sessions consisted of 20 trials/day over 1 month. The authors statistical analysis showed that SCNx rats acquired the response significantly faster than sham-operated animals. Figure modified from Vodolazhskaya et al. (2001). All figures redrawn from published data and are for illustration purposes only.

FIGURE 3

Recognition and spatial working memory were impaired in DPS hamsters. (A) Performance of ENT hamsters (red) and DPS (green) hamsters in the NOR (A1) and SA (A2) tasks at different times of day. * indicates that task performance was significantly different from chance (P < 0.05; dotted line). In contrast to task performance, exploration behavior, defined as the amount of time spent exploring objects (B1) and number of arm entries (B2), did not change across the day and did not differ among ENT and DPS hamsters. NOR task used a 60 min interval between sample and test phases. Figure modified from Ruby et al. (2008, .

FIGURE 4

SCN lesions rescue object recognition and spatial memory in DPS hamsters. (A) Sequence of experiment stages. ENT hamsters performed the NOR and SA tasks, followed by exposure to the DPS protocol (red triangle). 4 weeks later, arrhythmic animals underwent SCN ablation or Sham surgeries (red triangle). (B) Representative actogram of an SCNx animal; memory tasks were performed at the end of each experimental stage. NOR task used a 24-h interval between sample and test phases. (C) Memory was rescued only in animals with complete SCN ablation (blue circles). No differences were found in NOR exploration times or in SA arm entries; all tests were performed within 3 h before dark onset. Figure modified from Fernandez et al. (2014).

FIGURE 5

A neuroanatomical model of a proposed SCN-septum-dentate circuit. For the sake of clarity, only the elements of the model discussed in the text are illustrated here. The projections from each structure are shown as gray arrows. The SCN has reciprocal connections with the ventral lateral septum (LSv). GABAergic fibers from the LSv innervate the medial septum (MS) shown here without the diagonal band of Broca. MS fibers containing Ach, GABA, and glutamate project to all subregions of the hippocampus, but only the dentate gyrus is shown. Fibers from the LSv to the SCN are primarily vasopressinergic (VP). See text for details and citations. Image credit: Allen Institute.

FIGURE 6

Loss of circadian timing fragments theta oscillations. (A) Power density in the EEG averaged over 5 min of object exploration. Delta (δ) and theta (θ) power bands defined by dotted lines. Data for ENT, DPS, SCNx, and Sham groups are aligned by peaks in theta power. Theta dominance was defined by a theta/delta (T/D) ratio > 1.0 within each EEG epoch. (B) Theta dominance in the EEG during exploration of a novel object given for (B1) total amount of time T/D > 1.0, (B2) number of theta-dominated episodes, and (B3) the mean duration of individual episodes where T/D was > 1.0. Data are mean (± SEM). * indicates P < 0.05, **P < 0.01 compared to ENT. Figure modified from Loewke et al. (2020).

FIGURE 7

Increased synaptic inhibition and reduced cholinergic responsiveness in the dentate gyrus of DPS hamsters. (A) Paired pulse ratios from granule cells in the dentate gyrus at different interstimulus intervals (ISI) during simultaneous recordings of (A1) field excitatory postsynaptic potentials (fEPSPs) and (A2) populations spikes (POP) under control (Con) or carbachol bath (Carb, 10 μM) conditions. Note that we use area rather than amplitude to quantify the POP spike. This is because amplitude can be confounded both by the number of cells excited by the stimulus and by the synchrony of those cells whereas POP spike area is not. The carbachol-induced population spike increase was attenuated in cells from DPS hamsters compared to cells from ENT animals at ISIs from 40–100 Hz (**P < 0.01). (B1) Carbachol normally suppresses evoked inhibition as it did here in the control group (ENT, red); however carbachol had no effect on cells from DPS hamsters (green) where evoked inhibition remained high (**P < 0.001). (B2) The number of spontaneous IPSP events in cells from DPS hamsters was 2x that observed in cells from ENT animals under control conditions at intervals < 500 ms (KS test, P = 0.014). Carbachol increased spontaneous IPSPs by the same magnitude in both groups. Figure modified from McMartin et al. (2021).

FIGURE 8

Nest building failed to elicit increases in dentate Ach in a DPS hamster. In vivo microdialysis was used to quantify Ach concentrations during nest building in a single hamster before and after the DPS protocol. A microdialysis probe inserted into the dentate gyrus withdrew cerebrospinal fluid (CSF) samples at 15 min intervals, beginning 45 min before nesting material was provided, and continuing for 60 min afterward. Neostigmine (200 nM) was added to the samples to prevent enzymatic degradation of Ach. (A) Single-plotted actogram of a hamster before and after the DPS protocol (red arrow). Days and times of microdialysis given (red and green circles). (B) Ach concentrations (nM). Dotted lines indicate time when the hamster was placed in the microdialysis chamber and when nesting material was provided. LLOQ, lower level of quantification. Samples quantified courtesy of Brains On-Line (San Francisco, CA, United States).