Synaptic dysfunction and abnormal behaviors in mice lacking major isoforms of Shank3

Abstract

SHANK3 is a synaptic scaffolding protein enriched in the postsynaptic density (PSD) of excitatory synapses. Small microdeletions and point mutations in SHANK3 have been identified in a small subgroup of individuals with autism spectrum disorder (ASD) and intellectual disability. SHANK3 also plays a key role in the chromosome 22q13.3 microdeletion syndrome (Phelan-McDermid syndrome), which includes ASD and cognitive dysfunction as major clinical features. To evaluate the role of Shank3 in vivo, we disrupted major isoforms of the gene in mice by deleting exons 4-9. Isoform-specific Shank3(e4-9) homozygous mutant mice display abnormal social behaviors, communication patterns, repetitive behaviors and learning and memory. Shank3(e4-9) male mice display more severe impairments than females in motor coordination. Shank3(e4-9) mice have reduced levels of Homer1b/c, GKAP and GluA1 at the PSD, and show attenuated activity-dependent redistribution of GluA1-containing AMPA receptors. Subtle morphological alterations in dendritic spines are also observed. Although synaptic transmission is normal in CA1 hippocampus, long-term potentiation is deficient in Shank3(e4-9) mice. We conclude that loss of major Shank3 species produces biochemical, cellular and morphological changes, leading to behavioral abnormalities in mice that bear similarities to human ASD patients with SHANK3 mutations.

Figure 1.

Generation and characterization of Shank3^e4–9 mice. (A) A diagram of the structure of the murine Shank3 gene showing promoters and alternative splicing sites (the asterisk indicates alternatively spliced exons). Five intragenic promoters were identified (see also Supplementary Material, Fig. S1B): ANK, ankyrin repeats; SH3, Src homology 3 domain; PDZ, postsynaptic density protein, Drosophila disk large tumor suppressor (DlgA) and Zonula occludens-1 protein (Zo-1) domain proline-rich domain; SAM, sterile α-motif domain. (B) Genomic map and structure of the Shank3 gene covering exons 1–10 and the targeting construct for the deletion of Shank3 exons 4–9. RT-F = forward primers and RT-R = reverse primer for RT-PCR analysis in Supplementary Material, Fig. S1A4. (C) Genotyping of Shank3^e4–9 mice by Southern blot with XbaI and the 3′ flanking probe. The 10.5 kb fragment is the wild-type band; the 6.4 kb fragment is the mutant band. (D) Western blot analysis with Shank3^e4–9 (–/–) brain samples; the 190, 170 and 140 kDa bands are seen in the Shank3^+/+ (+/+) brain and only the 140 and 170 kDa bands are in Shank3^e4–9 samples. (E) RT-PCR analysis for specific Shank3 transcripts in Shank3^e4–9 mice. Shank3a and Shank3b transcripts were absent in Shank3^e4–9 mice. Other transcripts (Shank3c–e) from promoters 3–5 were present in Shank3^e4–9 mice.

Figure 2.

Sociability and interaction as well as ultrasonic communications are abnormal in Shank3^e4–9 mice. (A and B) No genotype or sex differences are evident for preferences between the identical NS objects. In the NS1–Soc1 pairing, Shank3^+/+ (+/+) males and females prefer the novel Soc stimulus; Shank3^e4–9 (–/–) males and females show no or reduced preferences for this stimulus. When presented with familiar and novel Soc stimuli, all mice prefer the novel social partner. (C) In the dyadic test, Shank3^e4–9 mice participate in bidirectional social exchanges with C3H partners for shorter periods of time than Shank3^+/+ controls and their partners. (D) Shank3^e4–9 mice take longer to initiate their first social interaction than Shank3^+/+ mice. (E and F) Both male (E) and female (F) Shank3^e4–9 mice and their respective C3H partners show reduced times in social interactions than the respective Shank3^+/+ targets and their C3H partners; n = 10 mice/genotype/sex. *P< 0.05, Shank3^+/+ versus Shank3^e4–9 mice; #P< 0.05, females versus males within genotype; ‡P< 0.05, compared with the NS1–NS1 test within genotype; ¤P< 0.05, compared with the NS1–Soc1 test within genotype; &P< 0.05, target versus partner within genotype; ^vP< 0.05, Shank3^+/+ social partner versus Shank3^e4–9 partner (see Supplementary Material, Fig. S2 and Table S1). (G) Shank3^e4–9 males emit more and Shanke^e4–9 females produce fewer calls relative to Shank3^+/+ controls. (H) Representative spectrographs of ultrasonic calls for male and female Shank3^+/+ and Shank3^e4–9 mice. (I and J) The calls emitted by Shank3^+/+ mice are primarily of long duration, whereas those by Shank3^e4–9 mice are mostly of short duration. (K and L) Shank3^+/+ males produce more mid-range and high-frequency calls of longer duration than Shank3^e4–9 males. Durations did not vary across frequencies for Shank3^e4–9 mice; however, durations of high-frequency calls by Shank3^e4–9 females were shorter than those by other groups; n = 10 mice/genotype/sex. *P< 0.05, Shank3^+/+ versus Shank3^e4–9 mice; ^#P< 0.05, females versus males within genotype; ^!P< 0.05, compared with short duration or low-frequency calls; =P< 0.05, compared with moderate duration or mid-range frequency calls.

Figure 3.

Shank3^e4–9 mice display abnormalities in motor performance. (A) Shank3^e4–9(–/–) mice make more foot-misplacements than Shank3^+/+ (+/+) controls. (B) Shank3^e4–9 mice move more slowly than Shank3^+/+ mice; Shank3^e4–9 males are slower than mutant females. (C) Spontaneous locomotor activity in the open field is lower in Shank3^e4–9 males than in other groups. (D and E) Motor learning on the accelerating rotorod is deficient in Shank3^e4–9 males compared with Shank3^+/+ males over all five trials and it is perturbed on trials 4 and 5 compared with Shank3^+/+ and Shank3^e4–9 females. (F) Repetitive frequencies of object explorations from inside and outside the nest were higher for Shank3^e4–9 than Shank3^+/+ mice. (G) Shank3^+/+ males and females typically left the nest to explore the novel object. Shank3^e4–9 males spent similar amounts of time exploring the objects from inside and outside the nest than Shank3^+/+ males; Shank3^e4–9 females spent the most time exploring objects from the nest. (H) Shank3^e4–9 spent more time self-grooming than Shank3^+/+ controls; n = 10 mice/genotype/sex. *P< 0.05, Shank3^+/+ versus Shank3^e4–9 mice; ^#P< 0.05, females versus males within genotype; ⁺P< 0.05, Shank3^+/+ females versus Shank3^e4–9 males; ^†P< 0.05, in nest compared with out of the nest.

Figure 4.

Shank3^e4–9 mice are deficient in learning and memory. (A) To reach the hidden platform in water maze, Shank3^e4–9 mice swam over longer distances on the first 3 days of acquisition testing than Shank3^+/+ mice. When the hidden platform was moved to a new location, swim distance was increased on days 9–11 for Shank3^e4–9 mice. (B and C) Visible platform testing with naïve (left) and water-maze experienced (Exp.) (right) Shank3^+/+ and Shank3^e4–9 mice. Naïve Shank3^e4–9 males (B, left) and females (C, left) swam over longer distances to reach the visible platform than respective Shank3^+/+ controls. Swim distances for experienced Shank3^e4–9 males (B, right) and females (C, right) were similar to those of the respective Shank3^+/+ controls. (D and E) Swim distances for Shank3^+/+ (D) and Shank3^e4–9 (E) mice during acquisition (days 2, 4 and 6) and reversal probe tests (days 8, 10 and 12). During acquisition, all mice showed a marked preference for the NE quadrant. At reversal, Shank3^+/+ mice, but not Shank3^e4–9 mice, developed a significant preference for the SW quadrant. (F) Shank3 mice were evaluated in the novel object recognition test consisting of training (Train), and tests for STM, LTM and RM. During training, neither genotype showed any preference for either of the identical objects. Shank3^+/+ mice preferred the novel object on the STM, LTM and RM tests; preferences were reduced for Shank3^e4–9 mice. (G) Total numbers of object contacts were similar between genotypes. Object contacts declined for Shank3^+/+ mice over LTM and RM testing; contacts remained high for Shank3^e4–9 mice. (H) In the STFP test, no genotype differences were noted for STM. Shank3^e4–9 mice showed a dramatically reduced preference for the demonstrator diet during the LTM and RM tests. For all tests, n = 10 mice/genotype/sex. *P< 0.05, Shank3^+/+ versus Shank3^e4–9 mice; ^#P< 0.05, females versus males within genotype; ^P< 0.05, naïve Shank3^e4–9 mice versus experienced Shank3^e4–9 mice; ^§P< 0.05, versus NE quadrant; ^XP< 0.05, versus SW quadrant; ^†P< 0.05, compared with the STM test.

Figure 5.

Dendritic spine morphology was altered in Shank3^e4–9 mice. (A) Low (left) and high (right) resolution images of EGFP-expressing pyramidal neurons from Shank3^+/+ (+/+) and Shank3^e4–9 (–/–) mice. Quantification of spine density (B), spine length (C) and spine-head area (D). A significant difference was revealed in spine length but not for spine density or spine-head area (spine length: 1.23 ± 0.07 μm, 678 spines from 7 Shank3^e4–9 mice; and 0.76 ± 0.06 μm, 1183 spines from 8 Shank3^+/+ mice; *P< 0.002, Shank3^+/+ versus Shank3^e4–9 mice). (E and F) Representative electron micrographs of hippocampal CA1 striatum radiatum synapses from Shank3^+/+and Shank3^e4–9 mice. The PSD is visible as an electron-dense layer adjacent to the post-SPM (scale bar = 0.1 μm). A total of 71 synapses from 3 Shank3^+/+ mice and 69 synapses from 3 Shank3^e4–9 mice were measured. There were no significant genotype changes in PSD length or thickness. (G and H) Representative images from Golgi staining of 4-week (G) and 10-week-old (H) Shank3^+/+ and Shank3^e4–9 mice. Scale bar = 5 μm. (I and J) Spine density (I) and spine length (J) in 4-week-old Shank3^e4–9 mice (n = 15 cells from 3 +/+ mice and n = 14 cells from 3 –/– mice). A total of 731 spines from Shank3^+/+ and 649 spines from Shank3^e4–9 mice were measured (*P= 0.03 for spine density and P< 0.003 for spine length). (K and L) Spine density (K) and spine length (L) in 10-week-old Shank3^e4–9 mice (n = 40 cells from 3 +/+ mice; n = 31 cells from 3 –/– mice). A total of 992 spines from Shank3^+/+ and 707 spines from Shank3^e4–9 mice were measured (*P= 0.01 for spine length).

Figure 6.

Altered protein composition in PSD and SPM fractions of Shank3^e4–9 mice. Immunoblot analyses of PSD (A) and SPM (B) fractions from individual Shank3^+/+ and Shank3^e4–9 mice for the indicated proteins. Pan-Shank (Shank1–3) antisera revealed an absence of the Shank3 band in Shank3^e4–9 mice. Homer1b/c and GKAP in the PSD and GluA1 and NR2A in the SPM are reduced in Shank3^e4–9 mice. (C and D) Quantification of proteins in PSD fractions based on results shown in (A) and (B), respectively, normalized using actin for protein quantification. There is a significant reduction in the levels of the indicated proteins in Shank3^e4–9 compared with Shank3^+/+ samples. Homer1b/c (n = 16 each/genotype, *P< 0.001), GKAP (n = 11 each/genotype,*P= 0.03), GluA1 (n = 8 each/genotype,*P< 0.002) and NR2A (n = 8 each/genotype, *P= 0.02). No sex difference was observed. There is no significant differences for other proteins including Homer1a (n = 11 each/genotype), GluA2 (n = 8 each/genotype) and NR2B (n = 8 each/genotype). (E) Hippocampal neurons in dissociated culture (16–18DIV) from Shank3^+/+ or Shank3^e4–9 mice (n = 8 each/genotype) were immunostained with bassoon antibody (red) and with GKAP, Homer1b/c, GluA1(C-terminus) or NR2A antibodies (green). Merged images shown on the right. Scale bar = 5 µm. (F) Normalized integrated density for the indicated proteins (Shank3^+/+= 1.00). Significant differences were found for Homer1b/c and GluA1 proteins between Shank3^+/+ and Shank3^e4–9 samples (*P< 0.01 for Homer1b/c and *P< 0.001 for GluA1). No significant differences were found for GKAP and NR2A.

Figure 7.

Impaired synaptic plasticity and activity-dependent AMPAR GluA1 distribution in Shank3^e4–9 mice. (A) Summary graph of LTP experiments and representative fPSP traces in Shank3^+/+ (+/+) and Shank3^e4–9 (–/–) mice (+/+, n = 12 slices from 8 mice; –/–, n = 8 slices from 8 mice). Tetanic stimulation (100 Hz, 1 s × 2) was applied at 15 min. LTP was significantly impaired in Shank3^e4–9 mice. (B) Input–output relationships of fPSPs of Shank3^+/+ (+/+, n = 23 slices from 8 mice) and Shank3^e4–9 mice (–/–, n = 23 slices from 8 mice) were not significantly different. (C) Summary graph of the fiber volley (FV) amplitude during the input–output test of Shank3^+/+ (+/+, n = 8 slices from 8 mice) and Shank3^e4–9 mice (–/–, n = 9 slices from 8 mice). Insert depicts FV versus fPSP relationship. (D) Paired-pulse ratio at different inter-stimulus intervals in Shank3^+/+ (+/+, n = 13 slices from 8 mice) and Shank3^e4–9 mice (–/–, n = 14 slices from 8 mice). (E) Sample traces and summary graphs depicting mIPSC amplitude and frequency in Shank3^+/+ (n = 13 cells from 5 mice) and Shank3^e4–9 (n = 14 cells from 5 mice) animals. No significant differences were observed between genotypes. (F) Sample traces and summary graphs depicting mEPSC amplitude and frequency in Shank3^+/+ (n = 14 cells from 5 mice) and Shank3^e4–9 (n = 14 cells from 5 mice) animals. No significant differences were observed between genotypes. (G) Attenuated response of activity-dependent distribution of surface GluA1. Hippocampal neurons in dissociated culture (16–18DIV) from Shank3^+/+ and Shank3^e4–9 mice were treated with a chemical LTP protocol. Neurons were sequentially stained for surface GluA1 (N-terminus) (green) and bassoon (red) under non-permeant and permeant conditions, respectively. A significant increase of surface GluA1 staining after cLTP was seen in Shank3^+/+ neurons; this increase was much attenuated in Shank3^e4–9 neurons. (H) The normalized integrated density of surface GluA1. Staining intensity was significantly increased after cLTP in cultured neurons from Shank3^+/+, but not from Shank3^e4–9 mice. Shank3^+/+, n = 17 cells; Shank3^e4–9, n = 12 cells. *P <0.001, Shank3^+/+ versus Shank3^e4–9 mice. ^P< 0.01, cLTP-stimulated Shank3^+/+ versus Shank3^e4–9 mice.