Nacc1 Mutation in Mice Models Rare Neurodevelopmental Disorder with Underlying Synaptic Dysfunction

Generation of Nacc1^+/R284W mice

The Nacc1 genes in human and mouse are ∼87% homologous and the site encoding the arginine to tryptophan (R>W) in human is conserved in mouse. The human gene encodes one transcript, and the mouse has two, designated as long and short which were identified first in rodent brain (Cha et al., 1997; Korutla et al., 2002; Mackler et al., 2003). Only the long isoform in mouse contains sequence corresponding to R>W in human (Fig. 1). Utilizing CRISPR-Cas9, we generated the first knock-in mouse model of the disorder through introducing the homologous nucleotide substitutions in the Nacc1 gene (human: rs1060505041, cDNA 892C>T, p.298R>W; mouse: cDNA 850C>T and 852T>G, p284R>W; Fig. 1). In brief, a DNA oligonucleotide donor encoding the base pair mutation analogous to the human gene along with a single guide RNA and SpCas9 protein were injected into C57BL/6 single-cell mouse embryos. One founder (F17) was selected for breeding and backcrossing in the C57BL/6 strain. Germline transmission through F9 generation was confirmed by sequencing, and most data reported here were collected in F3–F6 progeny. Mice were genotyped by DNA sequencing of PCR products from tail or ear DNA. In some experiments, post necropsy tissues were used for genotyping. Guide RNA, the donor DNA sequence, and primer sequences are reported in Table 1.

Postnatal litter size, survival at different ages, and weight gain

The size of litters from Nacc1^+/+ and Nacc1^+/R284W females mated to Nacc1^+/R284W males were not different suggesting the mutation did not affect fertility (Fig. 2a). To determine if genotype distribution differed from the expected 1:2:1 Mendelian outcome, litters from Nacc1^+/R284W × Nacc1^+/R284W mating pairs were scored for genotype at different ages [embryonic day (E) 17.5, P7, and P21]. At E17.5 and P7, there was no change in the distribution of Nacc1^+/+, Nacc1^+/R284W, and Nacc1^R284W/R284W genotypes (Table 3). Mice genotyped at P21 (their weaning period) showed reduced percentages of the Nacc1^R284W/R284W genotype for both males and females (Table 3). These data suggested that between P7 and P21, mice with two mutant alleles may undergo premature death. Longer survival periods were not examined systematically, however, two other groups from F5 generation (11 heterozygotes) and F6 generation (4 homozygotes) were aged past 8 months with no premature deaths, and 24 Nacc1^+/R284W mice from the F1 generation that were aged to 13 months and 20 from F4 and F5 aged to between 11 and 17 months at UMass Chan had no premature deaths. Overall, these results indicate that Nacc1^+/+ and Nacc1^+/R284W mice have normal fertility and survival but Nacc1^R284W/R284W mice could be at increased risk of neonatal death closer to weaning.

Weights were measured across several different cohorts of mice between E17.5 to 6 months in age at the two test sites (MGH and UMass Chan; Fig. 2b–h). There were no changes in weight at E17.5 between Nacc1^+/+, Nacc1^+/R284W, and Nacc1^R284W/R284W mice (Fig. 2b, UMass Chan site) and at P1 between Nacc1^+/+ and Nacc1^+/R284W mice (Fig. 2c, MGH site). P7 male Nacc1^R284W/R284W showed reduced weight compared with Nacc1^+/R284W males while Nacc1^+/R284W and Nacc1^R284W/R284W females showed reduced weight compared with Nacc1^+/+ females (Fig. 2d, UMass Chan site). At P14, Nacc1^R284W/R284W males also showed a reduced weight compared with Nacc1^+/R284W males and while the trend remained for female Nacc1^R284W/R284W, there was no change in female weight observed (Fig. 2e, MGH site). At P21, Nacc1^R284W/R284W males and females showed reduced weight compared with Nacc1^+/R284W and Nacc1^+/+ mice (Fig. 2f, UMass Chan site). Both P120 and P180 Nacc1^+/R284W males showed reduced weight compared with Nacc1^+/+ males (Fig. 2g, MGH, and 2h, UMass Chan, respectively) while only P180 Nacc1^+/R284W females showed a reduced weight compared with Nacc1^+/+ females (Fig. 2h, UMass Chan). Altogether this analysis suggested that mice with the Nacc1 R284W mutation from P7 to 6 months had delayed weight gain with Nacc1^R284W/R284W displaying more consistent reduced weight gain for both males and females while Nacc1^+/R284W mice showed sex-dependent changes as some ages. The delayed weight gain for Nacc1^R284W/R284W mimics an early symptom of delayed weight gain seen in patients that contributes to a failure to thrive (Schoch et al., 2017).

Absence of cataracts in Nacc1 mutant mice

Bilateral cataracts identified between birth and 10 years of age were found in five of seven of the individuals first identified with the NACC1 R298W mutation (Schoch et al., 2017). No cataracts were evident in nine Nacc1^+/R284W and four Nacc1^R284W/R284W mice aged 1.5–5months using the method of observation described by He et al. (2019). This indicates that formation of cataracts, at least at early ages, is not a phenotype in mice with the Nacc1 R284W mutation. Cataracts were not reported in two mouse models with deletion of the Nacc1 gene (Yap et al., 2013; da Silva-Buttkus et al., 2023).

Nacc1^+/R284W mice have behavioral and motor deficits

A timeline shows the order of a panel of tests performed in the same group of mice from 2.5 to 4 months of age (Fig. 3a). Thigmotaxis in normal mice is the tendency to stay along the periphery of an open field (Simon et al., 1994). Nacc1^+/R284W males were the same as Nacc1^+/+ males in time spent in the center and periphery of the field, whereas Nacc1^+/R284W females spent less time in the center and more time in the border than Nacc1^+/+ littermates (Fig. 3b–d), a possible sign of greater anxiety (Seibenhener and Wooten, 2015). However, Nacc1^+/R284W mice were not different from Nacc1^+/+ in total distance traveled (Fig. 3e).

Mice were evaluated at 3 months for hindlimb clasping and paw curling (Fig. 3f–j). Male and female Nacc1^+/R284W mice had significantly higher clasping scores and mean percentage of clasping than Nacc1^+/+ mice (Fig. 3h,i, respectively). To explore the onset age of this phenotype, we tested additional groups of younger (P14, F4–F5) and older (6-month-old, F2) mice from different cohorts. Male and female Nacc1^+/R284W mice examined at P14 displayed no changes from Nacc1^+/+ in percent clasping (Fig. 3g), whereas Nacc1^+/R284W females tested at age 6 months had a higher percent of clasping compared with age-matched Nacc1^+/+ females (Fig. 3j). Finally, homozygote mice bred beyond F6 generation exhibited hindlimb clasping. This included five of six F8 homozygotes at 3.2 months and 19/20 F10 homozygotes at 2.8–3 months old. Collectively, these data suggest that hindlimb clasping is a robust indicator of abnormality that develops after age P14. This phenotype has been reported in mouse models of other neurodevelopmental (Rett syndrome, Angelman's syndrome) and neurodegenerative disorders (Huntington's disease) and its presence signals dysfunction in multiple regions including the cortex, cerebellum, and caudate–putamen and the involvement of cerebello-cortical-reticular and cortico-striato-pallido-reticular pathways (Lalonde and Strazielle, 2011; Mandel-Brehm et al., 2015; Cruz et al., 2021).

Mice in the behavior cohort were next tested at 4 months for their ability to build nests. Scoring included both the percentage of nestlet used and a composite score for the nest built (Fig. 3k–m). The total amount of shredded nestlet used for building a nest was not different between Nacc1^+/R284W and Nacc1^+/+ male and female mice (Fig. 3l). Both sexes of Nacc1^+/R284W mice were able to build nests but Nacc1^+/R284W females scored poorly compared with Nacc1^+/+ females (Fig. 3k,m), which is based on how much of the nestlet provided to them is used for building a nest and the quality of the built nest. These results indicated that the innate ability to build nests was impaired in females with the mutation.

Gait assessed at 4 months revealed a significantly shorter hindlimb stride length in female Nacc1^+/R284W mice compared with that in Nacc1^+/+ mice (Fig. 3n,p). No difference between Nacc1^+/+ and Nacc1^+/R284W mice was uncovered in forelimb stride length (Fig. 3o) or in stride width and diagonal stride width (Fig. 3q–t).

Altogether, results from the behavioral analysis showed that Nacc1^+/R284W mice exhibited motor and behavioral impairments between 3 and 6 months of age. Females had more functional deficits than males in an open field measure of anxiety, in the incidence of clasping, in building nests, and in hindlimb stride length.

Performance on three other tests did not differ between genotypes (Fig. 4). The 2 d novel object recognition test, where two identical objects were presented on day 1 and then one was replaced by a novel object on day 2 (Fig. 4a–d), tests long-term memory recall in mice (≥24 h between tests 1 and 2; Taglialatela et al., 2009). Nacc1^+/+ and Nacc1^+/R284W mice showed no preference for object placement during the training phase on day 1 (Fig. 4b,c). Nacc1^+/+ and Nacc1^+/R284W mice also showed no change in their preference for the novel object on day 2 (Fig. 4d). The Y-maze assesses short-term spatial working memory (Kraeuter et al., 2019) through measuring correct and incorrect alternations (Fig. 4e–g). Performance in the Y-maze assessed as mean percent alteration between the three arms of the maze did not differ between genotypes (Fig. 4f) nor did the total number of arm entries (Fig. 4g). Finally, in tests of grip strength (Fig. 4h–j), female and male Nacc1^+/R284W mice displayed no changes in normalized average (Fig. 4i) or normalized maximum grip strength (Fig. 4j) compared with their respective Nacc1^+/+ mice. These results suggest that memory impairment and grip strength in Nacc1^+/R284W mice are not detected through these testing metrics.

Seizure-like behaviors appear in Nacc1^+/R284W and Nacc1^R284W/R284W mice at 7 months of age

Patients with NACC1 892C>T (pR298W) mutation have seizures and manifest abnormal EEG activity associated with myoclonic movements [Schoch et al. (2017), their supplementary patient report data]. Behavioral seizures, similar to those identified in the Racine seizure scale [Van Erum et al. (2019), their Fig. 1], were seen in some Nacc1^+/R284W mice starting from age 7 months in the home cage either spontaneously or in response to handling, a sudden noise or light stimulus. The seizure-like behaviors included momentary or sustained freezing (Fig. 4k,m; Movies 1, 3, respectively), whisker twitching and head bobbing (Fig. 4k,m; Movies 1, 3, respectively), and wild jumping (Fig. 4l, Movie 2) and were seen in Nacc1^+/R284W and Nacc1^R284W/R284W mice and not in Nacc1^+/+ mice of similar ages. Seizures were seen in the home cages in later generations of mice. A homozygote mouse at F9 was noted to have a mild seizure in the cage at 4.5 months and a severe seizure at 5.5 months at which time the mouse was killed.

Nacc1^+/R284W mice display cortical EEG epileptiform discharges during wakefulness and REM sleep and changes in spectral power distribution

We recorded cortical EEG and EMG activity in Nacc1^+/+ and Nacc1^+/R284W mice at 3.5–4 months of age (Fig. 5a). Nacc1^+/+ littermates (n = 11, 6 males and 5 females) showed normal EEG activity (Fig. 5b,g). However, male and female Nacc1^+/R284W mice (n = 9, 5 males and 4 females) displayed single or repetitive high-amplitude epileptiform discharges in the EEG during wakefulness and REM sleep (Fig. 5c,h). During wakefulness, epileptiform discharges coincided with low EMG activity and behavioral arrest (Fig. 5, Table 4, Movie 4). Characteristics of epileptiform discharges that occurred during wakefulness and REM sleep including mean number and duration of bursts measured in 48-, 24- and 1 h periods did not differ significantly between females and males (Table 4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Epileptiform discharges in Nacc1^+/R284W mice. a, Still image (bottom right) from a video showing mouse in chamber tethered to a recording cable and traces of epileptiform discharges that coincided with low EMG activity and behavioral arrest. See Movie 4. b, g, Examples of normal EEG and EMG during wakefulness (b) and REM sleep (g) in the same Nacc1^+/+ mouse. Sample traces show epileptiform discharges during wakefulness (c) and during REM sleep (h) in the same Nacc1^+/R284W mouse. Note the absence of muscle movement recorded by EMG that coincides with the EEG epileptiform discharges during wakefulness. d–f, i–k, Plots show the cortical EEG power spectral distribution expressed as a percent of total power for each frequency (0.5–60 Hz) during wakefulness (d–f) and during REM sleep (i–k) and for females (d,i), males (e,j), and combined females and males (f,k) respectively. The inserted bar graphs show cortical EEG power bands (δ, 0.5–4.5 Hz; θ, 4.5–10 Hz; σ, 10–15 Hz; β, 15–30 Hz; γ, 30–60 Hz). Episodes containing epileptiform discharges were analyzed over a 24 h period and compared with wakefulness cortical EEG power distribution during a circadian time of high waking drive (19:00–22:00; d–f) or with REM sleep cortical EEG power distribution during a circadian time of high sleeping drive (10:00–13:00; i–k), respectively. d–f, i–k, In female and male Nacc1^+/R284W mice, the cortical EEG power distribution of the epileptiform discharges was significantly different from normal wakefulness or REM sleep. A two-way ANOVA followed by a post hoc Bonferroni’s test, n = 4 male Nacc1^+/R284W mice (red), n = 5 female Nacc1^+/R284W mice (blue), n = 9 female + male Nacc1^+/R284W mice (purple). For all panels, asterisks indicate p values with the following cutoffs: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

The spectral power distribution differed between periods of epileptiform discharges and those of normal wakefulness (Fig. 5d–f) and REM sleep (Fig. 5i–k) with percentage of total power significantly reduced in the δ band (1–3 Hz) and increased in the β band (13–30 Hz) in males and females (Fig. 5f,k). Other significant changes in power spectra of the bursting periods occurred in males during wakefulness (γ, 55–60 Hz) and REM sleep (θ, 4.5–10 Hz, and σ, 1–3 Hz; Fig. 5e,j). Power spectra analysis of epileptiform discharges during wakefulness also showed rhythmic activity in the θ frequency range (4.5–10 Hz) in both genders and was associated with harmonics (at 16 and 24 Hz), characteristic of EEG hypersynchrony, a feature also described in middle-aged APP/PS1 mice (Jin et al., 2018). These data indicate that Nacc1^+/R284W mice have frequent epileptiform discharges during wakefulness and REM sleep which are accompanied by changes in the spectral power distribution.

Analysis of the 24 h circadian period revealed that Nacc1^+/R284W mice displayed an hourly percentage of wakefulness, SWS, and REM sleep similar to Nacc1^+/+ mice except female Nacc1^+/R284W mice who spent a greater percent of time in SWS (Fig. 6a–c). With some exceptions described below, the spectral power distributions of the EEG during wake–sleep cycles were similar in female and male Nacc1^+/R284W mice and Nacc1^+/+ mice (Fig. 6). During REM sleep, female Nacc1^+/R284W mice (n = 4) had a deficit in the percentage of total power associated with θ band as compared with female Nacc1^+/+ mice (n = 5; Fig. 6i). This occurred during the period when mice are highly awake (19:00–22:00), indicating a deficit in hippocampal synchronization (Buzsaki, 2002). During SWS, male Nacc1^+/R284W mice (n = 5) had significantly higher δ power as compared with male Nacc1^+/+ mice. This happened during the light period (10:00–13:00), when mice are mostly asleep, indicating a deeper SWS and more sleep pressure (n = 5; Fig. 6n). Altogether, these results indicated that Nacc1^+/R284W mice had some changes in their sleep–wake cycles and associated spectral power distribution that differed in females and males.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Sleep–wake phenotypes in female and male Nacc1^+/R284W mice. a–c, Wakefulness (a), SWS (b), and REM sleep (c) hourly percentage in female mice. d–i, Cortical EEG power spectral distribution and bands (inserted histogram), expressed in percentage of total (0.5–60 Hz) power of wakefulness (d,g), SWS (e,h), and REM sleep (f,i), during a circadian time of high sleeping drive (10:00–13:00; d–f) and during a circadian time of high waking drive (19:00–22:00; g–i). δ, 0.5–4.5 Hz; θ, 4.5–10 Hz; σ, 10–15 Hz; β, 15–30 Hz; γ, 30–60 Hz. In i, θ power is reduced in female Nacc1^+/R284W mice (red, n = 4) compared with that in female littermate control (Nacc1^+/+, black, n = 5) mice. *p < 0.05, two-way ANOVA followed by a post hoc Bonferroni’s test. j–l, Wakefulness (j), SWS (k), and REM sleep (l) hourly percentage in male mice. m–r, Cortical EEG power spectral distribution and bands (inserted histogram), expressed in percentage of total (0.5–60 Hz) power of wakefulness (m,p), SWS (n,q), and REM sleep (o,r) during a circadian time of high sleeping drive (10:00–13:00; m–o) and during a circadian time of high waking drive (19:00–22:00; p–r). δ, 0.5–4.5 Hz; θ, 4.5–10 Hz; σ, 10–15 Hz; β, 15–30 Hz; γ, 30–60 Hz. Male Nacc1^+/R284W mice (blue, n = 5) and male littermate wild-type (Nacc1^+/+, black, n = 6) mice. In n, ,δ power is increased in male Nacc1^+/R284W mice *p < 0.05, two-way ANOVA followed by a post hoc Bonferroni’s test.

δ waves contribute to the restorative aspects of SWS and to consolidate memories from synaptic maturation (Malkani and Zee, 2022). Enhanced (or increased) δ power occurs in human neurodevelopmental disorders including Rett syndrome (Ammanuel et al., 2015; Roche et al., 2019) and in a mouse model of Angelman's syndrome using the 129 mouse strain differing from the C57BL/6 strain (Sidorov et al., 2017) used in our study. δ power buildup was speculated to result from delayed synaptic maturation (Ammanuel et al., 2015) and may signal an imbalance in cortical inhibitory and excitatory activity. This imbalance could in turn affect cognitive function, prompting the idea for its potential as a biomarker to track disease progression and assess therapeutic interventions (Ostrowski et al., 2021).

The Nacc1 R284W mutation elevates NACC1 protein levels and alters expression of its two isoforms in the brain

Western blot assays detecting NACC1 protein were performed with the polyclonal antibody from Abcam (#29047) which detects both long and short NACC1 isoforms and with polyclonal antibody from ThermoFisher (41656) which recognizes primarily the long isoform. It is worth noting again that only the long isoform encodes the R284W mutation (Fig. 1). NACC1 in lysates made from crude homogenates of mouse motor cortex of P7, P14, and 3- and 6-month-old mice expressed short and long isoforms that migrated to ∼57 and 67 kD, respectively, in Nacc1^+/+, Nacc1^+/R284W, and Nacc1^R284W/R284W mice (Fig. 7a–d). There was a slightly faster migration of both isoforms in mice with the mutation (best seen in Fig. 7b,c). A significant rise in short NACC1 protein levels occurred in P7 and P14 Nacc1^+/R284W and Nacc1^R284W/R284W mice, and in short and long NACC1 protein levels in Nacc1^+/R284W mice at 3 and 6 months compared with that in Nacc1^+/+ mice (Fig. 7a–d). These changes resulted in a net increase in total levels of NACC1 in mutant mice. The increase in NACC1 in 3-month-old Nacc1^+/R284W mice contrasted with a decline in FoxG1 (BF1), a protein present in adult pyramidal neurons (Fig. 7c).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Increased NACC1 protein in the cortex of Nacc1^+/R284W and NACC1^R284W/R284W mice correlates with changes in RNA isoforms. a–d, Shown are Western blots (WBs) of crude homogenates and quantitative analysis for P7 (a), P14 (b), 3-month-old (c), and 6-month-old (d) mice. Long (arrow) and short (arrowhead) NACC1 were detected in all age groups. Numbers on the left of WBs show migration of molecular mass markers in kDa. F and M indicate samples are from females or males, respectively. Graphs show mean protein levels ± SD relative to loading controls: GAPDH was used for P14 and 3 month data and actin for P7 and 6 month data. Females are represented as open circles and males closed circles. In the P7 mice [a, n = 8 Nacc1^+/+(4 female, 4 male), n = 8 Nacc1^+/R284W (4 female, 4 male), n = 8 Nacc1^R284W/R284W (4 female, 4 male)], there was a significant increase in the short isoform of NACC1 in Nacc1^+/R284W and Nacc1^R284W/R284W mice compared with Nacc1^+/+ mice with NACC1 antibodies from Abcam (one-way ANOVA, F = 23.62, p < 0.0001, Tukey's multiple-comparisons test, **p < 0.01, ****p < 0.0001, df = 21; males: one-way ANOVA, F = 125.6, p < 0.0001; females: one-way ANOVA, F = 11.15, p = 0.0037) and ThermoFisher 41656 (one-way ANOVA, F = 28.95, p < 0.0001, Tukey's multiple-comparisons test, **p < 0.01, ****p < 0.0001, df = 21; males: one-way ANOVA, F = 6.786, p = 0.0160; females: one-way ANOVA, F = 28.33, p = 0.0001). P7 samples were run on the same gel before genotyping was complete and white line indicate movement of lanes to group genotypes together. In the P14 mice (b, n = 5 Nacc1^+/+, n = 5 Nacc1^+/R284W, n = 3 Nacc1^R284W/R284W), there was a significant increase in total NACC1 (one-way ANOVA, F = 62.89, p < 0.0001, Tukey's multiple-comparisons test, p < 0.0001, df = 10) and the short isoform of NACC1 (one-way ANOVA, F = 534.1, p < 0.0001, Tukey's multiple-comparisons test, p < 0.0001, df = 10) in the Nacc1^+/R284W and Nacc1^R284W/R284W mice compared with that in Nacc1^+/+ mice. In the 3-month-old mice [c, Nacc1^+/+ (n = 10, 5 female, 5 male) and Nacc1^+/R284W (n = 8, 4 female, 4 male)], there was a significant increase in total NACC1 (unpaired t test, p < 0.0001, t = 14.32, df = 16; males: p < 0.0001, t =8.515, df = 7; females: p < 0.0001, t = 11.66, df = 7), the long isoform of NACC1 (unpaired t test, p < 0.0001, t = 7.106, df = 16; males: p < 0.0001, t =14.87, df = 7; females: p < 0.0001, t = 11.83, df = 7), and the short isoform of NACC1 (unpaired t test, p < 0.0001, t = 19.66, df = 16; males: p = 0.0054, t = 3.967, df = 7; females: p = 0.0003, t = 6.558, df = 7) and a decrease in FoxG1 (unpaired t test, p = 0.0009, t = 4.108, df = 15; males: p = 0.0149, t = 3.206, df = 7; females: p = 0.0374, t = 2.662, df = 6). In the 6-month-old mice [d, Nacc1^+/+ (n = 8, 4 female, 4 male) and Nacc1^+/R284W (n = 9, 5 female, 4 male)], there was a significant increase in total NACC1 (unpaired t test, p = 0.0053, t = 3.254, df = 15; males: p = 0.025, t = 2.968, df = 6; females: p = 0.1612, t = 1.567, df = 7), the long isoform of NACC1 (unpaired t test, p = 0.0479, t = 2.154, df = 15; males: p = 0.0866, t = 2.047, df = 6; females: p = 0.3907, t = 0.9149, df = 7), and the short isoform of NACC1 (unpaired t test, p = 0.0002, t = 4.860, df = 15; males: p = 0.0035, t = 4.659, df = 6; females: p = 0.041, t = 2.500, df = 7) and an increase in NACC1 Thermo (unpaired t test, p = 0.0253, t = 2.485, df = 15; males: p = 0.02393, t = 1.306, df = 6, females: p = 0.0267, t = 2.794, df = 7). e, f, NACC1 distribution in subcellular fractions. Nuclear and cytoplasmic fractions were prepared from the cortex of 3-month-old mice. Graph in e shows signal intensities of the nuclear fraction [mean ± SD, Nacc1^+/+ (n = 7, 3 female, 4 male) and Nacc1^+/R284W (n = 7, 4 female, 3 male)] normalized to GAPDH for levels of long (unpaired t test, p = 0.0227, t = 2.613, df = 12; males: p = 0.1268, t = 1.830, df = 5; females: p = 0.1525, t = 1.687, df = 5) and short (unpaired t test, p = 0.0056, t =3.368, df = 12; males: p = 0.0699, t = 2.298, df = 5; females: p = 0.0884, t = 2.112, df = 5). Histone deacetylase 3 (HDAC3) and 20S proteasome levels were unchanged between groups. Graph in f shows signal intensity of the cytoplasmic fraction [mean ± SD, Nacc1^+/+ (n = 8, 4 female, 4 male) and Nacc1^+/R284W (n = 8, 4 female, 4 male)] normalized to histone H3 with a significant increase in short NACC1 (unpaired t test, p = 0.0161, t = 2.737, df = 14; males: p = 0.1273, t = 1.769, df = 6; females: p = 0.0481, t = 2.475, df = 6) in the Nacc1^+/R284W mice. g, Western blot images from which data were obtained for graphs shown in e and f. h, NACC1 in synaptosome fractions. WB shows long and short NACC1 (at arrow and arrowhead, respectively) in synaptosomes of 3-month-old Nacc1^+/+, Nacc1^+/R284W, and Nacc1^R284W/R284W mice. i, Enrichment of pre- and postsynaptic markers, SNAP25 and PSD95, respectively, was confirmed in the synaptosome fractions. The numbers at the top of the blot in g and i identify each mouse. j, k, qRT-PCR analysis reveals altered expression of Nacc1 isoforms in Nacc1^+/R284W and Nacc1^R284W/R284W mice. j, Schematic shows the regions in Nacc1 that generate the long and short isoforms and the position of the spliced isoform. The red line indicates where the mutation is found. The striped area in the long isoform highlights the region not found in the short isoform. The position of the primers used to generate total Nacc1 are highlighted in green, the long isoform-specific primer in purple, and the short isoform-specific primer position in blue. ActB was the loading control in all qPCR analysis. k, Bar graphs show fold change in expression of total, long, and short isoform Nacc1 expression in P14 and 6-month-old mice. Statistics: A one-way ANOVA with Tukey's post hoc analysis in P14 samples, an unpaired Student’s t test for long isoform, and a Mann–Whitney test for total NACC1 and short isoform. For all panels, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

To determine the subcellular compartment where the increase in NACC1 isoforms occurred, we examined nuclear and cytoplasmic fractions prepared from cortex of crude homogenates of 3-month-old mice. There were higher levels of long and short NACC1 in nuclear fractions and in short NACC1 in cytoplasmic fractions in Nacc1^+/R284W mice compared with those in Nacc1^+/+ mice but no difference between genotypes in the relative distribution of nuclear and cytoplasmic NACC1 or its known interactors 20S proteasome and HDAC3 (Fig. 7e–g; Korutla et al., 2005; Shen et al., 2007). These results suggested that the mutation did not affect the nuclear cytoplasmic transport of NACC1 and some of its interactors. Where both sexes were examined at the same age (P7, 3 months, and 6 months; Fig. 7a,c–f), the increase in NACC1 in mutant mice favored a significant increase in females compared with that in males. In P7 and 3-month-old mice, the short isoform was increased in crude homogenates in both males and females (Fig. 7a,c), and in the nuclear–cytoplasmic fractionation, the short isoform was increased only in the cytoplasmic fraction of females (Fig. 7f). In 3-month-old mice, levels of NACC1 in cytoplasmic fractions were significantly elevated only in Nacc1^+/R284W females (Fig. 7f). The long isoform was significantly increased in both males and females of 3-month-old mice (Fig. 7c). Finally, synaptosome fractions prepared from the cortex of 3-month-old mice also expressed the short and long isoforms, in all genotypes (Fig. 7h,i), lending support to the idea that WT and mutant NACC1 are positioned to directly influence synaptic function.

Altogether, these findings showed that the two NACC1 isoforms increased in an age-dependent manner with the short isoform rising earlier than the long isoform which harbors the mutation. The R284W mutation did not affect the relative nuclear cytoplasmic distribution of NACC1 or its localization within the synaptic compartment.

Next, we addressed if the increases in NACC1 protein levels in mutant mice involved changes in the levels of RNA encoding each NACC1 isoform. qPCR assays were performed with RNA prepared from the cortex of P14 and 6-month-old mice. Total Nacc1 RNA levels, measured with a primer pair that detected both isoforms, did not differ between genotypes (Fig. 7j,k). The long isoform was diminished in the cortex of P14 Nacc1^R284W/R284W compared with that of Nacc1^+/+ mice, whereas RNA specific to the short isoform increased at P14 in Nacc1^R284W/R284W and at P14 and 6 months in Nacc1^+/R284W mice (Fig. 7k). The net effect was approximately a fourfold decrease in the ratio of long/short isoform RNA levels in Nacc1^+/R284W mice compared with that in Nacc1^+/+ mice at both ages without a change in total Nacc1 RNA expression between genotypes. Potential contributors to augmented levels of NACC1 isoforms in Nacc1^+/R284W mice with no net change in RNA levels include altered RNA stability, reduced clearance of the mutant protein, or enhanced sensitivity of anti-NACC1 antibodies to secondary structure or post-translational modification of mutant protein. For example, a putative site for SUMOylation in NACC1 at lysine residue K167 of human protein along with SUMO-interacting motifs mediates post-translational modification that may be important for NACC1 function in transcription and localization to nuclear bodies (Tatemichi et al., 2015). Also, Daniel et al. (2023) showed that levels of exogenously expressed WT mouse NACC1 (using sequence for the long isoform) was increased when coexpressed with SUMO.

NACC1 immunoreactivity is increased in neurons of Nacc1^+/R284W mice

The results above showed that Nacc1^+/R284W mice displayed behavioral and motor deficits, seizures, cortical EEG abnormalities, and increased levels of NACC1 isoforms in the cortex. To assess overall brain architecture, we stained sections from brains of Nacc1^+/+ and Nacc1^+/R284W mice with cresyl violet and Luxol fast blue and cross-sectional areas of major gray and white matter regions were delineated (Fig. 8a–f). No regional differences were obtained between genotypes for gray and white matter except for the caudate–putamen which was slightly but significantly smaller in Nacc1^+/R284W mice compared with Nacc1^+/+ mice (Fig. 8e). It is possible that the caudate–putamen in mutant mice has fewer projections from the motor cortex and other cortical areas resulting in its diminished size.

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

No gross morphological changes in Nacc1^+/R284W mice. a–f, Mouse brain sagittal sections stained with cresyl violet for gray matter and Luxol fast blue for white matter. a, c, Representative cresyl violet-stained sagittal sections of a Nacc1^+/+ mouse (a) and Nacc1^+/R284W mouse (c). b, d, Representative Luxol fast blue-stained sagittal section of a Nacc1^+/+mouse (b) and a Nacc1^+/R284W mouse (d). Brain regions quantified for differences in area are identified in b. Acronyms (arrows point to white matter tracts quantified): Ctx, cortex; CPu, caudate–putamen; Hip, hippocampus; Th, thalamus; Cb, cerebellum; P+M, pons and medulla; C.C., corpus callosum; Fx, fornix; Ant, anterior commissure. e–f, Quantification of area for gray and white matter using Luxol fast blue matched Nacc1^+/+ and Nacc1^+/R284W images. e, Area quantification of gray matter regions. Caudate–putamen showed a significantly reduced area in Nacc1^+/R284W mice compared with Nacc1^+/+ mice (unpaired Student’s t test, p = 0.0072, t = 5.052, df = 4). f, Area quantification of white matter regions. No changes for matched images were identified.

Next, we examined immunofluorescence labeling of NACC1 in the cortex using the same polyclonal antibody that detected long and short NACC1 isoforms (Abcam #29047) by Western blot shown in Figure 7. NACC1 localization was widespread in the brain and more intense in the neocortex of Nacc1^+/R284W mice than that of Nacc1^+/+ mice (Fig. 9a–d). NACC1 colocalized with the nuclear-enriched neuronal protein NeuN (Rbfox3) in all cortical layers. There was a patchy distribution of nuclear NACC1 (Fig. 9b, inset) that was previously shown in neurons in vitro to correspond to nuclear speckles (aka SC35 domains; Hall et al., 2006; Shen et al., 2007). Quantification of NACC1 immunoreactivity in the motor cortex revealed more intense staining in all cortical cell layers of Nacc1^+/R284W compared with that of Nacc1^+/+ mice and more in layer VI of Nacc1^R284W/R284W than that of Nacc1^+/R284W mice (Fig. 9d,e). In addition to pyramidal cells, neocortical neurons positive for parvalbumin, a calcium ion binding protein that is a marker for fast spiking GABA inhibitory interneurons, also had elevated NACC1 immunoreactivity in Nacc1^+/R284W and Nacc1^R284W/R284W mice (Fig. 9f). Increased nuclear NACC1 also appeared in neurons in caudate–putamen (Fig. 9h–j) and nucleus accumbens (Fig. 9k–m) and CA1 region of the hippocampus (Fig. 10c) of Nacc1^+/R284W and Nacc1^R284W/R284W mice. Altogether, these data showed widespread neuronal nuclear localization of NACC1 in the brain of Nacc1^+/+ mice that was amplified by the NACC1 mutation in mutant mice and included cortical and hippocampal pyramidal neurons, which are glutamatergic excitatory cells, and parvalbumin-containing neurons which are GABAergic inhibitory neurons. The increased NACC1 immunoreactivity in the brain of mutant mice is consistent with results by Western blot using the same antibody from Abcam (Fig. 7) except that the two isoforms cannot be distinguished by immunohistochemical labeling as they are by Western blot.

• Download figure
• Open in new tab
• Download powerpoint

Figure 9.

NACC1 immunofluorescence is increased in neuronal nuclei of Nacc1^+/R284W and Nacc1^R284W/R284W mice. Cresyl violet-stained images in a and c are mouse brain sections seen in the coronal plane adapted from the Allen Mouse Brain Atlas (http://mouse.brain-map.org). a, The boxed area of the visual cortex indicates the brain region corresponding to the 10× image to the right. For the 10× immunofluorescence image, NACC1 was labeled with rhodamine red (red) and NeuN with AlexaFluor888 (green). NACC1 colabeling with NeuN appears uniform throughout all layers of the cortex and more intense in Nacc1^+/R284W mice. b, High magnification z-plane maximum projections from 30 µm sections imaged from panel a. NACC1 can be seen as nuclear speckles (inset) in Nacc1^+/+ mice. In Nacc1^+/R284W mice, the staining is much more intense. c, The boxed area of the motor cortex indicates the brain region corresponding to the 10× image to the right. For the 10× immunofluorescence image, NACC1 was labeled with rhodamine red (red) and NeuN with AlexaFluor488 (green). NACC1 colabeling with NeuN appears uniform throughout all layers of the cortex and more intense in Nacc1^+/R284W and Nacc1^R284W/R284W mice. d, Representative images of NACC1 and NeuN staining used for quantification in e. Images are single confocal z-planes of layer V of the motor cortex of 2-month-old mice used for quantitative data. NACC1 is labeled in rhodamine red (red) and NeuN with AlexaFluor488 (green). e, Graphs show increased fold change of NACC1 intensity in motor cortex layers 2/3, 5, and 6 of P60 mice (each dot represents a nucleus; red lines depict median ± IQR, n = 120 cells per layer per mouse from 4 mice per genotype; statistics: layer 2/3, Kruskal–Wallis test with Dunn's multiple comparisons (p < 0.0001, Kruskal–Wallis statistic = 154.3); layer 5, Kruskal–Wallis test with Dunn's multiple comparisons (p < 0.0001, Kruskal–Wallis statistic = 219.8); layer 6, Kruskal–Wallis test with Dunn's multiple comparisons (p < 0.0001, Kruskal–Wallis statistic = 198.3)]. f, Single confocal z-plane images of parvalbumin labeled neurons (arrow in white box) of the motor cortex of 2-month-old mice shows increased NACC1 staining intensity in Nacc1^+/R284W and Nacc1^R284W/R284W mice. NACC1 is labeled with rhodamine red and parvalbumin with AlexaFluor488. g, A cresyl violet-stained representative image adapted from the Allen Mouse Brain Atlas (http://mouse.brain-map.org) representing the areas of caudate–putamen and nucleus accumbens (boxed regions 1 and 2, respectively) where images in h–m were taken. h–j, NACC1 immunoreactivity in the caudate–putamen colocalizes with neuronal marker NeuN and appears increased in the Nacc1^+/R284W and Nacc1^R284W/R284W genotypes compared with that in Nacc1^+/+ genotypes. Insets show only the NACC1 channel. k–m, NACC1 immunoreactivity in neurons of the nucleus accumbens are colocalized with NeuN. NACC1 immunoreactivity is increased in the Nacc1^+/R284W and Nacc1^R284W/R284W genotypes compared with that in Nacc1^+/+ genotypes. Insets show only the NACC1 channel. For all images shown, a single z-plane of the middle section of the tissue was used. Scale bars in the bottom left of h and k are 50 µm and apply to all images. NeuN was labeled with AlexaFluor488 and identifies differentiated neurons (green). NACC1 was labeled with rhodamine red (red). For all panels, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

• Download figure
• Open in new tab
• Download powerpoint

Figure 10.

NACC1 immunoreactivity is detected in glial nuclei and some processes. a, Cresyl violet-stained image of brain coronal section at level of the visual cortex adapted from the Allen Mouse Brain Atlas (http://mouse.brain-map.org). Boxed regions are where images were taken. Images in panels b and c are from the corpus callosum and hippocampus (CA1 region), respectively, of 3-month-old mice. Images in b and c were captured as maximum z-plane projections of 30 µm sections. NACC1 is labeled with rhodamine red and NeuN and GFAP with Bodipy (green). b, In the corpus callosum, NACC1 and GFAP immunoreactivities are colocalized in astrocytic processes which are reduced in Nacc1^+/R284W mice. c, In the hippocampus CA1 region, GFAP-labeled astrocytic processes (arrows) that have NACC1 immunoreactive nuclei (arrowheads) are reduced in Nacc1^+/R284W mice. d, Immunofluorescent single z-plane confocal labeling of NACC1 in nuclei of glia, identified by the presence of S100b (mature astrocytes) or olig2 (mature oligodendrocytes) in Nacc1^+/+, Nacc1^+/R284W, and Nacc1^R284W/R284W mice. Images were taken from 2-month-old mice from the motor cortex (s100b) or corpus callosum (Olig2) analogous to the brain region shown in Figure 9c. White arrowheads in boxes identify corresponding cells within the larger 40× image and the inset where only the NACC1 channel is visible in red. Immunofluorescent staining intensity for NACC1 is generally low. It should be noted in Nacc1^R284W/R284W mice that there is a s100b-labeled cell (yellow box) that has more nuclear NACC1 immunofluorescence than in Nacc1^+/+. e, NACC1 is not detected in Iba1-labeled microglia of Nacc1^+/+ and Nacc1^+/R284W mice at 3 months of age. Images show localization of immunoreactive Iba1, a marker for microglia in Nacc1^+/+ and Nacc1^+/R284W mice in the primary visual cortex (V1). The first two images from the left show Iba1 alone (green) and the third and fourth images are from sections that were costained for Iba1 (green), NACC1 (red), and Hoechst dye (blue). No colocalization of Iba1 and NACC1 is evident.

In contrast to neurons, astrocytes in the corpus callosum and hippocampus identified by the presence of immunoreactive glial fibrillary acidic protein (GFAP) had similar levels of nuclear NACC1 labeling in Nacc1^+/+ and Nacc1^+/R284W mice (Fig. 10c, arrowhead). However, a surprising finding was that NACC1 also colocalized with GFAP-positive glial processes and GFAP staining intensity in these processes was reduced in Nacc1^+/R284W mice compared with that in Nacc1^+/+ mice (Fig. 10c, arrows). It was unclear if the processes were present in these glial cells or had been diminished by the mutation. Most astrocytes in the cortex that were identified with the specific marker calcium binding protein B (S100B) and oligodendroglia marked by oligodendrocyte transcription factor 2 (Olig2) had similarly low levels of nuclear NACC1 immunoreactivity in Nacc1^+/+ and Nacc1^+/R284W mice (Fig. 10d), although a few cells in Nacc1^R284W/R284W had more intense staining (Fig. 10d, yellow arrow). Microglia immunoreactive for ionized calcium binding adaptor molecule 1 (Iba1) did not show colocalized NACC1 in either the cell body or nucleus (Fig. 10e). In summary, NACC1 was detected at similarly low levels in most nuclei of astrocytes and oligodendroglia of Nacc1^+/+ and mutant mice and was absent in microglia.

RNA-seq and pathway analysis in the cortex of Nacc1^+/R284W and Nacc1^R284W/R284W mice reveal pervasive molecular changes at the transcriptome level

Due to the role of NACC1 as a transcriptional repressor and its prevalence in the nucleus of cortical neurons described above, we used RNA-seq to determine the impact of the Nacc1R284W mutation on gene regulation. RNA extraction and sequencing from samples of the motor cortex of P14 Nacc1^+/+, Nacc1^+/R284W, and Nacc1^R284W/R284W mice were performed (n = 3 per genotype and sex) as detailed in Materials and Methods. This age was selected for the analysis due to the peak in synapse formation (Farhy-Tselnicker and Allen, 2018). Nacc1 was not found to be differentially expressed by RNA sequencing in Nacc1^+/R284W or Nacc1^R284W/R284W mice. Consistent with changes in levels of Nacc1 isoforms seen by qPCR (Fig. 7j,k), RNA sequencing showed the percentage of reads were reduced for long Nacc1 isoform and increased for short Nacc1 isoforms (Fig. 11a–c).

• Download figure
• Open in new tab
• Download powerpoint

Figure 11.

RNA-sequencing reads in P14 mice depict percent change in Nacc1 isoform expression in a mutation dose response. a, Representative female Nacc1^+/+ (red), Nacc1^+/R284W (gray), and Nacc1^R284W/R284W (orange) sashimi plots from P14 RNA-sequencing data. Below the plots is a map (blue) showing the corresponding splice sites in the Nacc1 gene of the short and the long isoforms. Each sashimi plot shows the number of reads spanning both the long and the short isoform exon–exon junctions. On top of each plot are the long and short exon–exon reads as well as the total Nacc1 reads for the gene. b, c, Percent of reads for long and short Nacc1 isoforms in the cortex of P14 and 6-month-old Nacc1^+/+, Nacc1^+/R284W, and Nacc1^R284W/R284W mice. Data shown as mean ± SDs as percentage of reads obtained from RNA-seq data for the isoform indicated compared with the total number of short and long isoform reads from the same mouse sample. b, The long isoform for P14 mice shows a significant reduction in the percent of reads in a gene mutation dose-dependent manner (one-way ANOVA, p < 0.0001, F = 41.39, with Tukey's multiple comparison). c, The short isoform shows an increase in the isoform's expression (one-way ANOVA, p < 0.0001, F = 41.39, with Tukey's multiple comparison). p values are represented as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Volcano plots (Fig. 12a,b) display the number and distribution of DEGs with a nominal p value cutoff of <0.05 (complete list of DEGs can be found in Extended Data Fig. 12-1). These genes were used for pathway and GO analysis (discussed below; a complete list of GO, KEGG, and Reactome datasets can be found in Extended Data Figs. 12-2, 12-3, 12-4, respectively). More DEGs were identified in P14 Nacc1^R284W/R284W mice (3,822 DEGs of 27,968 total transcripts) than those in Nacc1^+/R284W (3,268 DEGs of 28,188 transcripts) mice and after applying the Bonferroni p-adjusted cutoff of p < 0.05 (1,355 for Nacc1^R284W/R284W and 1,109 for Nacc1^+/R284W). Of these DEGs, 41 were significantly changed across the three genotypes supporting a mutant gene dose effect, and 38 of the 41 were downregulated (Tables 5, 6).

• Download figure
• Open in new tab
• Download powerpoint

Figure 12.

RNA-sequencing results show widespread changes in transcripts of P14 Nacc1^+/R284W and Nacc1^R284W/R284W mice. a, b, Punches from the motor cortex of 6 Nacc1^+/+, 6 Nacc1^+/R284W, and 6 Nacc1^R284W/R284W (3 males and 3 females from each genotype) underwent RNA sequencing. Volcano plots of DEGs found in Nacc1^+/R284W (a) and Nacc1^R284W/R284W (b) mice compared with wild-type at the nominal p value cutoff of <0.05 are plotted. Blue indicates downregulated genes and red indicates upregulated genes. Nonsignificant genes are gray and the top 10 up- and downregulated genes are labeled. All DEGs can be found in Extended Data Fig. 12-1. c, d, Venn diagrams showing overlap between Nacc1^+/R284W and Nacc1^R284W/R284W DEGs at the nominal p value cutoff for upregulated genes, c, and downregulated genes, d. Top 10 genes identified from the volcano plot are depicted for each group they belong to. e–h, GO terms derived from upregulated and downregulated nominal p value cutoff DEGs from Nacc1^+/R284W and Nacc1^R284W/R284W were compared and common GO terms extracted. The top 5 GO terms are shown from the biological process (BP, green), cellular component (CC, blue), and molecular function (MF, orange) categories which are common to both mutant genotypes. Black boxes highlight common terms found in the top 5 common GO terms in both Nacc1^+/R284W and Nacc1^R284W/R284W results. Numbers found in the bubbles represent the number of DEG genes enriched in that pathway. All enriched GO terms, KEGG pathway, and Reactome pathways results can be found in Extended Data Figs.12-2, 12-3, and 12-4 respectively.

The top 10 upregulated and downregulated DEGs of Nacc1^+/R284W and Nacc1^R284W/R284W mice each compared with those of Nacc1^+/+ mice are depicted in Figure 12a,b. Common top 10 upregulated and downregulated genes between Nacc1^+/R284W and Nacc1^R284W/R284W are shown in Figure 12c,d, respectively. Common upregulated genes between Nacc1^+/R284W and Nacc1^R284W/R284W include Kirrel2, Scel, Egfl6, Ephx1, and Pdzph1. Among these Kirrel2 which encodes Kirrel-like nephrin family adhesion molecule 2 is important in cell adhesion and axon targeting (Vaddadi et al., 2019). Common downregulated genes between Nacc1^+/R284W and Nacc1^R284W/R284W include Mfge8, Cbs, Lfng, Lcat, Cst3, and Entpd2, all of which except Lcat were affected in a mutant gene dose-dependent manner. Two of these transcripts, Mfge8, which encodes lactadherin, and Lcat which encoded lecithin–cholesterol acetyltransferase, are important in glial function (Boddaert et al., 2007; Hirsch-Reinshagen et al., 2009).

Sex-specific DEG changes were compared for Nacc1^+/R284W and Nacc1^R284W/R284W mice (n = 3 mice per sex for each genotype). The number of DEGs unique to Nacc1^+/R284W and Nacc1^R284W/R284W females was greater than that in corresponding males for upregulated transcripts (157 vs 6 in Nacc1^+/R284W mice and 157 vs 120 in Nacc1^R284W/R284W mice) and downregulated transcripts (82 vs 17 in Nacc1^+/R284W mice and 182 vs 69 in Nacc1^R284W/R284W mice). Although the top 3 DEGs for each sex and genotype group were different, three of the transcripts in Nacc1^+/R284W mice encoded ribosomal proteins and included upregulated Rps27 in males and downregulated Rps27rt in males and Rps28 in females. Also noteworthy in Nacc1^R284W/R284W females among the top 3 downregulated DEGs were Mag and Mobp, which encode proteins that are enriched in oligodendrocytes and are important in white matter formation.

GO term analysis revealed hundreds of enriched terms: upregulated genes in Nacc1^+/R284W mice compared with those in Nacc1^+/+ mice enriched for 470 total terms [314 biological processes (BP), 68 cellular components (CC), and 88 molecular functions (MF)], while downregulated genes enriched for 136 total terms (85 BP, 43 CC, and 8 MF). Upregulated genes from Nacc1^R284W/R284W mice compared with Nacc1^+/+ mice enriched for 406 total GO terms (295 BP, 46 CC, 65 MF), while downregulated genes enriched for 221 total terms (173 BP, 29 CC, and 19 MF). To determine high-confidence GO terms changed in both mutant genotypes, we compared the upregulated and downregulated GO terms between Nacc1^+/R284W and Nacc1^R284W/R284W mice (Fig. 12e–h). The top upregulated GO terms common to Nacc1^+/R284W and Nacc1^R284W/R284W mice were related to dendrite development, synaptic organization including the postsynaptic membrane and ion channel activity (Fig. 12e,f, respectively, black box). Common downregulated GO terms between Nacc1^+/R284W and Nacc1^R284W/R284W mice enrich for various metabolic functions, mitochondrial function, and structural constituents of the ribosome (Fig. 12g,h, respectively, black box). When all DEGs were included for the GO analysis (up and down), there was still enrichment for postsynaptic terms for Nacc1^+/R284W and Nacc1^R284W/R284W mice. GO analysis also highlighted DEGs encoding multiple types of potassium, calcium, and sodium channels, among them transcripts linked to seizure phenotypes (Table 7).

Sex-specific GO terms were compared for each genotype (n = 3 mice per sex for each genotype). In Nacc1^+/R284W mice the top 3 unique terms enriched for upregulated genes in females were muscle tissue development (BP), presynaptic active zone (CC), and delayed rectifier potassium channel activity (MF) and in males were cell recognition (BP), adherens junctions (CC), and PDZ domain binding (MF). The top GO terms enriched for downregulated transcripts in Nacc1^+/R284W females included mitochondrial ATP synthesis coupled electron transport chain (BP), inner mitochondrial membrane protein complex (CC), and rRNA binding (MF) whereas those in males were angiogenesis (BP), extracellular matrix (CC), and integrin binding (MF). In Nacc1^R284W/R284W mice, unique GO terms with the most upregulated DEGs in females were positive regulation of synapse assembly (BP), filopodium (CC), and receptor binding (MF) and in males were positive regulation of catabolic process (BP), nuclear matrix (CC), and proximal promotor sequence specific DNA binding (MF). Unique GO terms with the most downregulated DEGs in females were ensheathment of neurons (BP), cell to cell junction (CC), and heparin binding (MF) whereas in males they were energy derivation by oxidation of organic compounds (BP), mitochondrial inner membrane (CC), and cell adhesion mediator activity (MF). These top GO term results suggested that both sexes of mutant mice had transcript changes that reflected diverse areas of dysfunction with females showing more synapse-related changes and males more adhesion and extracellular matrix alterations.

Remarkably, most DEGs encoding transcripts enriched in glia were downregulated in P14 Nacc1^+/R284W mice and Nacc1^R284W/R284W mice (Table 8). These included DEGs Mfge8 and Lcat as noted above as well as Gfap and Aldh1l1. The latter two normally have their highest levels in mice at P14 when synapse formation and maturation are most active and glia play an important role (Farhy-Tselnicker and Allen, 2018). Aldh1l1 also exhibited a gene mutation dose effect (Table 8) as did its protein product ALDH1L1, and both declined progressively in the cortex of P14 Nacc1^+/R284W mice and Nacc1^R284W/R284W mice (Fig. 13a,b).

• Download figure
• Open in new tab
• Download powerpoint

Figure 13.

Levels of ALDH1L1 are reduced in the cortex of P14 Nacc1^+/R284W mice. Western blots in a and quantitative data in b for ALDH1L1 levels in the motor cortex of P14 mice. The lysates are from the same brains of female mice used for RNA-seq analysis shown in Figure 12. Graphs in b depict mean ± SD, n = 5 Nacc1^+/+, 5 Nacc1^+/R284W, 3 Nacc1^R284W/R284W, one-way ANOVA (p < 0.0001, F = 38.52) with Tukey's multiple-comparisons test. c–f, Reverse transcriptase quantitative PCR confirmation of RNA-sequencing DEGs. c, Gfap expression was reduced in Nacc1^+/R284W and Nacc1^R284W/R284W mice (one-way ANOVA, p < 0.0001, F = 22.59 with Tukey's multiple-comparisons test). d, Cort expression was down in Nacc1^+/R284W and Nacc1^R284W/R284W mice (one-way ANOVA, p < 0.0001, F = 18.18 with Tukey's multiple-comparisons test). e, Hsd11b1 expression was reduced in Nacc1^+/R284W and Nacc1^R284W/R284W mice (one-way ANOVA, p < 0.0001, F = 48.23 with Tukey's multiple-comparisons test). One female Nacc1^+/R284W mouse was removed due to being an outlier (ROUT Q = 1%). f, Kcne1l1 was down in Nacc1^+/R284W and Nacc1^R284W/R284W mice (one-way ANOVA, p < 0.0001, F = 27.55 with Tukey's multiple-comparisons test). One female Nacc1^+/R284W mouse was removed due to being an outlier (ROUT Q = 1%). All graphs show mean ± SD. Females are represented as open circles and males closed circles. p values are represented as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Many transcripts encoding neuropeptides were also downregulated. These included the genes encoding cortistatin (Cort), somatostatin receptor 2 (Sstr2), neurotensin (Nts), neurotensin receptor 1 and 2 (Ntsr1 and Ntsr2, respectively), and substance P (Tac1). In addition to their roles in synaptic function in the brain, some of these peptides play an important role in GI tract gut motility including cortistatin, neurotensin (and its receptor), and substance P (Yau, 1978; Gonzalez-Rey et al., 2006; Steinhoff et al., 2014; Ratner et al., 2016). GI dysfunction is a significant problem in patients with NACC1^+/R298W (Schoch et al., 2017). NACC1 is localized to the epithelium of the GI tract (Yap et al., 2012) and could directly influence the action of these peptides.

To validate the results of the RNA-seq, qPCR was performed for four transcripts of interest that had a high fold change (Gfap, Cort, Hsd1b11, and Kcne1l1). Results showed that the four genes tested had reductions in fold change expression that correlated with those obtained by RNA-seq (Fig. 13c–f).

In summary, results of RNA-seq transcript and pathway analysis in Nacc1^+/R284W and Nacc1^R284W/R284W mice uncovered altered transcriptomes that implicate dysregulation of distinct cellular functions including neurotransmission. These molecular changes likely contribute to phenotypes displayed in Nacc1^+/R284W mice including behavioral and motor deficits, behavioral seizures, and aberrant cortical EEG activity.

The cortex of Nacc1^+/R284W mice have altered levels of proteins that function at synapses but no change in synaptic density

RNA-seq and pathway analysis described above suggested the Nacc1 R>W mutation altered synaptic connectivity. In support of this idea, Western blots of crude homogenates from the cortex of Nacc1^+/R284W showed significant changes in levels of synaptic proteins examined at P14, 3 months, and 6 months. At P14, where only lysates from females were examined, altered levels were detected in the vesicle membrane proteins synaptophysin and VAMP2, the presynaptic protein SNAP25, and in homozygotes the receptor GABRb1 (GABA type A receptor subunit; Fig. 14a). The transcripts encoding two of these proteins, VAMP2 and synaptophysin, were not differentially expressed based on results from RNA-seq performed in the same mice indicating the protein changes were post-translational. Older mice also had synaptic protein changes. Males at 3 months and females at 6 months had a reduction in Prosapip1 (Fig. 14b,c), a scaffold protein enriched in the postsynaptic density of dendritic spines (Wendholt et al., 2006; Reim et al., 2016). Also in 6-month-old mutant mice presynaptic SNAP25 was significantly changed in males and the postsynaptic protein Homer in females (Fig. 14c). These findings support the idea that protein changes exist pre- and postsynaptically in Nacc1^+/R284W mice and depend on age and gender.

• Download figure
• Open in new tab
• Download powerpoint

Figure 14.

Nacc1^+/R284W mice have changes in levels of synaptic proteins but normal neuropil ultrastructure. a–c, Western blots of synaptic proteins detected in the cortex of crude homogenates from P14 mice [Nacc1^+/+ (n = 5 females), Nacc1^+/R284W (n = 5 females), and Nacc1^R284W/R284W (n = 3 females)], 3-month-old mice [Nacc1^+/+ (n = 3 males) and Nacc1^+/R284W (n = 2 males, 1 female)], and 6-month-old mice [Nacc1^+/+ (n = 8, 4 males, 4 females) and Nacc1^+/R284W (n = 9, 4 males, 5 females)]. Graphs show means and SDs for signal intensity normalized to GAPDH in a and b and to actin in c. Females are represented as open circles and males closed circles. In the P14 mice, there was a significant increase in VAMP1 (one-way ANOVA, p = 0.0144, F = 6.683 with Tukey's multiple-comparisons test *p = 0.0159), synaptophysin (unpaired t test, *p = 0.0308, t = 2.617, df = 8), SNAP25 (one-way ANOVA, p = 0.002, F = 12.41 with Tukey's multiple-comparisons test, Nacc1^+/+ vs Nacc1^+/R284W **p = 0.0044, Nacc1^+/+ vs Nacc1^R284W/R284W **p = 0.0048), and GABAa beta1 (one-way ANOVA, p = 0.0127, F = 6.964 with Tukey's multiple-comparisons test *p = 0.01) in the Nacc1^+/R284W or Nacc1^R284W/R284W mice compared with that in Nacc1^+/+ mice. In the 3-month-old mice (b), there was a significant decrease in Prosapip1 in the Nacc1^+/R284W compared with that in Nacc1^+/+ mice (unpaired t test, *p = 0.0144, t = 4.137, df = 4). In the 6-month-old mice (c), there was a significant increase in Homer (unpaired t test, *p = 0.0106, t = 2.919, df = 15; males: p = 0.2898, t = 1.161, df = 6; females: p = 0.001, t = 5.407, df = 7) and SNAP25 (unpaired t test, *p = 0.0117, t = 2.868, df = 15; males: p = 0.0004, t = 6.952, df = 6; females: p = 0.8844, t = 0.1508, df = 7) and a significant decrease in Prosapip1 (unpaired t test, *p = 0.0195, t = 2.615, df = 15; males: p < 0.0001, t = 9.172, df = 6; females: p = 0.8042, t = 0.2575, df = 7) in the Nacc1^+/R284W compared with that in Nacc1^+/+ mice. d, e, Electron microscopic images of motor cortex neuropil from 3-month-old Nacc1^+/+ and Nacc1^+/R284W mice that are examples of those used for analysis of synaptic density. Arrows and arrowheads point to postsynaptic sites at spines and dendrites, respectively. Scale bar, 500 nm. f–h, Number of synaptic contacts and total and average lengths of synapses in Nacc1^+/+ and Nacc1^+/R284W mice. Box and whisker plots are results from Mann–Whitney U nonparametric test. The bar in the box represents the median and the whiskers represent the 5–95% percentile, f, p = 0.6738, g, p = 0.2895, h, p = 0.5931, n = 102 images from 2 Nacc1^+/+ mice and n = 101 images from 2 Nacc1^+/R284W mice.

Finally, we looked at the density of synapses in cortex neuropil of 3-month-old Nacc1^+/+ and Nacc1^+/R284W mice and observed axospinous and axodendritic synapses in both genotypes (Fig. 14d,e) and no significant difference in the number of synapses or total or average length of synaptic contacts between genotypes (Fig. 14f–h). Thus, despite evidence for changes in transcripts and proteins that implicate synaptic impairment in Nacc1^+/R284W, synaptic architecture at the ultrastructural level in 3-month-old mice was intact.

Summary of results

In all, the comprehensive description provided above shows that Nacc1^+/R284W mice exhibit multiple phenotypes in postnatal and adult life at the behavioral, electrophysiological, and molecular level. Some phenotypes were affected by age, sex, and mutant gene dose. There are robust readouts such as hindlimb clasping, epileptiform discharges, and changes in brain levels of Nacc1 transcript, protein, and localization, as well as other transcripts and proteins, that will be useful as quantifiable endpoints for evaluating underlying mechanisms of neurodevelopmental disease and therapeutic interventions.