Restoring RCAN1 Dosage Mitigates Sleep and EEG Abnormalities in a Down Syndrome Model

Abstract

Background.

Approximately 60% of individuals with Down syndrome (DS) have sleep abnormalities independent of breathing obstruction. Previously, we demonstrated increased wakefulness and decreased NREM sleep in the Dp(16)1Yey+ (Dp16) DS model mouse. In this study, we determined if increased RCAN1 levels mediate sleep disruption in Dp16 mice.

Methods.

We examined sleep architecture and electroencephalogram (EEG) patterns in young and aged Dp16 mice in which we genetically restored disomic levels of Rcan1 (Dp16 Rcan1^2N, shortened to Dp16–2N). Approximately equal numbers of male and female mice were used for each age group. We also examined gene expression and anxiety-like behaviors in aged mice.

Results.

We found that young Dp16 and Dp16–2N mice differ slightly from WT mice in sleep architecture and EEG characteristics. However, with age, more severe sleep deficits manifest in Dp16 mice and are partially rescued by Rcan1 dosage correction. Aged Dp16 mice exhibit significantly less mean EEG total power across different sleep states and activity phases. In contrast, WT and Dp16–2N mice show no age-related differences across states and stages. Aged Dp16 mice diverge from aged WT mice across multiple wake and sleep frequency bands during dark and light phases. In contrast, the EEG characteristics of aged Dp16–2N differ only modestly from those of WT mice.

Conclusions.

Combined, our data demonstrate that restoring Rcan1 gene levels mitigates some sleep architecture disruptions and EEG differences observed in aged Dp16 mice.

Introduction

Down syndrome (DS) cause by triplication of all or part of human chromosome 21 (Hsa21) (Desai, 1997; Parker et al., 2010) is the most common genetic cause of intellectual disability (Antonarakis et al., 2004; Freeman et al., 2008). Approximately 60% of DS individuals experience sleep disturbances (Fan et al., 2017), which arise early and worsen with age (Fuca et al., 2022; Hanna et al., 2022). While obstructive sleep apnea (OSA) contributes (Gimenez et al., 2021), OSA-independent disruptions such as delayed non-rapid eye movement (NREM) onset, sleep fragmentation, and reduced REM sleep are also frequent (Andreou et al., 2002; Fernandez and Edgin, 2013) and contribute to impaired cognition and memory (Heller and Ruby, 2019). DS individuals also display abnormal electroencephalogram (EEG) oscillations linked to cognitive deficits (Lopez-Loeza et al., 2016; Velikova et al., 2011) which may serve as biomarkers of neural dysfunction (Levenga et al., 2018; Salem et al., 2015).

Several mouse models have been used to study DS-related sleep phenotypes. The Ts65Dn model exhibits increased wakefulness and reduced NREM sleep (Colas et al., 2008) and includes a non-syntenic Mmu17 segment, whereas Dp(16)1Yey/+ (Dp16) mice carry a duplication orthologous to Hsa21 (Liu et al., 2011; Yu et al., 2010) and are free from unrelated loci. We previously reported that aged Dp16 mice show increased wakefulness and reduced NREM sleep compared to wild-type (WT) controls (Levenga et al., 2018). Among the triplicated genes on Hsa21 is Regulator of Calcineurin 1 (Rcan1) (Hoeffer et al., 2007), chronically overexpressed in DS brains (Sun et al., 2011). RCAN1 modulates Calcineurin A (CaN) activity as both inhibitor and facilitator (Fuentes et al., 2000; Wong et al., 2015), and its overexpression leads to cognitive impairments and neurodegeneration (Wong et al., 2015). RCAN1’s role in sleep regulation is evolutionarily conserved: its Drosophila homolog sarah (srh) is essential for sleep and circadian function (Kweon et al., 2018; Nakai et al., 2011) and altered RCAN1 expression affects circadian rhythmicity in mice (Rotter et al., 2014; Wong et al., 2022).

Given our prior findings that aged Dp16 mice have disturbed sleep (Levenga et al., 2018) and altered circadian activity linked to RCAN1 (Wong et al., 2022), we asked whether sleep abnormalities appear in young Dp16 mice and whether normalizing Rcan1 dosage rescues these deficits. Using in vivo EEG to assess sleep architecture in young and aged Dp16 mice and Dp16–2N animals with restored RCAN1 expression, we found that young (3–6 months) Dp16 mice exhibit mild sleep and EEG abnormalities, while aged (10–18 months) Dp16 mice show pronounced disruptions. Restoring Rcan1 to two copies significantly improved sleep and EEG measures in aged Dp16–2N mice.

To probe potential mechanisms underlying altered sleep and EEG rescue dynamics, we performed hippocampal transcriptomic analyses and identified genotype-dependent enrichment of GABAergic and neuroimmune pathways. Because sleep disruption in DS is frequently associated with anxiety-related phenotypes (Chawla et al., 2020; Fuca et al., 2023; Fuca et al., 2022; Ong et al., 2024), we also assessed open-field and elevated plus maze behavior in aged cohorts. We found sex-specific increases in both anxiety-like behavior and activity levels in aged Dp16 animals. Critically, these behaviors were reduced by the correction of Rcan1 gene dose.

Together, these findings indicate that Rcan1 dosage contributes to age-emergent sleep, EEG, transcriptomic, and anxiety-like abnormalities in Dp16 mice and identify RCAN1 as an important modulator of DS-related sleep dysfunction.

Methods

Mice

Dp(16)1Yey/+ (Dp16) mice (Yu et al., 2010) on a C57BL/6J background were bred with Rcan1 heterozygous mice (Vega et al., 2003) to generate wild-type (WT), Dp16, and Dp16 littermates carrying only two copies of the Rcan1 gene (Dp16–2N) as previously described ((Wong et al., 2022) and Fig. 1). All experiments were conducted using age-matched male and female littermates and from colonies backcrossed several generations in a C57BL/6J background (See Supplementary Materials for more detail). WT, Dp16, and Dp16–2N mice were divided into two cohorts by age: young mice (3–4 months) and aged mice (10–18 months) and grouped by genotype. In this study, Ns = 18 young WT mice, 15 young Dp16 mice, 17 young Dp16–2N mice, 23 aged WT mice, 27 aged Dp16 mice, and 19 aged Dp16–2N mice were used. Mice were group housed in plastic cages measuring 31×18×18 cm on a 12:12h light:dark cycle, with lights on at 07:00 and an ambient temperature of 22–24°C. Food and water were available ad libitum. After headstage implantation surgery and during recovery and EEG recording, mice were individually housed. This was to prevent injury to the subject after surgery due to cage mate aggression and to prevent damage to the headstage from cage mates during recording. Mice were individually housed for a maximum of 12–19 days. All procedures were approved by the University of Colorado Boulder Institutional Animal Care and Use Committee and conform to the National Institutes of Health guidelines.

Figure 1. Rcan1 gene dose correction brings both transcript and protein levels back to WT range.

(A) RNA-Seq data confirm that Dp16 mice have three copies of Rcan1, and Rcan1 transcript counts are brought to WT levels by Rcan1 gene dose correction in Dp16–2N mice. N: WT=12, Dp16=8, Dp16–2N=10 (B&C) Western blot analysis confirmed RCAN1 isoforms are overexpressed in Dp16 mice, and Rcan1 dose correction in Dp16–2N mice restores RCAN1 protein to WT levels (N = 6 mice/genotype). β-Actin = loading control. See Table 1 for a statistics summary.

Western Blotting

Brain tissues were isolated from adult 8-week-old mice from each experimental genotype and blotted using procedures described previously (Wong et al., 2015). Briefly, 20 μg of protein samples were prepared in Laemmli buffer, separated using 4–12% Bis–Tris gradient gels, and transferred to PVDF membranes. Blots were blocked for 1 h at room temperature (RT) and were incubated with primary antibodies to probe for RCAN1 (1:2000; Sigma), and β-actin (1:20000; Abcam) immunoreactivity. Blots were incubated for up to 72 hours at 4°C in 0.2% I-Block (Tropix) in Tris-buffered saline with 0.1% Tween-20 (TBS-T). Blots were then washed with TBS-T, incubated with HRP-conjugated goat anti-mouse or anti-rabbit secondary antibodies (1:5000–1:20 000, Promega) in I-Block solution at RT for 1 h, and washed again with TBS-T. Immunoreactive signals were imaged (ProteinSimple) using enhanced chemiluminescence (GE Healthcare) and normalized by loading control levels.

Surgery

Head-mount and electrode implantation were performed according to manufacturer protocols (Pinnacle Technology, Lawrence, KS) as described previously (Levenga et al., 2018). Briefly, a prefabricated head mount (Pinnacle Technology, #8201) with two EEG channels and one EMG channel was affixed to the skull with cyanoacrylate and four stainless steel screws (2.5 mm anteriorly and 3.0 mm posteriorly, Pinnacle Technology #8209 and #8212) which also served as EEG electrodes. The front edge of the head mount was placed 3.0 mm rostral of bregma to ensure all recording electrodes rested on the cerebral cortex. EMG wire electrodes were inserted into the nuchal muscles. The head mount was secured with dental cement (Metabond, Parkell, NY). All animals received buprenorphine (1 mg/kg) with physiological saline (0.1 ml/kg) subcutaneously before concluding surgery. Animals were individually housed in recording chambers (20 cm high x 25.4 cm diameter, Pinnacle Technology) with nesting material, food, and water ad libitum and allowed 7–10 days to recover and acclimate (Levenga et al., 2018).

Recording Apparatus

Before EEG/EMG recording, animals were habituated for 3–5 days in the recording chamber, which contained the recording cable. The recording cable connected a six-pin preamplifier (Pinnacle Technology #8202) in the head mount to a swivel commutator (Pinnacle Technology #8204), which was linked to a secondary amplifier and data conditioner (Pinnacle Technology DCAS #8206). EEG and EMG signals were sampled at 1000 Hz. The EEG signal was low-pass filtered at 100 Hz and high-pass filtered at 0.1 Hz; the EMG signal was low-pass filtered at 6 kHz and high-pass filtered at 3 Hz. Data were collected over two 24-h periods using Sirenia Sleep Pro (Pinnacle Technology V1.7.4).

Electroencephalogram analysis

As previously described (Levenga et al., 2018), the initial scoring of recorded EEG data used semi-automated cluster analyses based on EEG and EMG activity over 24 h. These recordings were divided into 10-s epochs that were designated as awake (wake: high frequency, low amplitude EEG, variable EMG activity), slow-wave sleep (NREM: low frequency, high amplitude EEG, low EMG activity) or rapid-eye-movement sleep (REM: high frequency, low amplitude EEG, no EMG activity) states. To ensure accuracy, two independent evaluators blind to genotype manually analyzed the epochs with Sirenia Sleep Pro software (Pinnacle Technology) and determined sleep states. A bout was defined as at least two consecutively scored epochs in either wake, NREM, or REM without an intervening transition. The time in each state and the length of bouts were quantified using Sirenia Sleep Pro (Pinnacle Technology). EEG data between 0.5 and 60 Hz from artifact-free epochs were analyzed with a Fast Fourier transform (FFT) using a Hann window function to produce a spectral analysis in 0.25 Hz bins. Bins were grouped into standard frequency bands, including delta (0.5–4 Hz), theta (4–8 Hz), alpha (8–15 Hz), beta (15–30 Hz), and gamma (30–60 Hz) for each animal (Sirenia Sleep Pro, Pinnacle Technology). The data for each animal were calculated by averaging the scores from two days of recording. To control the differences in power per bandwidth across animals, relative power per bandwidth was calculated as a percentage of the total power summed across 0.5 to 60 Hz for each subject. Total power per genotype by state and phase was determined by calculating the mean total power of all mice in the group (e.g., young Dp16, wake, light phase) summed over 0.5 to 60 Hz. Mean relative power per bandwidth was calculated for genotype, sleep state, and phase.

Transcriptomics

To identify differences in gene expression between male and female WT, DP16, and Dp16–2N mice aged 10–18 months, we performed RNA sequencing as previously described (Lombardi et al., 2025). Female (WT N=7, DP16 N=4, Dp16–2N N=5) and male (WT N=5, DP N=4, Dp16–2N N=5) mice were perfused and their hippocampi were extracted and flash frozen on dry ice for RNA extraction. RNA was extracted using TRIzol reagent after homogenization with a Dounce homogenizer, per the manufacturer’s instructions. Universal Plus mRNA-Seq library preparation kit with NuQuant was used to create Poly(A)-selected libraries on extracted RNA. Libraries were sequenced on Illumina NovaSeqX to produce >40M 150bp paired-end FASTQ reads. Reads were trimmed and filtered for quality using fastp (Chen et al., 2018), aligned to the MMU10 Mus musculus genome using STAR (Dobin et al., 2013), and read counts were generated using featureCounts (Liao et al., 2014). Differential gene expression analysis was performed using DESeq2 with a significance cut-off of BH-corrected adjusted p<0.1 (Love et al., 2014). Gene set enrichment analysis was performed using the ClusterProfiler package (Yu et al., 2012). Transcriptomic results are available in Supplemental tables 1 and 2.

Behavioral Assays

Behavioral tests were performed as described previously (Wong et al., 2020). Briefly, the order of behavioral tests was as follows: open-field arena (OFA) followed by elevated plus maze (EPM) in 5 independent cohorts with mixed genotypes testing. For OFA, mice explore a white Plexiglas arena (40 × 40 cm²) for 10 min with 180 lux overhead lighting and 55 dB of white noise in the testing room throughout. Data was collected and analyzed using the Ethovision XT video-tracking system (Noldus, Wageningen, Netherlands), with the center zone defined as the area 10 cm from the arena walls. For EPM, a white 30-cm arm-length EPM arena was used for testing. Mice were placed in the center zone of the maze and activity was recorded for 5 min by a video camera. Subject movements were analyzed using EthoVision XT (Noldus). Illumination levels during testing were maintained at 195 lux, with 55 dB of white noise present in the testing room throughout. Male and female mice were tested and analyzed independently for both behavioral measures to prevent confounds in the testing room (Wong et al., 2020).

Statistical Analysis

Statistical analyses were performed with SPSS (IBM Corp., Armonk, NY), Prism (GraphPad, Boston, MA), and Excel (Microsoft Corp., WA). Main effects were examined using repeated measures ANOVA (rANOVA), one-way ANOVA, and two-way ANOVA, and comparisons between groups were made with Tukey’s HSD test with an alpha level of 0.05. Data are presented as the mean ± SEM and are available on request. Table 1 includes a summary of statistical analyses generated for this study.

TABLE 1 –

Statistics Summary

Figure	Dependent Variable	Genotype	Sex	Genotype x Sex	Partial Eta2	N/group and post hoc testing
1A Two Way ANOVA	RCAN1 transcript counts	F(2,24)=33.40 p<.001	F(1,24)=9.88 p=0.004 Dp16–2N females express more RCAN1 than males; p=0.037	F(2,24)=0.77 p=0.475	Geno=0.736 Sex=0.292 Geno x Sex=0.060	WT=11 (5M, 6F) Dp16=9 (5M, 4F) Dp16–2N=10 (5M, 5F) RCAN1 levels WT vs. Dp16 p<.001 Dp16 vs Dp16–2N p<.001 WT vs Dp16–2N p=.685
1C Two Way ANOVA	RCAN1 protein levels	F(2,12)=7.745 p=0.007	F(1,12)=0.057 p=0.815	F(2,12)=1.187 p=0.339	Geno=0.563 Sex=0.005 Geno x Sex= 0.165	WT=6 (3M, 3F) Dp16=6 (3M, 3F) Dp16–2N=6 (3M, 3F) RCAN1 levels WT vs. Dp16 p=.009 Dp16 vs Dp16–2N p= .024 WT vs Dp16–2N p=.841
2A Two-way ANOVA	WD	F(2,43)=3.432 p=0.041	F(1,43)=0.104 p=0.749	F(2,43)=0.797 p=0.457	Geno=0.138 Sex=0.002 GenoxSex=0.036	WT=18 (12M, 6F) Dp16=15 (8M, 7F) Dp16–2N=16 (9M, 7F) Wake Dark: WT vs. Dp16 p=0.026 Dp16 vs. Dp16–2N p=0.985 WT vs. Dp16–2N p=0.035 NREM Dark: WT vs. Dp16 p=0.017 Dp16 vs. Dp16–2N p=0.961 WT vs. Dp16–2N p=0.030
	WL	F(2,43)=1.522 p=0.230	F(1,43)=1.276 p=.265	F(2,43)=0.545 p=0.584	Geno=0.066 Sex=0.029 GenoxSex= 0.025
	NRD	F(2,43)=3.728 p=0.032	F(1,43)=0.464 p=0.499	F(2,43)=0.903 p=0.413	Geno=0.148 Sex=0.011 GenoxSex=0.040
	NRL	F(2,43)=1.089 p=0.346	F(1,43)=0.482 p=0.491	F(2,43)=0.225 p=0.800	Geno=0.048 Sex=0.011 GenoxSex=0.010
	RD	F(2,43)=0.045 p=0.942	F(1,43)=2.279 p=0.138	F(2,43)=1.353 p=0.269	Geno=0.003 Sex=0.050 GenoxSex=0.059
	RL	F(2,43)=1.546 p=0.225	F(1,43)=0.893 p=0.350	F(2,43)=0.524 p=0.596	Geno=0.067 Sex=0.020 GenoxSex=0.024
2B Two-way ANOVA	WD	F(2,64)=9.477 p<0.001	F(1,64)=4.586 p=0.036	F(2,64)=3.956 p=0.024	Geno=0.228 Sex=0.067 GenoxSex= 0.110	WT=23 (12M, 11F) Dp16=27 (14M, 13F) Dp16–2N=20 (9M, 11F) Wake Dark: WT vs. Dp16 p<0.001 Dp16 vs. Dp16–2N p=0.011 WT vs. Dp16–2N p=0.667 Wake Light: WT vs. Dp16 p=0.024 Dp16 vs. Dp16–2N p=0.005 WT vs. Dp16–2N p=0.797 NREM Dark: WT vs. Dp16 p<0.001 Dp16 vs. Dp16–2N p=0.016 WT vs. Dp16–2N p=0.360 NREM Light: WT vs. Dp16 p=0.003 Dp16 vs. Dp16–2N p=0.011 WT vs. Dp16–2N p=0.937 REM Light: WT vs. Dp16 p=0.064 WT vs. Dp16–2N p=0.005 Genotype x Sex Wake Dark: Dp16 F vs. WT M p =0.053 Dp16 F vs. WT F p=0.065 Dp16 F vs. Dp16–2N M p=0.002 Dp16–2N F vs. Dp16–2N M p=0.016 Dp16–2N M vs. Dp16 F p=0.002 Dp16–2N M vs. Dp16 M p=0.002 Dp16 M vs. WT F p=0.070 Dp16 M vs. WT M p=0.058 NREM Dark Dp16 F vs. Dp16−2N M p=0.004 Dp16 F vs. WT F p=0.028 Dp16 F vs. WT M p=0.026 Dp16–2N F vs. Dp16–2N M p=0.026 Dp16–2N M vs. Dp16 M p=0.005 Dp16 M vs. WT F p=0.033 Dp16 M vs. WT M p=0.030
	WL	F(2,64)=6.226 p=0.003	F(1,64)=0.827 p=0.366	F(2,64)=0.178 p=0.837	Geno=0.163 Sex=0.013 GenoxSex= 0.006
	NRD	F(2,64)=10.637 p<0.001	F(1,64)=4.071 p=0.048	F(2,64)=3.573 p=0.034	Geno=0.249 Sex=0.060 GenoxSex= 0.100
	NRL	F(2,64)=7.346 p=0.001	F(1,64)=1.244 p=0.229	F(2,64)=0.114 p=0.892	Geno=0.187 Sex=0.019 GenoxSex= 0.004
	RD	F(2,64)=1.662 p=0.198	F(1,64)=1.348 p=0.250	F(2,64)=1.261 p=0.290	Geno=0.049 Sex=0.021 GenoxSex= 0.038
	RL	F(2,64)=5.594 p=0.006	F(1,64)=0.130 p=0.720 Dp16–2N females have greater WD (p=0.004) and less NRD than Dp16–2N males (p=0.004)	F(2,64)=0.056 p=0.945	Geno=0.149 Sex=0.002 GenoxSex=0.002

Figure 3 Bouts	State Phase	Genotype	Sex	Geno x Sex	Partial Eta2	N/group Post hoc testing
3A Young Mean Bouts Two-way ANOVA	WD	F(2,43)=0.096 p=0.908	F(1,43)=0.138 p=0.712	F(2,43)=1.049 p=0.359	Geno=0.004 Sex=0.003 GxS=0.045	WT=18 (12M, 6F) Dp16=15 (8M, 7F) Dp16–2N=17 (10M, 7F) REM Dark: WT vs. Dp16 p=0.042 WT vs. Dp16–2N p=0.049 Dp16 vs. Dp16–2N p<.001 REM Light: WT vs. Dp16–2N p=.010
	WL	F(2,43)=0.499 p=0.611	F(1,43)=1.376 p=0.247	F(2,43)=1.268 p=0.291	Geno=0.022 Sex=0.030 GxS=0.055
	NRD	F(2,43)=0.413 p=0.664	F(1,43)=0.918 p=0.343	F(2,43)=1.261 p=0.294	Geno=0.018 Sex=0.020 GxS=0.054
	NRL	F(2,43)=1.032 p=0.365	F(1,43)=0.980 p=0.328	F(2,43)=1.015 p=0.371	Geno=0.045 Sex=0.022 GxS=0.044
	RD	F(2,43)=10.833 p<0.001	F(1,43)=0.144 p=0.706	F(2,43)=0.886 p=0.420	Geno=0.330 Sex=0.003 GxS=0.039
	RL	F(2,43)=5.321 p=0.009	F(1,43)=1.689 p=0.201	F(2,43)=1.639 p=0.206	Geno=0.195 Sex=0.037 GxS=0.069
3B Young Mean Bout Length Two-way ANOVA	WD	F(2,43)=0.739 p=0.484	F(1,43)=4.822 p=0.033	F(2,43)=2.169 p=0.126	Geno=0.032 Sex=0.099 GxS=0.090	WT=18 (12M, 6F) Dp16=15 (8M, 7F) Dp16–2N=17 (10M, 7F) REM Dark: Dp16 vs. Dp16–2N p=.020 REM Light: Dp16 vs. Dp16–2N p=.005 Sex Effect: Females have longer Wake bouts in Dark phase than males
	WL	F(2,43)=1.293 p=0.285	F(1,43)=0.216 p=0.644	F(2,43)=0.229 p=0.796	Geno=0.056 Sex=0.005 GxS=0.010
	NRD	F(2,43)=3.133 p=0.053	F(1,43)=0.297 p=0.588	F(2,43)=0.728 p=0.489	Geno=0.125 Sex=0.007 GxS=0.032
	NRL	F(2,43)=0.023 p=0.977	F(1,43)=0.251 p=0.619	F(2,43)=0.313 p=0.733	Geno=0.001 Sex=0.006 GxS=0.014
	RD	F(2,43)=4.444 p=.017	F(1,43)=2.399 p=0.129	F(2,43)=0.848 p=0.435	Geno=0.168 Sex=0.006 GxS=0.037
	RL	F(2,43)=5.508 p=0.007	F(1,43)=0.852 p=0.361	F(2,43)=2.060 p=0.140	Geno=0.200 Sex=0.019 GxS=0.086
3C Aged Mean Bouts Two-way ANOVA	WD	F(2,63)=2.437 p=0.096	F(1,63)=0.812 p=0.371	F(2,63)=0.659 p=0.521	Geno=0.072 Sex=0.013 GxS=0.020	WT=23 (12M, 11F) Dp16=27 (13M, 13F) Dp16–2N =20 (8M, 12F) Wake Dark: WT vs. Dp16 p=0.089 Wake Light: WT vs. Dp16–2N p=0.039 NREM Dark: WT vs. Dp16 p=0.043 WT vs. Dp16–2N p=0.080 NREM Light: WT vs. Dp16 p=.049 WT vs. Dp16–2N p=0.007 REM Dark: WT vs. Dp16–2N p=0.071 REM Light: WT vs. Dp16–2N p=0.007 Dp16 vs. Dp16–2N p=0.021
	WL	F(2,63)=3.217 p=0.047	F(1,63)=0.369 p=0.546	F(2,63)=0.395 p=0.675	Geno=0.093 Sex=0.006 GxS=0.012
	NRD	F(2,63)=3.729 p=0.029	F(1,63)=0.953 p=0.333	F(2,63)=0.088 p=0.916	Geno=0.106 Sex=0.015 GxS=0.003
	NRL	F(2,63)=5.223 p=0.008	F(1,63)=0.424 p=0.517	F(2,63)=0.224 p=0.800	Geno=0.142 Sex=0.007 GxS=0.007
	RD	F(2,63)=2.672 p=0.077	F(1,63)=1.280 p=0.262	F(2,63)=0.951 p=0.392	Geno=0.020 Sex=0.020 GxS=0.029
	RL	F(2,63)=5.079 p=0.009	F(1,63)=0.333 p=0.566	F(2,63)=0.364 p=0.696	Geno=0.139 Sex=0.005 GxS=0.011
3D Aged Mean Bout Length Two-way ANOVA	WD	F(2,63)=1.460 p=0.240	F(1,63)=0.716 p=0.401	F(2,63)=1.988 p=0.145	Geno=0.044 Sex=0.011 GxS=0.059	WT=23 (12M, 11F) Dp16=27 (13M, 13F) Dp16–2N =20 (8M, 12F) NREM Dark: WT vs. Dp16 p<0.001 WT vs. Dp16–2N p<0.001 NREM Light: WT vs. Dp16 p<0.001 WT vs. Dp16–2N p=0.021 Sex Effect: Females have longer NREM bouts in Light phase than males REM Dark: WT vs. Dp16–2N p<0.001 Dp16 vs. Dp16–2N p=0.003 REM Light: WT vs. Dp16–2N p=0.064 Dp16 vs. Dp16–2N p=0.049
	WL	F(2,63)=0.169 p=0.845	F(1,63)=0.019 p=0.891	F(2,63)=1.845 p=0.167	Geno=0.005 Sex=0.0009 GxS=0.055
	NRD	F(2,63)=11.664 p<0.001	F(1,63)=0.939 p=0.336	F(2,63)=0.585 p=0.560	Geno=0.270 Sex=0.015 GxS=0.018
	NRL	F(2,63)=9.262 p<0.001	F(1,63)=4.188 p=0.047	F(2,63)=1.180 p=0.314	Geno=0.227 Sex=0.061 GxS=0.036
	RD	F(2,63)=9.347 p<0.001	F(1,63)=0.319 p=0.574	F(2,63)=0.763 p=0.470	Geno=0.229 Sex=0.005 GxS=0.024
	RL	F(2,63)=3.121 p=0.050	F(1,63)=0.516 p=0.475	F(2,63)=1.421 p=0.249	Geno=0.090 Sex=0.008 GxS=0.043

Figure 4	State Phase	Geno	Sex	Geno x Sex	Partial Eta2	N/group Post hoc testing
Young Avg Power by State Phase Two-way ANOVA	WD	Delta F(2,43)=4.940 p=0.012 Theta F(2,43)=0.684 p=0.510 Alpha F(2,43)=4.833 p=0.013 Beta F(2,43)=4.478 p=0.018 Gamma F(2,43)=2.746 p=0.077	Delta F(1,43)=0.000 p=0.990 Theta F(1,43)=0.514 p=0.485 Alpha F(1,43)=1.216 p=0.277 Beta F(1,43)=0.041 p=0.840 Gamma F(1,43)=0.001 p=0.972	Delta F(2,43)=0.504 p=0.608 Theta F(2,43)=0.618 p=0.544 Alpha F(2,43)=0.909 p=0.411 Beta F(2,43)=0.566 p=0.972 Gamma F(2,43)=1.042 p=0.362	Delta: G=0.202 S=0.00 GxS=0.025 Theta: G= 0.034 S=0.013 GxS=0.031 Alpha: G=0.199 S=0.030 GxS=0.045 Beta: G=0.187 S=0.001 GxS=0.051 Gamma: G=0.123 S=0.00 GxS=0.051	WT=14 (10M, 4F) Dp16=14 (8M, 6F) Dp16–2N =16 (9M, 7F) Wake Dark Delta: WT vs. Dp16 p=0.009 Dp16 vs. Dp16–2N p=0.018 Alpha: WT vs. Dp16–2N p=0.013 Beta: WT vs. Dp16 p=0.021 WT vs. Dp16–2N p=0.018 Gamma: WT vs. Dp16 p=0.020 NREM Dark Gamma: WT vs. Dp16 p=0.060 WT vs. Dp16–2N p=0.001 REM Dark Alpha WT vs. Dp16–2N p=0.001 Dp16 vs. Dp16–2N p=0.077 Beta WT vs. Dp16–2N p=0.073 Male vs Female Beta: p=0.007 Theta: p=0.073 For Beta Males > Female Wake Light Alpha: WT vs. Dp16 p=0.004 Dp16 vs. Dp16–2N p=0.056 Beta: WT vs. Dp16 p=0.042 WT vs. Dp16–2N p=0.016 Gamma: WT vs. Dp16 p=0.055 NREM Light Gamma: WT vs. Dp16 p=0.063 WT vs. Dp16–2N p=0.008 REM Light Alpha: WT vs. Dp16–2N p=0.055 Beta: WT vs. Dp16–2N p=0.007 Male vs Female Beta: p=0.002 Theta: p=0.086 (trend) For Beta Males > Female
	NRD	Delta F(2,43)=1.272 p=0.292 Theta F(2,43)=0.397 p=0.675 Alpha F(2,43)=2.019 p=0.147 Beta F(2,43)=0.867 p=0.428 Gamma F(2,43)=6.675 p=0.003	Delta F(1,43)=0.007 p=0.933 Theta F(1,43)=0.099 p=0.755 Alpha F(1,43)=0.038 p=0.847 Beta F(1,43)=0.753 p=0.391 Gamma F(1,43)=0.029 p=0.865	Delta F(2,43)=1.331 p=0.276 Theta F(2,43)=0.138 p=0.872 Alpha F(2,43)=0.150 p=0.861 Beta F(2,43)=1.092 p=0.346 Gamma F(2,43)=2.871 p=0.069	Delta: G=0.063 S=0.00 GxS=0.065 Theta: G=0.020 S=0.003 GxS=0.007 Alpha: G=0.096 S=0.001 GxS=0.008 Beta: G=0.044 S=0.019 GxS=0.054 Gamma: G=0.260 S=0.001 GxS=0.131
	RD	Delta F(2,43)=0.442 p=0.646 Theta F(2,43)=1.041 p=0.363 Alpha F(2,43)=5.358 p=0.009 Beta F(2,43)=2.763 p=0.076 Gamma F(2,43)=1.644 p=0.207	Delta F(1,43)=0.008 p=0.930 Theta F(1,43)=3.401 p=0.073 Alpha F(1,43)=1.146 p=0.291 Beta F(1,43)=8.092 p=0.007 Gamma F(1,43)=0.011 p=0.918	Delta F(2,43)=1.364 p=0.276 Theta F(2,43)=0.929 p=0.404 Alpha F(2,43)=0.692 p=0.507 Beta F(2,43)=0.164 p=0.850 Gamma F(2,43)=0.879 p=0.423	Delta: G=0.023 S=0.00 GxS=0.067 Theta: G=0.052 S=0.082 GxS=0.047 Alpha: G=0.220 S=0.029 GxS=0.035 Beta: G=0.127 S=0.176 GxS=0.009 Gamma: G=0.207 S=0.000 GxS=0.044
	WL	Delta F(2,43)=1.871 p=0.168 Theta F(2,43)=1.865 p=0.169 Alpha F(2,43)=5.057 p=0.011 Beta F(2,43)=4.200 p=0.022 Gamma F(2,43)=1.724 p=0.192	Delta F(1,43)=0.430 p=0.516 Theta F(1,43)=0.007 p=0.936 Alpha F(1,43)=1.482 p=0.231 Beta F(1,43)=0.157 p=0.684 Gamma F(1,43)=0.143 p=0.694	Delta F(2,43)=0.563 p=0.574 Theta F(2,43)=0.186 p=0.831 Alpha F(2,43)=1.001 p=0.377 Beta F(2,43)=0.077 p=0.926 Gamma F(2,43)=0.861 p=0.431	Delta: G=0.090 S=0.011 GxS=0.029 Theta: G=0.089 S=0.00 GxS=0.010 Alpha: G=0.210 S=0.038 GxS=0.050 Beta: G= 0.181 S=0.004 GxS=0.004 Gamma: G=0.083 S=0.004 GxS=0.043
	NRL	Delta F(2,43)=1.479 p=0.241 Theta F(2,43)=1.584 p=0.218 Alpha F(2,43)=2.057 p=0.142 Beta F(2,43)=1.462 p=0.244 Gamma F(2,43)=3.699 p=0.034	Delta F(1,43)=0.015 p=0.903 Theta F(1,43)=0.004 p=0.948 Alpha F(1,43)=0.006 p=0.938 Beta F(1,43)=0.433 p=0.515 Gamma F(1,43)=0.866 p=0.358	Delta F(2,43)=1.172 p=0.321 Theta F(2,43)=0.506 p=0.607 Alpha F(2,43)=0.128 p=0.880 Beta F(2,43)=1.593 p=0.217 Gamma F(2,43)=2.627 p=0.085	Delta: G=0.072 S=0.00 GxS=0.058 Theta: G=0.077 S=0.00 GxS=0.026 Alpha: G=0.098 S=0.00 GxS=0.007 Beta: G= 0.071 S=0.011 GxS=0.077 Gamma: G=0.163 S=0.022 GxS=0.121
	RL	Delta F(2,43)=0.585 p=0.562 Theta F(2,43)=1.405 p=0.258 Alpha F(2,43)=1.913 p=0.162 Beta F(2,43)=4.555 p=0.017 Gamma F(2,43)=0.977 p=0.386	Delta F(1,43)=0.001 p=0.977 Theta F(1,43)=3.113 p=0.086 Alpha F(1,43)=1.204 p=0.279 Beta F(1,43)=10.568 p=0.002 Gamma F(1,43)=0.408 p=0.527	Delta F(2,43)=1.458 p=0.245 Theta F(2,43)=1.520 p=0.232 Alpha F(2,43)=1.165 p=0.323 Beta F(2,43)=1.382 p=0.263 Gamma F(2,43)=0.504 p=0.608	Delta: G=0.030 S=0.00 GxS=0.071 Theta: G=0.069 S=0.076 GxS=0.074 Alpha: G=0.091 S=0.031 GxS=0.058 Beta: G=0.193 S=0.218 GxS=0.068 Gamma: G=0.049 S=0.011 GxS=0.026

Figure 5	State Phase	Geno	Sex	Geno x Sex	Partial Eta2	N/group Post hoc testing
Aged Avg Power by State Phase Two-way ANOVA	WD	Delta F(2,48)=5.032 p=0.011 Theta F(2,48)=2.690 p=0.111 Alpha F(2,48)=2.195 p=0.093 Beta F(2,48)=7.524 p=0.002 Gamma F(2,48)=4.002 p=0.025	Delta F(1,48)=2.355 p=0.132 Theta F(1,48)=1.499 p=0.227 Alpha F(1,48)=3.325 p=0.075 Beta F(1,48)=0.087 p=0.769 Gamma F(1,48)=6.422 p=0.015	Delta F(2,48)=1.888 p=0.164 Theta F(2,48)=0.755 p=0.476 Alpha F(2,48)=2.406 p=0.102 Beta F(2,48)=1.301 p=0.906 Gamma F(2,48)=1.669 p=0.200	Delta: G=0.190 S=0.052 GxS= 0.081 Theta: G= 0.111 S=0.034 GxS=0.034 Alpha: G=0.093 S=0.072 GxS=0.101 Beta: G=0.259 S=0.002 GxS=0.057 Gamma: G=0.157 S=0.130 GxS=0.072	WT=18 (10 M, 9F) Dp16=15 (6M, 9F) Dp16–2N =16 (9M, 7F) Wake Dark Delta: WT vs. Dp16 p=0.018 Dp16 vs. Dp16–2N p=0.012 Theta: WT vs. Dp16–2N p=0.075 Beta: WT vs. Dp16 p=0.076 WT vs. Dp16–2N p<0.001 Gamma: WT vs. Dp16 p=0.009 Sex Effect: Females have higher WD gamma than males NREM Dark Delta: WT vs. Dp16 p=0.004 WT vs. Dp16–2N p=0.002 Theta: WT vs. Dp16 p=0.019 WT vs. Dp16–2N p=0.012 Gamma: WT vs. Dp16–2N p<0.001 Dp16 vs. Dp16–2N p=0.042 REM Dark Delta: WT vs. Dp16 p=0.053 Theta: WT vs. Dp16 p=0.006 Dp16 vs. Dp16–2N p=0.024 Alpha: WT vs. Dp16–2N p=0.071 Wake Light Beta: WT vs. Dp16 p=0.003 WT vs. Dp16–2N p=0.005 Gamma: WT vs. Dp16 p=0.072 NREM Light Delta: WT vs. Dp16 p<0.001 WT vs. Dp16–2N p<0.001 Theta: WT vs. Dp16 p=0.009 WT vs. Dp16–2N p=0.009 Alpha: WT vs. Dp16 p=0.019 Sex Effect: Females have higher NRL gamma than males Alpha Geno x Sex Dp16–2N F vs. Dp16–2N M p=0.050 Dp16–2N F vs. WT F p=0.076 Dp16–2N F vs. WT M p=0.019 WT M vs Dp16 M p=0.069 Dp16–2N females have higher alpha power than WT or Dp16–2N males Gamma: WT vs. Dp16 p<0.001 WT vs. Dp16–2N p<0.001 Dp16 vs. Dp16–2N p=0.027 REM Light Delta Geno x Sex Dp16–2N F vs. Dp16–2N M p=0.048 Dp16–2N females have lower delta power Alpha: WT vs. Dp16–2N p=0.016 Beta: WT vs. Dp16–2N p=0.063 Gamma: WT vs. Dp16–2N p=0.081
	NRD	Delta F(2,48)=8.396 p<0.001 Theta F(2,48)=5.303 p=0.009 Alpha F(2,48)=2.492 p=0.095 Beta F(2,48)=1.548 p=0.224 Gamma F(2,48)=10.379 p<0.001	Delta F(1,48)=1.034 p=0.315 Theta F(1,48)=3.640 p=0.063 Alpha F(1,48)=0.304 p=0.584 Beta F(1,48)=0.800 p=0.376 Gamma F(1,48)=0.012 p=0.912	Delta F(2,48)=0.077 p=0.926 Theta F(2,48)=0.035 p=0.966 Alpha F(2,48)=1.286 p=0.287 Beta F(2,48)=0.186 p=0.831 Gamma F(2,48)=0.362 p=0.699	Delta: G=0.281 S=0.023 GxS= 0.004 Theta: G= 0.198 S=0.078 GxS=0.002 Alpha: G=0.104 S=0.007 GxS=0.000 Beta: G=0.067 S=0.018 GxS=0.009 Gamma: G=0.326 S=0.00 GxS=0.017
	RD	Delta F(2,48)=3.181 p=0.051 Theta F(2,48)=5.255 p=0.009 Alpha F(2,48)=2.193 p=0.124 Beta F(2,48)=2.506 p=0.093 Gamma F(2,48)=1.827 p=0.173	Delta F(1,48)=0.002 p=0.966 Theta F(1,48)=1.119 p=0.296 Alpha F(1,48)=0.918 p=0.343 Beta F(1,48)=1.133 p=0.293 Gamma F(1,48)=0.524 p=0.473	Delta F(2,48)=1.536 p=0.227 Theta F(2,48)=0.298 p=0.744 Alpha F(2,48)=1.322 p=0.277 Beta F(2,48)=0.240 p=0.788 Gamma F(2,48)=0.262 p=0.770	Delta: G=0.129 S=0.00 GxS= 0.067 Theta: G= 0.196 S=0.025 GxS=0.014 Alpha: G=0.093 S=0.007 GxS=0.000 Beta: G=0.067 S=0.021 GxS=0.058 Gamma: G=0.078 S=0.012 GxS=0.012
	WL	Delta F(2,48)=1.257 p=0.295 Theta F(2,48)=0.846 p=0.436 Alpha F(2,48)=0.117 p=0.890 Beta F(2,48)=8.021 p<0.001 Gamma F(2,48)=2.234 p=0.119	Delta F(1,48)=0.368 p=0.547 Theta F(1,48)=0.657 p=0.422 Alpha F(1,48)=0.983 p=0.327 Beta F(1,48)=0.383 p=0.539 Gamma F(1,48)=1.116 p=0.297	Delta F(2,48)=0.849 p=0.435 Theta F(2,48)=0.202 p=0.818 Alpha F(2,48)=1.073 p=0.351 Beta F(2,48)=0.174 p=0.841 Gamma F(2,48)=1.002 p=0.376	Delta: G=0.055 S=0.008 GxS= 0.038 Theta: G= 0.038 S=0.015 GxS=0.009 Alpha: G=0.005 S=0.022 GxS=0.048 Beta: G=0.272 S=0.009 GxS=0.008 Gamma: G=0.094 S=0.025 GxS=0.045
	NRL	Delta F(2,48)=13.912 p<0.001 Theta F(2,48)=5.799 p=0.006 Alpha F(2,48)=4.921 p=0.012 Beta F(2,48)=2.271 p=0.115 Gamma F(2,48)=29.845 p<0.001	Delta F(1,48)=2.852 p=0.099 Theta F(1,48)=0.938 p=0.338 Alpha F(1,48)=1.463 p=0.233 Beta F(1,48)=0.150 p=0.701 Gamma F(1,48)=4.346 p=0.043	Delta F(2,48)=0.733 p=0.486 Theta F(2,48)=0.767 p=0.471 Alpha F(2,48)=4.455 p=0.018 Beta F(2,48)=0.780 p=0.465 Gamma F(2,48)=0.476 p=0.625	Delta: G=0.393 S=0.062 GxS= 0.033 Theta: G= 0.212 S=0.021 GxS=0.034 Alpha: G=0.186 S=0.033 GxS=0.171 Beta: G=0.096 S=0.003 GxS=0.035 Gamma: G=0.581 S=0.092 GxS=0.022
	RL	Delta F(2,48)=0.563 p=0.574 Theta F(2,48)=1.143 p=0.328 Alpha F(2,48)=3.596 p=0.036 Beta F(2,48)=2.761 p=0.074 Gamma F(2,48)=2.537 p=0.091	Delta F(1,48)=0.484 p=0.490 Theta F(1,48)=0.028 p=0.867 Alpha F(1,48)=0.565 p=0.456 Beta F(1,48)=0.340 p=0.563 Gamma F(1,48)=0.577 p=0.452	Delta F(2,48)=3.251 p=0.048 Theta F(2,48)=1.305 p=0.282 Alpha F(2,48)=2.084 p=0.137 Beta F(2,48)=0.440 p=0.647 Gamma F(2,48)=0.318 p=0.729	Delta: G=0.026 S=0.011 GxS= 0.131 Theta: G= 0.050 S=0.001 GxS=0.057 Alpha: G=0.143 S=0.013 GxS=0.088 Beta: G=0.114 S=0.008 GxS=0.020 Gamma: G=0.106 S=0.013 GxS=0.015

Figure 6 Transcriptomics	Normalization method	N/Group	# Genes with adjusted p-value < 0.1
Male Dp16 vs. WT	DESeq2	Dp16=4 WT=5	89
Male Dp16–2N vs. WT	DESeq2	Dp16–2N =5 WT=5	91
Female Dp16 vs. WT	DESeq2	WT=7 Dp16=4	107
Female Dp16–2N vs. WT	DESeq2	WT=7 Dp16–2N =5	115
Male Dp16–2N vs. Dp16	DESeq2	Dp16–2N =5 Dp16=4	3
Female Dp16–2N vs. Dp16	DESeq2	Dp16=4 Dp16–2N =5	4

Figure 7 Behavior	Dep Variable	Sex	Geno	Partial Eta2	N/group Post hoc testing
One-way ANOVA (Aged Open Field Analysis)	Distance Moved	Male	F(2,36)=0.403 p=0.671	0.023	OFA WT=25 (14M, 11F) Dp16=22 (11M, 11F) Dp16–2N=21 (12M, 9F) Distance WT F vs. Dp16 F p=0.057 Peripheral/Center WT M vs. Dp16 M p=0.041 Dp16 M vs. Dp16–2N M p=0.085
	Distance Moved	Female	F(2,30)=2.993 p=0.066	0.176
	Periphery/Center Time	Male	F(2,36)=3.699 p=0.035	0.179
	Periphery/Center Time	Female	F(2,30)=1.594 p=0.221	0.102
One-way ANOVA (Aged Elevated Plus Maze)	Distance Moved	Male	F(1,33)=1.506 p=0.240	0.100	EPM WT=21 (10M, 11F) Dp16=20 (9M, 11F) Dp16–2N =19 (10M, 9F) Distance WT F vs. Dp16–2N F p<0.001 Dp16 F vs. Dp16–2N F p=0.006 Closed Arm Dp16 M vs. Dp16–2N M p=0.050 WT F vs. Dp16–2N F p=0.002 Dp16 F vs. Dp16–2N F p=0.011 Open Arm WT M vs. Dp16 M p=0.034 Dp16 M vs. Dp16–2N M p=0.044 WT F vs. Dp16–2N F p=0.024 Dp16 F vs Dp16–2N F p=0.015 Center Zone WT M vs. Dp16–2N M p=0.047
	Distance Moved	Female	F(1,29)=10.395 p<0.001	0.426
	Closed Arm Time	Male	F(1,29)=3.471 p=0.046	0.205
	Closed Arm Time	Female	F(1,29)=8.259 p=0.002	0.371
	Open Arm Time	Male	F(1,29)=4.388 p=0.022	0.245
	Open Arm Time	Female	F(1,30)=5.464 p=0.010	0.281
	Center Zone Time	Male	F(1,33)=3.908 p=0.032	0.224
	Center Zone Time	Female	F(1,30)=2.727 p=0.083	0.163

Supp Figure 1	Dependent Variable		%Time/Hr x Geno	%Time/Hr x Sex	%Time/Hr x Geno x Sex	Partial Eta2	N/Group and Post hoc testing
1A Rep Meas. -ANOVA	Young Wake	Between subjects	Light: F(2,43)=0.621 p=0.542 Dark: F(2,43)=3.392 p=0.043	Light: F(1,43)=1.035 p=0.315 Dark: F(1,43)=1.909 p=0.174	Light: F(2,43)=0.126 p=0.882 Dark: F(2,43)=0.697 p=0.504	Geno=0.028 Sex=0.024 GxS=0.006 Geno=0.136 Sex=0.043 GxS=0.031	WT=18 (12M, 6F) Dp16=15 (8M, 7F) Dp16–2N=16 (9M, 7F) Wake Dark: WT vs. Dp16 p=0.024 WT vs. Dp16–2N p=0.030
1B Rep Meas – ANOVA	Young NREM	Between subjects	Light: F(2,43)=0.669 p=0.517 Dark: F(2,43)=0.755 p=0.186	Light: F(1,43)=1.065 p=0.308 Dark: F(1,43)=1.751 p=0.390	Light: F(2,43)=0.088 p=0.916 Dark: F(2,43)=0.779 p=0.465	Geno=0.030 Sex=0.024 GxS=0.004 Geno=0.075 Sex=0.017 GxS=0.035
1C Rep Meas. - ANOVA	Young REM	Between subjects	Light: F(2,43)=2.663 p=0.081 Dark: F(2,43)= 0.040 p=0.961	Light: F(1,43)=0.597 p=0.444 Dark: F(1,43)=1.019 p=0.318	Light: F(2,43)=1.975 p=0.151 Dark: F(2,43)=1.014 p=0.371	Geno=0.110 Sex=0.014 GxS=0.084 Geno=0.002 Sex=0.023 GxS=0.045
1D Rep Meas.-ANOVA	Aged Wake	Between subjects	Light: F(2,64)=4.470 p=0.015 Dark: F(2,64)=5.276 p=0.008	Light: F(1,64)=0.323 p=0.572 Dark: F(1,64)=2.006 p=0.161	Light: F(2,64)=0.021 p=0.979 Dark: F(2,64)=2.031 p=0.140	Geno=0.123 Sex=0.005 GxS=0.001 Geno=0.142 Sex=0.030 GxS=0.060	WT=23 (12M, 11F) Dp16=27 (14M, 13F) Dp16–2N =20 (9M, 11F) Wake Light: WT vs. Dp16 p=0.046 Dp16 vs. Dp16–2N p=0.025 Wake Dark: WT vs. Dp16 p=0.009 Dp16 vs. Dp16–2N p=0.080 NREM Light: WT vs. Dp16 p=0.004 Dp16 vs. Dp16–2N p=0.053 NREM Dark: WT vs. Dp16 p=0.004 Dp16 vs. Dp16–2N p=0.063 REM Light: WT vs. Dp16 p=0.046
1E Rep Meas.-ANOVA	Aged NREM	Between subjects	Light: F(2,64)=5.890 p=0.004 Dark: F(2,64)=6.242 p=0.003	Light: F(1,64)=0.259 p=0.589 Dark: F(1,64)=2.135 p=0.149	Light: F(2,64)=0.062 p=0.940 Dark: F(2,64)=1.652 p=0.200	Geno=0.155 Sex=0.005 GxS=0.002 Geno=0.163 Sex=0.032 GxS=0.049
1F Rep Meas.-ANOVA	Aged REM	Between subjects	Light: F(2,64)=2.821 p=0.067 Dark: F(2,64)= 1.097 p=0.357	Light: F(1,64)=0.055 p=0.760 Dark: F(1,64)=0.124 p=0.726	Light: F(2,64)=0.379 p=0.729 Dark: F(2,64)=0.410 p=0.665	Geno=0.081 Sex=0.001 GxS=0.010 Geno=0.050 Sex=0.002 GxS=0.013

Supp Figure 2	Dependent Variable		Bouts/Hr x Geno	Bouts/Hr x Sex	Bouts/Hr x Geno x Sex	Partial Eta2	N/Group and Post hoc testing
2A Rep Meas. -ANOVA	Young Wake	Between subjects	Light: F(2,43)=0.083 p=0.920 Dark: F(2,43)=0.801 p=0.455	Light: F(1,43)=.490 p=0.488 Dark: F(1,43)=1.726 p=0.196	Light: F(2,43)=1.561 p=0.221 Dark: F(2,43)=1.412 p=0.254	Geno=0.004 Sex=0.011 GxS=0.066 Geno=0.035 Sex=0.038 GxS=0.060	WT=18 (12M, 6F) Dp16=15 (8M, 7F) Dp16–2N=16 (9M, 7F) REM Light Dp16 vs Dp16–2N <0.001 Dp16 vs WT = 0.045 Dp16–2N6 vs WT = 0.053 REM Dark Dp16 vs Dp16–2N = 0.061 Dp16–2N vs WT = 0.011
2B Rep Meas – ANOVA	Young NREM	Between subjects	Light: F(2,43)=0.406 p=0.669 Dark: F(2,43)=0.083 p=0.920	Light: F(1,43)=0.003 p=0.956 Dark: F(1,43)=0.490 p=0.488	Light: F(2,43)=0.990 p=0.380 Dark: F(2,43)=1.561 p=0.221	Geno=0.018 Sex=0.000 GxS=0.043 Geno=0.004 Sex=0.011 GxS=0.066
2C Rep Meas. - ANOVA	Young REM	Between subjects	Light: F(2,43)=10.512 p<0.001 Dark: F(2,43)=5.795 p=0.006	Light: F(1,43)=0.029 p=0.865 Dark: F(1,43)=1.662 p=0.204	Light: F(2,43)=0.324 p=0.725 Dark: F(2,43)=1.217 p=0.306	Geno=0.323 Sex=0.001 GxS=0.015 Geno=0.208 Sex=0.036 GxS=0.052
2D Rep Meas.-ANOVA	Aged Wake	Between subjects	Light: F(2,64)=2.893 p=0.063 Dark: F(2,64)=2.184 p=0.121	Light: F(1,64)=0.002 p=0.964 Dark: F(1,64)=0.951 p=0.333	Light: F(2,64)=2.834 p=0.067 Dark: F(2,64)=0.090 p=0.914	Geno=0.087 Sex=0.000 GxS=0.085 Geno=0.067 Sex=0.015 GxS=0.003	WT=23 (12M, 11F) Dp16=27 (14M, 13F) Dp16–2N =20 (9M, 11F) Wake Light Dp16 vs WT = 0.062 NREM Light Dp16 vs WT = 0.016 Dp16–2N vs WT = 0.003 NREM Dark Dp16–2N vs WT = 0.029 REM Light Dp16 vs Dp16–2N = 0.077 Dp16–2N vs WT = 0.005 REM Dark Dp16–2N− vs WT = 0.049

Supp Figure 5 PV Cell Count	Genotype	Sex	Sex x Genotype	Partial Eta2	N (mouse) and n (slice)/Group Post hoc Testing
2Way ANOVA (CA1 HPC)	F(2,43) = 5.525 p=0.009	F(1,43)=0.622 p=0.434	F(2,43)=0.476 p=0.625	Geno=0.198 Sex=0.0139 GenoxSex= 0.021	WT F – 8 and 50 Dp16 F – 7 and 42 Dp16–2N F – 10 and 64 WT M – 10 and 59 Dp16 M – 9 and 54 Dp16–2N M – 6 and 35 Male WT vs Dp16 p=0.018 WT vs Dp16–2N p=0.080
2Way ANOVA (RT)	F(2,40)=1.304 p=0.283	F(1,40)=4.399p=0.042	F(2,40)=0.839 p=0.436	Geno=0.061 Sex=0.099 GenoxSex= 0.040	WT F – 8 and 48 Dp16 F – 9 and 50 Dp16–2N F – 8 and 45 WT M – 8 and 48 Dp16 M – 5 and 29 Dp16–2N M – 8 and 48 WT Female vs Male – p=0.026

^*All data were assessed using Levene’s test of homogeneity and Kolmogorov–Smirnov test

RESULTS

Rcan1 dose correction restores RCAN1 in Dp16 mice to WT levels

Rcan1 dose correction was verified by hippocampal mRNA and protein analyses (Fig. 1). Two-way ANOVA of Rcan1 transcript counts identified significant genotype (F_(2,24) =33.40, p<0.001) and sex (F₍₁,₂₄₎=9.88, p=0.004) effects but no interaction effect. Tukey’s HSD showed higher Rcan1 transcript levels in Dp16 than in WT (p<0.001) and Dp16–2N (p<0.001) mice, with no difference between WT and Dp16–2N (p=.685) (Fig. 1A). Within the Dp16–2N genotype, females expressed more Rcan1 than males (p=0.037). We identified a significant genotype effect on protein levels (Two-way ANOVA: F_(2,12) = 7.745, p = 0.007) but no sex or interaction effects. Tukey’s HSD showed that protein levels mirrored mRNA: RCAN1 levels were higher in Dp16 compared with WT (p=0.009) and Dp16–2N (p=0.024) mice, with no difference between WT and Dp16–2N (p=0.841) (Fig. 1C). Thus, Rcan1 dose correction normalizes hippocampal RCAN1 levels in Dp16–2N mice, establishing molecular rescue. See Table 1 for additional statistics.

Young Dp16 mice exhibit early indicators of altered sleep

We determined percent time spent awake and in NREM and REM sleep during the dark (active; 19:00–07:00) and light (inactive; 07:00–19:00) phases (Fig. 2A; Supplemental Fig. 1A–C). In the dark, genotype exerted an effect on wake (Two-way ANOVA: F_(2,43)=3.432, p=0.041) but not sex nor genotype x sex interaction, and Tukey’s HSD showed that percent time spent awake was greater in young Dp16 (p=0.024) and Dp16–2N (p=0.030) mice. Genotype affected percent time in NREM in the dark (Two-way ANOVA: F_(2,43)=3.728, p=0.032), which was reduced in both Dp16 (p=0.026) and Dp16–2N (p=0.035) compared with WT mice. Percent time in REM did not differ by genotype. These findings indicate that mild sleep abnormalities are already present in young adults but are not rescued by Rcan1 dose correction at this age. Because we observed more severe sleep deficits in aged Dp16 mice (Levenga et al., 2018), we next tested if Rcan1 gene dose correction improved sleep in aged Dp16 mice.

Figure 2. Sleep abnormalities are detected in young Dp16 mice, and Rcan1 gene dose correction partially rescues sleep impairments in aged Dp16 mice.

(A) Young Sleep Architecture. Summary percent times in sleep state and activity phase for all young groups. Young N: WT=18; Dp16=15; Dp16–2N=17. (B) Aged Sleep Architecture. Summary percent times in sleep state and activity phase for all aged groups. Aged N: WT=23; Dp16=27; Dp16–2N=19. W/D = awake dark phase; W/L = awake light phase; NR/D = NREM dark phase; NR/L = NREM light phase; R/D = REM dark phase; R/L = REM light phase. *p<0.05; **p<0.01; *** p<0.001. See Table 1 for a statistics summary.

Sleep deficits in aged Dp16 mice rescued by Rcan1 dose correction

In aged mice there were main effects of genotype (Two-way ANOVA: F_(2,64)=9.477, p<0.001), sex (F_(1,64)=4.586, p=0.036) and a genotype x sex interaction (F_(2,64)=3.956, p=0.024) on percent time spent awake in the dark. In the light, there was a main effect of genotype on wake (Two-way ANOVA: F_(2,64)=6.226, p=0.003) but not sex nor an interaction. In aged cohorts (Fig. 2B, Table 1) during both phases, Tukey’s HSD confirmed that Dp16 spent a greater percentage of time awake than WT (dark p<0.001; light p=0.024) and Dp16–2N mice (dark p=0.011; light p=0.005) while Dp16–2N and WT did not differ. In the dark, Dp16–2N males were awake less than Dp16–2N females (p=0.004), Dp16 females (p=0.002), and Dp16 males (p=0.002) (Tukey’s HSD; Fig. 2B; Table 1).

Percent time spent in NREM during the dark phase was affected by genotype (Two-way ANOVA: F_(2,64)=10.637, p<0.001), sex (F_(1,64)=4.07, p=0.048) and genotype x sex interaction (F_(2,64)=3.573, p=0.034). Tukey’s HSD showed that in the dark, Dp16 exhibited less percent time in NREM than WT (p<0.001) and Dp16–2N (dark p=0.016), however Dp16–2N did not differ from WT. Dark phase NREM showed genotype x sex interaction effects (Table 1): Dp16 females had less NREM than WT females (p=0.028) and WT males (p=0.026); Dp16 males had less NREM than WT males (p=0.030), Dp16–2N males (p=0.005), and WT females (p=0.033); and Dp16–2N females had less NREM than Dp16–2N males (p=0.004). In the light phase, percent time spent in NREM was affected by genotype (Two-way ANOVA: F_(2,64)=7.346, p=0.001) but not sex or genotype by sex interaction. Dp16 mice spent less time in NREM than WT (p=0.003) and Dp16–2N (p=0.011) in the light. The percent time in REM did not show a main effect of genotype or sex in the dark. There was a main effect of genotype in the light (Two-way ANOVA: F_(2,64)=5.594, p=0.006), with Tukey’s HSD revealing that both Dp16 (p=0.046) and Dp16–2N (p=0.005) percent times in REM exceeded WT (Supplemental Fig. 1F; Table 1). These results show a pronounced wake versus NREM disruption in aged Dp16 mice, indicating fragmentation, which was largely corrected by Rcan1 dose normalization. Because both genotypes are generated as littermates from the same breeder pairs, any ‘hybrid vigor/depression’ introduced by breeding structure would be shared across genotypes rather than tracking with Rcan1 dosage.

Bout frequency and duration: minimal disruption in young mice, but age-related changes in Dp16 mice emerge

After identifying abnormalities in the percent time spent in wake, NREM and REM sleep by young and aged mice, we examined bout frequency and duration of wake/sleep states during dark and light phases. In young mice, genotype did not affect wake and NREM bout frequencies during the dark and light phases, (Fig. 3A; Supplemental Fig. 2A,B) but did affect REM bout frequency (Two-way ANOVA: F_(2,43)=10.833, p<0.001; Fig. 3A; Supplemental Fig. 2C; Table 1). Tukey’s HSD showed that in the dark, Dp16 mice had fewer REM bouts than WT (p=0.042), while Dp16–2N animals showed more than both WT (p=0.049) and Dp16 (p<0.001). Genotype also influenced REM bout frequency in the light (Two-way ANOVA: F_(2,43)=5.321, p=0.009), as Dp16–2N mice also showed more REM bouts than WT (Tukey’s HSD p=0.01). Wake and NREM bout durations were similar across phases and genotypes, while REM duration showed effects of genotype in the dark (Two-way ANOVA: F_(2,43)=4.444, p=0.017) and in the light (F_(2,43)=5.508, p=0.007). Sex differences in wake dark bout duration were detected (Two-way ANOVA: F_(1,43)= 4.822, p=0.043), with females having longer bouts than males. REM sleep bout duration increased in Dp16 over Dp16–2N in the dark (p=0.020) and light (p=0.005); (Tukey’s HSD, Fig. 3B; Supplemental Fig. 2C; Table 1).

Figure 3. Young WT, Dp16, and Dp16–2Nmice differed only in REM sleep bout number or length. Aged mice exhibited significant differences across genotypes, sleep states and phases: aged Dp16 mice differed significantly from WT in the mean number of NREM bouts and bout length; Dp16–2N mice differed from WT mice during the light phase in all sleep states and differed from Dp16 in light phase NREM and REM sleep.

(A) There were no significant genotype differences in mean bouts per hour among young mice across Wake and NREM activity phases. Dp16 mice had fewer REM bouts in the dark than both WT and Dp16–2Nwhile Dp16–2Nhad more REM bouts than WT. In the light, Dp16–2Nhad more REM bouts than WT. (B) Mean bout length in seconds also did not differ significantly among groups during Wake and NREM across activity phases. Dp16–2N had significantly shorter bouts than Dp16 both in the dark and light phases. Young N: WT=18; Dp16=15; Dp16–2N=17. (C) Aged Dp16–2N mice displayed more Wake and NREM bouts in the light than aged WT mice, while Dp16 mice showed more bouts of NREM sleep than WT mice in both dark and light phases. Dp16–2N mice also showed more NREM bouts in the light than Dp16. Aged Dp16–2N mice had more REM bouts in the light than both WT and Dp16 mice. (D) Both Dp16 and Dp16–2N mice exhibited shorter NREM bout lengths than WT mice in the dark and light phases. REM sleep bout lengths were also shorter than WT in both Dp16 and Dp16–2N mice in the dark. Dp16–2N mice had shorter REM bout lengths than Dp16 in the light. Aged N: WT=23; Dp16=27; Dp16–2N=19. W/D = awake dark phase; W/L = awake light phase; NR/D = NREM dark phase; NR/L = NREM light phase; R/D = REM dark phase; R/L = REM light phase. *p<0.05; **p<0.01; *** p<0.001. See Table 1 for a statistics summary.

In aged mice, more bout differences were evident. In the light, genotype exerted an effect (Two-way ANOVA: F_(2,63)=3.217, p=0.047) but not in the dark, and Tukey’s HSD shows that Dp16–2N increased wake bout frequency over WT (p=0.039) but did not differ from Dp16 (Fig. 3C; Table 1). Genotype affected NREM sleep bout frequency in the dark (Two-way ANOVA: F_(2,63)=3.729, p=0.029) and in the light (F_(2,63)=5.223, p=0.008). Dp16 mice showed increased NREM bout frequency compared to WT in the dark (p=0.043) and light (p=0.049), while Dp16–2N NREM bout frequency differed from WT only in the light (p=0.007) (Tukey’s HSD; Fig. 3C; Table 1). REM bout frequency was also affected by genotype in the light (Two-way ANOVA: F_(2,63)=5.079, p=0.009) but not in the dark. Dp16–2N mice exhibited a greater frequency of REM bouts than WT (p=0.007) and Dp16 (p=0.021) (Tukey’s HSD; Fig. 3C; Table 1). Wake-bout duration was similar across aged genotypes (Fig. 3D) while NREM bout duration showed a genotype effect in the dark (Two-way ANOVA: F_(2,63)=11.664, p<0.001) and in the light (F_(2,63)=9.262, p<0.001) and a sex effect in the light (F_(1,63)=4.188, p=0.047). NREM bout duration was shorter for Dp16 in the dark (p<0.001) and light (p<0.001) and shorter for Dp16–2N in the dark (p<0.001) and light (p=0.021) than WT (Fig. 3D). Also, female mice from all genotype groups had longer NREM bouts during the light phase compared to males. Genotype also influenced REM sleep bout duration during dark (Two-way ANOVA: (F_(2,63)=9.347, p<0.001) and light (F_(2,63)=3.121, p=0.050) phases. In the dark, REM bout duration for Dp16–2N was shorter than for WT (p<0.001) and Dp16 (p=0.003), while in the light, it was longer than for Dp16 (p=0.049) (Fig. 3D; Table 1). Together, we observed only minimal REM bout changes in young mice but with aging, shorter, more fragmented NREM in Dp16 mice became evident. Rcan1 normalization selectively strengthens REM continuity.

EEG spectrum power distribution, disturbed in young mice, is partially rescued by Rcan1 correction in aged mice

We examined EEG spectrum power distribution during sleep and wake states in dark and light phases for both age groups. In young mice, genotype affected power in delta (Two-way ANOVA: F_(2,43)=4.940, p=0.012), alpha (F_(2,43)=4.833, p=0.013) and beta (F_(2,43)=4.478, p=0.018) band frequencies while awake during the dark phase. We found that Dp16 mice had less power in delta frequencies than WT (p=0.009) and Dp16–2N (p=0.018). Dp16–2N had less power in alpha band than WT (p=0.019), while both Dp16 (p=0.021) and Dp16–2N (p=0.018) mice exhibited greater power in beta frequencies than WT (Tukey’s HSD; Fig. 4A; Supplemental Fig. 3A; Table 1). For young mice in the light phase, genotype affected power while awake in the alpha (Two-way ANOVA: F_(2,43)=5.057, p=0.011) and beta (F_(2,43)=4.200, p=0.022) frequencies. Dp16 mice had reduced alpha power compared to WT (p=0.004), while WT had less power in beta than Dp16 (p=0.042) and Dp16–2N (p=0.016) (Tukey’s HSD; Fig. 4D; Supplemental Fig. 3D). These results indicate wake EEG shifts toward higher frequency power with reduced low frequency power in young Dp16 mice, and Rcan1 gene dosage correction only partially normalizes this pattern.

Figure 4. Percent total power differences emerged in young mice across EEG frequency bands during the wake state in both dark and light phases. Power differences among genotypes were few during NREM and REM sleep.

(A) Awake in the dark phase, young Dp16 mice exhibited less power in the delta band than WT and Dp16–2N mice while showing more power in the beta and gamma bands. Dp16–2N mice showed significantly less power than WT mice in alpha frequencies only. (B) During NREM, only Dp16–2N mice differed from WT in the gamma band. (C) Dp16–2N mice showed less power in alpha than WT during REM sleep in the dark. (D) In the light phase, Dp16 mice showed less power while awake in the alpha band than WT, and both Dp16 and Dp16–2N mice showed greater power than WT in the beta band. (E) During NREM sleep in the light phase, Dp16–2N mice exhibited more power than WT in the gamma band. There were no other differences among young groups in the light phase. (F) During REM sleep in the light Dp16–2N mice showed greater power than WT in beta frequencies. Sex had a significant effect on power in the beta band with males showing greater power than females. W/D = awake dark phase; W/L = awake light phase; NR/D = NREM dark phase; NR/L = NREM light phase; R/D = REM dark phase; R/L = REM light phase. *p<0.05; **p<0.01; *** p<0.001. Young N: WT=18; Dp16 N=15; Dp16–2N=17. See Table 1 for a statistics summary.

During NREM sleep in young mice, genotype had an effect only in the gamma band during both dark (Two-way ANOVA: F_(2,43)=6.675, p=0.003) and light (F_(2,43)=3.699, p=0.034) phases (Figs. 4B, 4E; Supplemental Figs. 3B, 3E; Table 1), with greater power in Dp16–2N than WT in the dark (p=0.001) and light (p=0.008) (Tukey’s HSD). Young mice were affected by genotypes during REM sleep in the alpha band during the dark phase (Two-way ANOVA: F_(2,43)=5.358, p=0.009) and beta band during the light phase (F_(2,43)=4.555, p=0.017). Dp16–2N mice exhibited less power in alpha than WT in the dark (p=0.001), but more power than WT in beta in the light (p=0.007) than WT (Tukey’s HSD; Figs. 4C, 4F; Supplemental Figs. 3C, 3F; Table 1). The amount of power in the beta frequencies was influenced by sex during both dark (Two-way ANOVA: F_(1,43)=8.092, p=0.007) and light (F_(1,43)=10.568, p=0.002) phases with males, of all genotypes, showing greater power than females in both periods.

During wake periods in the dark, genotype exerted an effect on power in the delta (Two-way ANOVA: F_(2,48)=5.032, p=0.011), beta (F_(2,48)=7.524, p=0.002) and gamma (F_(2,48)=4.002, p=0.025) frequencies in aged mice (Fig. 5A; Supplemental Fig. 4A; Table 1). We found that aged Dp16 mice had less power in delta frequencies than WT (p=0.018) or Dp16–2N (p=0.012). In the beta band, Dp16–2N showed greater power than WT (p<0.001), but in gamma, Dp16 showed greater power than WT (p=0.009), while WT and Dp16–2N were similar (Tukey’s HSD; Fig. 5A; Supplemental Fig. 4A; Table 1). Sex also exerted an effect on gamma band power (Two-way ANOVA: F_(1,48)=6.422, p=0.015), with females, regardless of genotype, displaying higher gamma power than males (Fig. 5A; Supplemental Fig. 4A; Table 1). In the light, genotype exerted an effect on power in the beta band (Two-way ANOVA: F_(2,48)=8.021, p<0.001). Aged Dp16 mice (p=0.003) and Dp16–2N (p=0.005) both exhibited greater power than WT (Tukey’s HSD; Fig. 5D; Supplemental Fig. 4D; Table 1).

Figure 5. Rcan1 dosage correction partially rescues EEG power differences exhibited in aged Dp16 mice during the dark and light phases.

(A) During wake states in the dark phase, both WT and Dp16–2N mice showed more power than Dp16 mice in the delta band. Dp16–2N mice exhibited greater power than WT in the beta band, and Dp16 showed greater power than WT in gamma frequencies. (B) During NREM sleep in the dark phase, WT mice showed greater power than aged Dp16 and Dp16–2N mice in the delta frequencies while Dp16 and Dp16–2N showed greater power than WT in theta frequencies. Both WT and Dp16 mice exhibited greater power in the gamma band. (C) Dp16 mice showed greater power than both WT and Dp16–2N mice in theta frequencies during REM sleep in the dark, but no other differences were observed. (D) While awake during the light phase, the only differences observed were in the beta frequencies. Both Dp16 and Dp16–2N mice exhibited greater power than WT. (E) During NREM sleep in the light phase, Dp16 and Dp16–2N mice exhibited less power in the delta frequencies and more power in the theta and gamma frequencies than WT mice. In the alpha band, Dp16 mice showed more power than WT and there was a genotype by sex interaction. Dp16–2N females exhibited significantly more power than WT males and Dp16–2N males. Dp16–2N mice also demonstrated more power in the gamma band than Dp16 mice. (F) During REM sleep in the light phase, the only difference in power distribution across frequency bands was in the alpha band. Dp16–2N mice showed less power in alpha than WT mice. There was a genotype by sex interaction in the delta band: Dp16–2N females had lower power than Dp16–2N males. W/D = awake dark phase; W/L = awake light phase; NR/D = NREM dark phase; NR/L = NREM light phase; R/D = REM dark phase; R/L = REM light phase. *p<0.05; **p<0.01; *** p<0.001. Aged N: WT=23; Dp16=27; Dp16–2N=19. See Table 1 for a statistics summary.

Aged mice also showed an effect of genotype on power in the delta (Two-way ANOVA: F_(2,48)=8.396, p<0.001), theta (F_(2,48)=5.303, p=0.009) and gamma (F_(2,48)=10.379, p<0.001) frequency bands during NREM in the dark. Percent total power in delta frequencies decreased in Dp16 (p=0.004) and Dp16–2N (p=0.002) compared to WT. Percent total power increased in the theta frequencies for Dp16 (p=0.019) and Dp16–2N (p=0.012). Dp16–2N gamma power was less than both WT (p<0.001) and Dp16 (p=0.042) (Tukey’s HSD; Fig. 5B; Supplemental Fig. 4B).

In the light during NREM sleep, genotype exerted an effect on percent total power in the delta (Two-way ANOVA: F_(2,48)=13.912, p<0.001), theta (F_(2,48)=5.799, p=0.006), alpha (F_(2,48)=4.921, p=0.012) and gamma (F_(2,48)=29.845, p<0.001) frequency bands. Percent total power was lower in Dp16 (p<0.001) and Dp16–2N (p<0.001) than WT mice in the delta frequencies while theta was higher in Dp16 (p=0.009) and Dp16–2N (p=0.009) mice. Dp16 mice also showed more power than WT in alpha (p=0.019) and gamma (p<0.001) bands and gamma power was higher in Dp16–2N than WT (p<0.001) mice. While we did not observe a main effect of sex on percent total power in the alpha frequencies, genotype interacted with sex to have a significant effect in this band (Two-way ANOVA: F_(2,48)=4.455, p=0.018). This interaction showed greater power in Dp16–2N females than WT (p=0.019) and Dp16–2N (p=0.050) males, but not WT females (Table 1). There was a significant effect of sex on percent total power in the gamma band (Two-way ANOVA: F_(1,48)=4.346, p=0.043) with female mice exhibiting greater power during NREM in the gamma band than males. Genotype affected percent total power in the theta band (Two-way ANOVA: F_(2,48)=5.255, p=0.009) during REM sleep in the dark. Dp16 showed greater theta power than WT (p=0.006) and Dp16–2N (p=0.024) aged mice (Tukey’s HSD; Fig. 5C; Supplemental Fig. 4C; Table 1). During REM sleep in the light, genotype exerted an effect in the alpha frequencies (Two-way ANOVA: F_(2,48)=3.596, p=0.036). Aged Dp16–2N mice had less power (p=0.016) in alpha than WT, and a genotype by sex interaction in delta frequencies (Two-way ANOVA: F_(2,48)=3.251, p=0.048) revealed greater power in Dp16–2N males vs. females (p=0.048) (Tukey’s HSD; Fig. 5F; Supplemental Fig. 4F).

The distribution of control of wake/sleep states across multiple brain circuits is clearly evident with the diverse effects of restoring Rcan1 gene dose seen in these results. Overall, in aged mice, Rcan1 dose correction reverses Dp16 delta deficits, while beta activity remains elevated. Because reduced delta power reflects decreased neuronal synchrony and diminished NREM intensity, this restoration suggests a partial recovery of deep sleep quality. Concurrently, Rcan1 correction lowers elevated theta, consistent with increased processing efficiency, whereas decreased alpha power in the light phase may indicate less awake quiet rest.

Gene expression differences across sex and genotype

RNA-seq revealed broad up-regulation of MMU16 genes in aged Dp16 and Dp16–2N hippocampus, consistent with trisomy (Waugh et al., 2023) (Fig. 6A,D). Several genes were found to be differentially expressed in Dp16 vs. WT mice throughout the genome (108 in females, 90 in males) (Supplemental Table 1). Expression of some genes was partially corrected in the Dp16–2N group compared with Dp16 (14 in females, 11 in males), including genes involved in inflammatory response, protein modification, and tight junctions (Supplemental Table 2). Gene Set Enrichment Analysis (GSEA) revealed pathways differentially enriched between groups (Fig. 6B,C), including suppression of GABAergic synapse-related sets in Dp16 and Dp16–2N males. Consistent with this, we found fewer parvalbumin-positive GABAergic cells in the aged hippocampus of Dp16 males (p=0.018) and a similar trend in Dp16–2N males (p=0.080) compared to WT males (Supplemental Figure 5). Additional GSEA signals included cytokine signaling, a hallmark of DS, and modulation of chemical synaptic transmission, a potential consequence of reduced GABAergic cell numbers. Fos showed sex-dependent modulation, with slight variation among females and larger differences among males (Fig. 6E). Normalization of Rcan1 dosage in Dp16–2N mice partially restored expression of a subset of genes with known roles in sleep/circadian regulation (e.g., Trpa1, Ghsr, Syt6, Slc14a2) and aging- or neurodegeneration-relevant pathways (e.g., Nes, Neurog2, Adamts1, Tnfrsf12a, Apold1, Eltd1, Selenov) (Babcock et al., 2021; Boldrini et al., 2018; Cole et al., 2022; Hara et al., 2017; Hsu et al., 2017; Kluge et al., 2010; Lagace et al., 2007; Masiero et al., 2013; Miguel et al., 2005; Nagy et al., 2021; Qiu et al., 2022; Roessingh and Stanewsky, 2017; Stritt et al., 2023; Yang et al., 2015; Zocher and Toda, 2023). Together, these results show transcriptomic shifts converging on inhibitory synapses and neuroimmune signaling, with male-biased effects paralleling EEG and PV+ patterns. Notably, several genes corrected by Rcan1 dose correction have roles in sleep and age-related processes.

Figure 6. Aged Dp16 and Dp16–2N male and female mice show transcriptionally distinct gene cluster profiles compared to WT.

(A) The percentages of differentially expressed genes (DEGs) belonging to each chromosome. (B) Clusters of DEGs enriched in different pathways based on Gene Set Enrichment Analysis (GSEA), separated by positive (activated) or negative (suppressed) enrichment scores. X-axis shows the comparisons by genotype and sex; the Y-axis lists the enriched pathways associated with the genes in those clusters. Color scale describes the adjusted P-value (red: lower, blue: higher), and dot size indicates the proportion of genes in the functional category that are annotated in the cluster. (C) GSEA plots for enriched pathways of interest in Male WT vs. Dp16–2N and Dp16 vs. Dp16–2N comparisons. The running enrichment score is shown by the green line while the vertical black lines represent where the members of the gene set appear in the ranked list of genes. (D) Volcano plots for female and male WT vs. Dp16 and WT vs. Dp16–2N groups. X-axis shows log2 fold change, Y-axis −log10 (adjusted-p). Significantly differentially expressed genes are shown in red for upregulated and green for downregulated (adjusted-p<0.1). (E). Normalized counts for Trpa1 and Fos, across genotype and sex. X-axis represents the different biological groups assessed in the RNAseq study. Y-axis displays normalized counts. * padj.< 0.1; ** padj.<0.05; *** padj.<0.01. See Table 1 for a statistics summary.

Reducing Rcan1 dosage ameliorates anxiety-like behavior in aged Dp16 mice

Because of the relationship between disrupted sleep and anxiety (Chawla et al., 2020; Fuca et al., 2023; Fuca et al., 2022; Ong et al., 2024), we assessed anxiety-like behavior in aged experimental groups using OFA and EPM assays. In OFA, genotype influenced time spent in the center versus the periphery (One-way ANOVA: F_(2,36)=3.699, p=0.035). We found that in males, Dp16 spent less time in the center relative to WT (p=0.041), with a trend relative to Dp16–2N (p=0.085). The mean total distance moved did not differ by genotype except for a trend in females across genotypes (One-way ANOVA: F_(2,30)=2.993, p=0.066) toward increased travel distance (Fig. 7A,B). In EPM, genotype exerted an effect on the amount of time spent in the open arm by male (One-way ANOVA: F_(1,29)=4.388, p=0.022) and female (F_(1,30)=5.464, p=0.010) mice. We found reduced open-arm time in male Dp16 compared to both WT (p=0.034) and Dp16–2N (p=0.044) males and in female Dp16 vs. Dp16–2N females (p=0.015). Dp16–2N females also showed increased open-arm time compared with WT females (p=0.024) (Fig. 7C).

Figure 7. Open Field Arena testing shows a small but significant increase in anxiety-like behavior in Dp16 mice that is partially rescued by Rcan1 dose correction in males. Elevated Plus Maze testing shows increases in anxiety-like behavior in Dp16 mice that is rescued by Rcan1-dose correction accompanied by an increase in distance moved in females.

(A) Male Dp16 mice show a significant decrease in center time during OFA testing. Rcan1 dose correction rescued this phenotype. (B) No differences in distance moved were seen in males or females, but females trended towards a significant genotype effect (p=0.066) and a trend towards greater distance moved in female Dp16 compared to female WT (p=0.057). (C) EPM testing revealed significant genotype effects in closed arm time for males (p=0.046) and females (p=0.002), open arm time for males (p=0.022) and females (p=0.010), center time for males (p=0.032), and a trend in center time for females (p=0.083). Male Dp16–2N mice had significantly less closed arm time than male Dp16 mice (p=0.050), and female Dp16–2N mice had significantly less closed arm time than both WT females (p=0.002) and Dp16 females (p=0.011). Male Dp16 mice had significantly lower open arm time than both WT males (p=0.034) and Dp16–2N males (p=0.044). In females, Dp16–2N mice had significantly higher open arm time than both female WTs (p=0.024) and female Dp16s (p=0.015). Male Dp16–2N mice had significantly higher center time compared to male WTs (p=0.047). (D) Females showed a significant genotype effect on distance moved (p<0.001). Female Dp16–2N mice moved significantly more compared to both WT (p<0.001) and Dp16 (p=0.006) females. See Table 1 for a statistics summary.

The time spent in the closed arm was also affected by genotype in both males (One-way ANOVA: F_(1,29)=3.471, p=0.046) and females (F_(1,29)=8.259, p=0.002). Dp16–2N males showed reduced closed-arm time compared to Dp16 males (p=0.050) and Dp16–2N females also exhibited reduced closed-arm time compared to Dp16 (p=0.011) and WT mice (p=0.002). The time spent in the center zone was also affected by genotype (One-way ANOVA: F_(1,33)=3.908, p=0.032). Center-zone time increased in Dp16–2N males vs. WT males (p=0.047) (Fig. 7C). Total distance moved did not differ among genotypes for males but did in females (One-way ANOVA: F_(1,33)=3.908, p<0.001). The total distance traveled increased in female Dp16–2N compared with both WT (p<0.001) and Dp16 (p=0.006) females (Fig. 7D). Taken together, these results show increased anxiety-like behavior in aged Dp16 males and some mitigation following Rcan1 correction, which aligns with parallel improvements in sleep architecture and EEG signatures.

DISCUSSION

We previously characterized the sleep architecture and EEG activity in aged Dp16 mice (Levenga et al., 2018). Here, we extend those findings to young Dp16 mice (3–6 months) and test whether restoring RCAN1 protein to disomic levels rescues disrupted sleep and EEG phenotypes in aged animals. By generating Dp16–2N mice, which retain the Dp16 trisomic interval but carry only two copies of Rcan1 (Vega et al., 2003; Yu et al., 2010), we show that RCAN1 dosage contributes substantially to the age-emergent sleep and electrophysiologic abnormalities observed in this model. These data strengthen the case that RCAN1 is one mechanistic contributor to DS-related sleep dysregulation, while also indicating that additional trisomic genes likely shape the full phenotype.

Sleep architecture

Individuals with DS have difficulty initiating and maintaining sleep (Andreou et al., 2002; Fan et al., 2017; Fernandez and Edgin, 2013; Nisbet et al., 2015). Consistent with our prior work (Levenga et al., 2018), aged Dp16 mice spent more time awake and less time in NREM than WT littermates, whereas young Dp16 mice showed only modest abnormalities. This pattern suggests that sleep disruption in Dp16 emerges more clearly with age than in early adulthood. Importantly, reducing RCAN1 dosage in aged Dp16–2N mice largely normalized wake and NREM measures, indicating that excess RCAN1 contributes to impaired sleep maintenance. A plausible mechanism is dysregulation of calcineurin-dependent signaling, as RCAN1 can either inhibit or facilitate calcineurin activity (Fuentes et al., 2000; Fuentes et al., 1995; Hoeffer et al., 2007; Hoeffer et al., 2013), and calcineurin-NFAT pathways influence circadian and sleep-related outputs (Aramburu et al., 2000; Dyar et al., 2015; Huang et al., 2012; Katz et al., 2008; Kweon et al., 2018; Lee et al., 2019; Nakai et al., 2011; Sachan et al., 2011). We therefore interpret the rescue in Dp16–2N mice as evidence that RCAN1 overexpression is one driver of sleep fragmentation in this model, rather than the sole determinant of all DS-related sleep phenotypes.

Age-dependent sleep abnormalities

Age-related sleep changes are well-described in humans and mice (Roffwarg et al., 1966; Soltani et al., 2019) and sleep disturbances emerge early in DS and worse across life (Bassell et al., 2015; Gimenez et al., 2021; Santos et al., 2022). Our “young” cohort reflects mature adulthood rather than juvenile development, which likely explains the modest EEG changes at this age. In that context, the stronger phenotype in aged Dp16 mice and the greater efficacy of Rcan1 dose correction at older ages are consistent with the idea that RCAN1 contributes more strongly to DS-related sleep dysfunction as aging-related stressors accumulate (Horvath et al., 2015). This interpretation also fits prior work showing altered circadian activity in Dp16 mice (Wong et al., 2022) and increased RCAN1 expression in aging and AD-related contexts (Wong et al., 2015).

These findings should also be viewed in the broader context of DS mouse models. Ts65Dn remains the most extensively characterized model and exhibits more robust sleep and EEG abnormalities, including increased wakefulness, reduced NREM sleep, and elevated sleep theta power, whereas Ts1Cje shows substantially milder sleep/EEG disruption (Bolla et al., 2025; Colas et al., 2008; Pittaras et al., 2025). Tc1 and TcMAC21 mice display increased sleep fragmentation (Banks et al., 2015; Tusk et al., 2025). Dp16 offers stronger construct validity than Ts65Dn because it confers trisomic dosage across the Mmu16 region syntenic to Hsa21 without the non-Hsa21 Mmu17 segment carried by Ts65Dn (Guedj et al., 2023; Gupta et al., 2016), but that cleaner genetic architecture may also yield a narrower or delayed phenotype spectrum in some domains. Thus, our inability to detect major developmental-stage sleep abnormalities in Dp16 does not diminish the value of the model; rather, it suggests that Dp16, Ts65Dn, Ts1Cje, and newer models such as Ts66Yah (Duchon et al., 2022) and TcMAC21 (Kazuki et al., 2020) each capture overlapping but distinct aspects of DS biology.

EEG comparisons

EEG frequency bands provide a complementary readout of state regulation and network function. In young Dp16 mice, the EEG phenotype was relatively subtle, suggesting that electrophysiologic abnormalities may precede or accompany later architectural disruption without yet producing large changes in state distribution. In aged Dp16 mice, however, the abnormalities were more pronounced, particularly during NREM, where reduced delta and increased theta, alpha, and gamma power point to altered network synchrony and diminished sleep quality (Achermann and Borbely, 2003; Tobler and Borbely, 1986).

Rcan1 dose correction improved several of these abnormalities, especially elevated theta and some alpha-band differences (Brown et al., 2012; Buzsáki, 2006), but did not normalize all measures. The incomplete rescue is informative. Reduced delta power, a surrogate of NREM intensity, persisted in some comparisons even when sleep architecture improved, suggesting that RCAN1 is a substantial contributor to the phenotype but not the only trisomic dosage driver. This is in line with the broader genetic logic of DS models, in which no single gene is expected to account for all circuit-level abnormalities.

The GABAergic findings also fit this interpretation. Our transcriptomic data implicated inhibitory synapse-related pathways (Fig. 6), and PV-cell analyses suggested modest sex- and genotype-dependent effects. Together, these results point more toward altered inhibitory circuit function than a simple loss of interneuron number (Supplementary Fig. 5, Petsche and Stumpf, 1962; Yoder and Pang, 2005). RCAN1-dependent calcium/calcineurin signaling could contribute to that dysfunction, but the persistence of some gamma-band abnormalities after Rcan1 correction argues that other triplicated genes, potentially including App and other Mmu16 loci, also shape the electrophysiologic phenotype (Hijazi et al., 2023).

Broader implications

Sleep regulation is polygenic, so it is unlikely that any single trisomic gene explains DS-associated sleep dysfunction in full (Dashti et al., 2019; Goodman et al., 2025; Jansen et al., 2019; Jones et al., 2019). We therefore view RCAN1 not as a solitary “sleep gene,” but as one mechanistically important node within a broader trisomy-sensitive network that links calcium signaling, neural state regulation, and behavioral function. Tailoring strategies to normalize neural RCAN1–CaN signaling may improve sleep continuity and cognition in DS.

Limitations and future directions

This study has several limitations. Dp16 does not recapitulate all features of human DS, including some anatomic contributors to OSA (Takahashi et al., 2020), and it also does not reproduce the full phenotype spectrum reported in other DS mouse models. That limitation is also a strength: because Dp16 avoids the extra non-Hsa21 segment present in Ts65Dn, it offers a cleaner system for testing the contribution of specific Hsa21-syntenic genes such as Rcan1. Our data therefore supports a model in which Dp16 is especially useful for studying adult and aging-related sleep and EEG abnormalities, while other models may better capture selected developmental phenotypes. Future work should compare sleep and EEG longitudinally across Dp16, Ts65Dn, Ts66Yah, and TcMAC21, and should incorporate cell-type-specific and region-specific manipulations to identify where RCAN1 dosage most strongly affects sleep-regulatory circuits. Overall, our findings support the conclusion that RCAN1 contributes importantly to sleep and EEG dysfunction in Dp16 mice, while also underscoring that DS model systems are complementary rather than interchangeable.

Highlights.

• Young Dp16 display sleep deficits; aged Dp16 show more severe sleep disruption

• RCAN1 dose correction rescues most sleep dysfunction in aged Dp16 animals

• RNA-Seq implicates GABAergic systems, PV+ cell counts show sex and genotype effects

• RCAN1 gene-dose correction rescues anxiety-like behavior in aged Dp16 mice

Data Availability Statement

The datasets generated for this study are available from the corresponding author upon request.