Loss of the neurodevelopmental gene Zswim6 alters striatal morphology and motor regulation

Abstract

The zinc-finger SWIM domain-containing protein 6 (ZSWIM6) is a protein of unknown function that has been associated with schizophrenia and limited educational attainment by three independent genome-wide association studies. Additionally, a putatively causal point mutation in ZSWIM6 has been identified in several cases of acromelic frontonasal dysostosis with severe intellectual disability. Despite the growing number of studies implicating ZSWIM6 as an important regulator of brain development, its role in this process has never been examined. Here, we report the generation of Zswim6 knockout mice and provide a detailed anatomical and behavioral characterization of the resulting phenotype. We show that Zswim6 is initially expressed widely during embryonic brain development but becomes restricted to the striatum postnatally. Loss of Zswim6 causes a reduction in striatal volume and changes in medium spiny neuron morphology. These changes are associated with alterations in motor control, including hyperactivity, impaired rotarod performance, repetitive movements, and behavioral hyperresponsiveness to amphetamine. Together, our results show that Zswim6 is indispensable to normal brain function and support the notion that Zswim6 might serve as an important contributor to the pathogenesis of schizophrenia and other neurodevelopmental disorders.

Keywords: Hyperactivity; Medium spiny neurons; Motor behavior; Neurodevelopment; Schizophrenia; Striatum; ZSWIM6.

Fig. 1. Zswim6 expression during striatal development

At embryonic day (E) 12.5 (A–C) and E14.5 (D–F) Zswim6 mRNA is strongly expressed in the subventricular zone of the lateral, medial and caudal ganglionic eminences but at low to undetectable levels in the cerebral cortex. Low-level expression is also seen in the amygdala (arrow in F) and thalamus (G–I). At E16.5 this expression expands to the cortical plate and medial habenula (arrow in I) and increases in intensity in the amygdala (arrow in H). (J) Image from the Allen Brain Atlas shows that expression of Zswim6 is present in the adult striatum but not in the overlying cerebral cortex. (K) Quantitative PCR (qPCR) shows that Zswim6 is expressed at higher levels in the LGE than in the MGE or cortex at E13.5 (N=3 mice for each region). (L) qPCR for Zswim4, Zswim5, and Zswim6 across striatal development shows that while Zswim4 and Zswim5 are largely downregulated by adulthood, Zswim6 expression is maintained (N=3 mice for each region). Abbreviations: LGE (lateral ganglionic eminence); MGE (medial ganglionic eminence); CGE (caudal ganglionic eminence); Cx (cortex); Th (thalamus); Am (amygdala); Hb (habenula); Hy (hypothalamus); St (striatum). Scale bars 300μm in A–C, 400μm in D–F, 500μm in G–I, 1000μm in J. ** p<.01. Error bars indicate SEM.

Fig. 2. Generation of Zswim6 knockout mice

(A) Schema of constructs used in generating a knockout (KO) allele for Zswim6 by flanking exon 4 with LoxP insertions, generating a mouse line, then crossing this line with Cre-deleter mice. (B) Southern blot analysis to identify correctly targeted embryonic stem (ES) cell clones obtained from hybrid C57BL/6x129/SvEv ES cells electroporated with the targeting construct. Four ES clones (Z1–Z4) were identified in which there were the expected 5′ and 3′ recombinations. HYB, hybrid ES cell; B6, C57BL/6 ES cell; 129, 129/SvEv ES cell. (C) In situ hybridization (ISH) for Zswim6 using a probe spanning from exon 3 to the beginning of exon 5 shows loss of expression in coronal sections from embryonic day (E) 14.5 forebrain of homozygous Zswim6 KOs. (D) RT-PCR reveals a loss of product created by primers flanking the exon 4 and 5 junction in the Zswim6 KO. (E) qPCR using a primer pair downstream from the deleted exon shows that there is a significant reduction in the relative abundance of Zswim6 mRNA in the MGE of E13.5 homozygous Zswim6 KO mice (N=3 mice). Abbreviations: WT (wild type); Neo (neomycin resistance cassette); E (exon). ***p<.001. Scale bar 400μm in C. Error bars indicate SEM.

Fig. 3. Decreased early postnatal survival and forebrain size in Zswim6 KOs

(A) Survival plot showing that the Zswim6 KO mice have a 40% mortality in the first few days after birth, followed by a very gradual decline through 4 weeks. Data is from approximately 100 mice from each group. (B) Analysis of the % of litters that are wild type (WT), heterozygous (HET) or KO for Zswim6 at embryonic day (E) 14.5, post-natal day (P) 0, and P21 is consistent with loss of KO pups after birth. (N=79 mice for P21, Χ² (2) = 27.911) (C) Zswim6 KOs also show a small decrease of weight relative to controls at P21 that persists into adulthood (N=10 mice). (D) Overhead surface area of the cerebral cortex (Cx) at P80 shows a significant decrease in the Zswim6 KOs, but no change in the cerebellum (CRBL; N=5 brains). Volumetric measurements based on analysis of coronal tissue sections show a significant reduction of volume in both the Cx (N=5 WT, 6 KO brains) and the striatum (St) at P80 (N=5 brains), as well as in the striatum at P21 (N=4 brains), with a borderline non-significant reduction at P0 (N=5 brains). *p<.05, ** p<.01, ***p<.001, ^#p<.0001. Error bars indicate SEM.

Fig. 4. Zswim6 knockouts have a reduced number of medium spiny neurons

(A) Cell counts for the medium spiny neuron marker CTIP2 in adult Zswim6 KOs relative to WT controls shows a significant reduction of cell number in the KOs. On the other hand cell density is unchanged, consistent with cell loss combined with reduced neuropil (i.e. total dendritic length; see Fig. 5). Representative immunostaining for CTIP2 is shown on the far right, from a region of striatum indicated by the dotted yellow box. (B) In contrast to the effects on medium spiny neurons, cell counts for several populations of striatal interneurons (choline acetyl transferase, ChAT; parvalbumin, PV; somatostatin, SST) are unchanged (N=3 brains). Representative immunostaining for ChAT, SST, and PV are shown to the right, from the same region of striatum indicated by the dotted yellow box shown above in Fig. 4A. (C) Semi-quantitative RT-PCR for striatal transcripts (shown relative to control values), including the dopamine receptors Drd1 and Drd2, are not altered in the Zswim6 KOs (N=3 mice). *p<.05. Scale bars 75μm in A and B. Error bars indicate SEM.

Fig. 5. Dendritic abnormalities in striatal medium spiny neurons of Zswim6 knockouts

(A) The number of primary dendrites, dendritic branch points or nodes, dendritic ends, and total length are decreased in the KOs, whereas there is no difference in mean length. (B) Consistent with the results in (A), Sholl analysis reveals decreased cumulative length and number of intersections in the KOs. (C) Shows an example of the reconstructions from Golgi-stained sections used in these analyses. (D) Spine density is also decreased on the medium spiny neurons of Zswim6 KOs (N=3 brains per group for all Golgi analyses). Scale bars 50μm in C, 5μm in D. *p<.05, ** p<.01, ***p<.001. Error bars indicate SEM.

Fig. 6. Behavioral defects in Zswim6 knockout mice

(A) Zswim6 KOs achieve significantly lower terminal speeds and shorter times to fail (defined as falling off or making one complete backward revolution while holding on) on the rotarod test (N=14 WT, 15 KO mice). (B) In the open field the KOs have more beam breaks, more rearing events, and decreased thigmotaxis (N=16 WT, 19 KO mice). (C) Force plate analysis shows greater distance traveled, decreased low mobility bouts, increased hind limb jumps, and a significant absolute directional bias in the turning behaviors of Zswim6 KO mice referenced to the center of the actometer floor (N=9 WT, 7 KO mice). (Di) Schematic illustrating use of the force plate actometer to quantitatively measure grooming events. (Top) Low mobility bouts (LMBs) are defined as 5-second blocks where the center of gravity does not extend beyond a circle with 15mm radius. (Middle) Fast Fourier transform yields a characteristic power spectra associated with phase 4 grooming (marked by peaks at ~6Hz and 12Hz, see arrowheads). (Bottom) Nomenclature for the pattern of phase 4 grooming events. (Dii) Mutant mice spend a greater fraction of LMBs grooming but do not show alterations in the power of these grooming events (Diii). (Div) Mutant mice exhibit a shift towards greater repetitions of phase 4 grooming, as defined in Di. (E) Amphetamine (2mg/kg; administered at 30 minutes) does not alter motility of controls (dark blue and purple lines show wild-types treated with amphetamine or saline, respectively; N=12 WT-amph, 13 KO-amph, 4 WT-saline, 4 KO-saline mice per group). In contrast, 2mg/kg amphetamine significantly increases the motility of the Zswim6 KOs over their moderately elevated baseline. *p<.05, **p<.01, ***p<.001. Error bars indicate SEM.