Altered sleep architecture, rapid eye movement sleep, and neural oscillation in a mouse model of human chromosome 16p11.2 microdeletion

Abstract

Sleep abnormalities are common among children with neurodevelopmental disorders. The human chr16p11.2 microdeletion is associated with a range of neurological and neurobehavioral abnormalities. Previous studies of a mouse model of human chr16p11.2 microdeletion (chr16p11.2df/+) have demonstrated pathophysiological changes at the synapses in the hippocampus and striatum; however, the impact of this genetic abnormality on system level brain functions, such as sleep and neural oscillation, has not been adequately investigated. Here, we show that chr16p11.2df/+ mice have altered sleep architecture, with increased wake time and reduced time in rapid eye movement (REM) and non-REM (NREM) sleep. Importantly, several measurements of REM sleep are significantly changed in deletion mice. The REM bout number and the bout number ratio of REM to NREM are decreased in mutant mice, suggesting a deficit in REM-NREM transition. The average REM bout duration is shorter in mutant mice, indicating a defect in REM maintenance. In addition, whole-cell patch clamp recording of the ventrolateral periaqueductal gray (vlPAG)-projecting gamma-aminobutyric acid (GABA)ergic neurons in the lateral paragigantocellular nucleus of ventral medulla of mutant mice reveal that these neurons, which are important for NREM-REM transition and REM maintenance, have hyperpolarized resting membrane potential and increased membrane resistance. These changes in intrinsic membrane properties suggest that these projection-specific neurons of mutant mice are less excitable, and thereby may play a role in deficient NREM-REM transition and REM maintenance. Furthermore, mutant mice exhibit changes in neural oscillation involving multiple frequency classes in several vigilance states. The most significant alterations occur in the theta frequency during wake and REM sleep.

Keywords: GABAergic neurons; LPGi; REM sleep; chr16p11.2 microdeletion; neural oscillation; theta rhythm; vlPAG.

Figure 1.

Schematic diagram of the electrode configuration, setup of polysomnography, and representative traces of EEG and EMG, and hypnograms. (A) Schematic diagram showing the positions of EEG miniscrews (highlighted in yellow) and EMG electrodes (represented by wires). (B) Photographs showing a mouse under polysomnographic recording during the day and night. (C) Representative traces of wake EEG and EMG as well as power spectral plot of EEG. Notice the low amplitude EEG and flat power spectral plot. (D) Representative traces of NREM EEG and EMG as well as power spectral plot of EEG. Notice the relatively high amplitude EEG and a prominent peak in the delta frequency (0.5–4 Hz) on the power spectral plot. (E) Representative traces of REM EEG and EMG as well as power spectral plot of EEG. Notice the low amplitude EEG and a prominent peak in the theta frequency (6–9 Hz) on the power spectral plot. The EEG and EMG traces between two adjacent vertical lines are 4 s in duration. The spectral plots shown in C, D, and E are derived from the EEG in the shaded epochs. The three vigilance and sleep states, that is, wake, NREM, and REM, are color coded for clarity. (F) Representative hypnograms of wildtype and mutant mice.

Figure 2.

Reduced NREM and REM sleep, and increased arousal in mutant mice. Mutant mice exhibited increased awake time (A), reduced total sleep time (B), reduced NREM sleep time (C), and reduced REM sleep time (D). The ratios of REM to total sleep time (E) and REM to NREM sleep time (F) are reduced in mutant mice. Time courses of wake (G), total sleep (H), NREM sleep (I), and REM sleep (J). Data are presented as mean ± SEM. Wildtype, n = 10. Mutant, n = 9. Statistical significance is determined by MANOVA (day and night in panels A–F), two-tailed, unpaired Student’s t-test (24 h in panels A–F), and two-way ANOVA with post hoc Bonferroni correction (panels G–J). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 3.

Bout analysis of wake, REM, and NREM sleep. Bout numbers of wake (A), NREM sleep (B), and REM sleep (C). (D) REM/NREM bout ratio. Bout durations of wake (E), total sleep (F), NREM sleep (G), and REM sleep (H). Data are presented as mean ± SEM. Wildtype, n = 10. Mutant, n = 9. Statistical significance is determined by two-tailed, unpaired Student’s t-test (24 h in panels A–H), and MANOVA (day and night in panels A–H). ns: not significant. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 4.

The distribution plot of REM and NREM bout duration. Distributions of NREM bout number (A and C) and normalized NREM percentage (B and D) as a function of NREM duration during the day (A and B) and night (C and D). Distributions of REM bout number (E and G) and normalized REM percentage (F and H) as a function of REM duration during the day (E and F) and night (G and H). Please note that the shortest NREM and REM duration bins are 16–32 s. Data are presented as mean ± SEM. Wildtype, n = 10. Mutant, n = 9. Statistical significance is determined by two-way ANOVA with post hoc Bonferroni correction. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 5.

Intrinsic membrane properties and current–voltage (I–V) relationship of vlPAG-projecting GABAergic neurons in LPGi of ventral medulla. (A) Schematic diagram highlighting the ventral lateral periaqueductal gray (vlPAG). (B) A representative fluorescent image showing bilateral vlPAG injected with CTB-488. (C) Schematic diagram of medulla demonstrating LPGi in ventral medulla (reference needed). Laterodorsal tegmental nucleus (LDT) and dorsal raphe nucleus (DRN) are shown as reference. (D) A representative fluorescent image illustrating unilateral LPGi labeled with retrograde CTB-488. The gigantocellular reticular nucleus (Gi), inferior olivary (IO) nucleus, and the nucleus of ambiguus (Amb) are shown for reference. (E) A representative fluorescent image showing a CTB-488 labeled LPGi neuron under patch clamp recording. (F) Post-recording detection of a patched neuron (filled with biocytin via patch pipette) with AMCA-conjugated avidin D. Notice that well-circumscribed contour of the neuron in contrast to irregular background staining. (G) Post-recording verification of Cre expression in the same patched neuron as in (F). Notice that multiple neurons are immunoreactive for Cre. (H) A merged image of (F) and (G). Colocalization of AMCA signal and Cre expression confirmed that the patched neuron is indeed GABAergic. (I) Resting membrane potential (RMP) is mildly reduced in mutant LPGi neurons. (J) Membrane resistance (Rm) is increased in mutant LPGi neurons. (K) Membrane capacitance (Cm) is indistinguishable between wildtype and mutant LPGi neurons. (L) Representative traces of current-voltage relationship in wildtype and mutant LPGi neurons. Both peak (denoted by arrows) and steady-state responses (the last 100 ms denoted by horizontal bars) are subjected to analysis. No statistical differences are present between wildtype and mutant LPGi neurons in peak current (M) and steady state (N) current responses. Schematic diagrams in panels A and C are from Paxinos and Frankin [46]. Arrows and insets in (E) to (H) denote and highlight a patched neuron. Scale bars are 200 microns in (A) to (D), and 30 microns in (E) to (H). Eighteen neurons from 6 wildtype mice and 15 neurons from 5 mutant mice were analyzed. Statistical significance was determined by two-tailed, unpaired Student’s t-test. ns: not significant, *p < 0.05, ***p < 0.001.

Figure 6.

Power spectral analysis. Normalized day-time spectral analysis of oscillation classes during wake (A), NREM (B), and REM sleep (C). Normalized nighttime spectral analysis of oscillation classes during wake (D), NREM (E), and REM sleep (F). Normalized day-time spectral analysis of 1 Hz bin during wake (G), NREM (H), and REM sleep (I). Normalized nighttime spectral analysis of 1 Hz bin during wake (J), NREM (K), and REM sleep (L). The oscillation frequencies are defined as delta (0.5–4 Hz), theta (6–9 Hz), alpha (9–12 Hz), beta (12–30 Hz), and gamma (30–59 Hz). Data are presented as mean ± SEM. Wildtype, n = 9. Mutant, n = 9. Statistical significance is determined by Mann–Whitney test (panels A–F), and two-way ANOVA with post hoc Bonferroni correction (panels G–L). ns: not significant. *p < 0.05, **p < 0.01, ***p < 0.001.