Adenomatous polyposis coli protein deletion leads to cognitive and autism-like disabilities

Abstract

Intellectual disabilities (IDs) and autism spectrum disorders link to human APC inactivating gene mutations. However, little is known about adenomatous polyposis coli's (APC's) role in the mammalian brain. This study is the first direct test of the impact of APC loss on central synapses, cognition and behavior. Using our newly generated APC conditional knock-out (cKO) mouse, we show that deletion of this single gene in forebrain neurons leads to a multisyndromic neurodevelopmental disorder. APC cKO mice, compared with wild-type littermates, exhibit learning and memory impairments, and autistic-like behaviors (increased repetitive behaviors, reduced social interest). To begin to elucidate neuronal changes caused by APC loss, we focused on the hippocampus, a key brain region for cognitive function. APC cKO mice display increased synaptic spine density, and altered synaptic function (increased frequency of miniature excitatory synaptic currents, modestly enhanced long-term potentiation). In addition, we found excessive β-catenin levels and associated changes in canonical Wnt target gene expression and N-cadherin synaptic adhesion complexes, including reduced levels of presenilin1. Our findings identify some novel functional and molecular changes not observed previously in other genetic mutant mouse models of co-morbid cognitive and autistic-like disabilities. This work thereby has important implications for potential therapeutic targets and the impact of their modulation. We provide new insights into molecular perturbations and cell types that are relevant to human ID and autism. In addition, our data elucidate a novel role for APC in the mammalian brain as a hub that links to and regulates synaptic adhesion and signal transduction pathways critical for normal cognition and behavior.

Figure 1

Adenomatous polyposis coli (APC) conditional knockout in mouse forebrain neurons. (a) Western blot showing APC protein levels are dramatically decreased in hippocampal, cortical and striatal, but not in cerebellar, lysates of APC conditional knock-out (cKO) mice at 3 months of age. Similar decreases were seen at 1 month (data not shown). APC deletion is driven by CAMKII promoter-dependent expression of Cre recombinase in forebrain postmitotic excitatory neurons and striatal inhibitory interneurons. (b) Histogram of decreased APC levels in the indicated brain region lysates of APC cKO mice, relative to control littermate levels. Signals are normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as a loading control (*P < 0.05, **P < 0.01, Student’s t-test, n = 5 APC cKO mice and 4 control littermates).

Figure 2

Impaired learning and memory in adenomatous polyposis coli (APC) conditional knock-out (cKO) mice. (a) Representative traces of APC cKO (red) and control littermate (blue) mice in the Barnes maze task. Control littermates showed rapid improvement and learned the location of the goal hole (green, a), using wall-mounted spatial cues, after 2 days, with 2 trials per day, as measured by (b) latency to find the goal, (c) number of errors committed before reaching the goal and (d) path efficiency. APC cKO mice showed slower improvement, (a, b) requiring 3–4 days to learn the location of the goal hole, (c) committing more errors and (d) taking less efficient paths to the hole throughout most of the test period. No differences were found in (e) average speed between the cKOs and control littermates during the task. (a, f) Probe trial on day 5 shows that both cKO mice and controls display a strong preference for the goal location, as measured by the number of visits to the goal hole. (a, g) Probe trial on day 12 shows that cKOs do not retain preference for the goal location, whereas controls do. One week after training (probe trial day 12), APC cKO mice (b) take longer time to reach the goal location, (c) commit more errors and (d) take a less efficient path, relative to control littermates, suggesting impaired long-term memory formation. *P < 0.05, Sidak-Bonferroni corrected t-test, n = 12 APC cKO mice and 14 control littermates. (h) In the continuous spontaneous alternation task to assess working memory, control littermates showed a preference to alternate, whereas APC cKOs did not; they alternated close to the chance rate of 50%. *P < 0.05, Student’s t-test, n = 7 APC cKO mice and 6 control littermates.

Figure 3

Repetitive behaviors and reduced social interest in adenomatous polyposis coli (APC) conditional knock-out (cKO) mice. (a, b) In the repetitive novel object contact task, (a) APC cKO mice repeated a particular preferred sequence of visits to the four objects more frequently, after normalizing for the total number of object visits. **P < 0.01, Student’s t-test. (b) The cKO mice preserved their direction of movement, traveling clockwise or counterclockwise, significantly more than control littermates, based on differences in the slope of log plots of the directionality of movement, normalized for the total number of object visits P = 0.001345, linear regression of difference in slopes F(1,158) = 10.6425, n = 8 APC cKO mice and 10 controls. For comparison, the plot shows the slope for total preservation of movement in one direction (top dotted line) versus totally random directional movements to visit objects (bottom dotted line). (c–f) In the social versus non-social olfaction task, both APC cKOs and control littermates displayed the ability to distinguish odors. (c) They exhibited similar levels of interest in non-social odors, such as vanilla, and showed the typical habituation pattern, spending less time sniffing the same odor over the three successive exposures (d), linear regression of differences in the slopes; vanilla, P = 0.08796, F(1,56) = 3.01581), water (data not shown), P = 0.3199 F(1,56) = 1.00705; banana, P = 0.9119, F(1,56) = 0.012352 (e) APC cKO mice, compared with control siblings, exhibited significantly less interest in social odors (novel male cage odor, *P < 0.05, Student’s t-test, n = 11 cKO mice and 9 controls). Further, the control mice showed the typical habituation pattern on the successive trials. In contrast, APC cKO mice showed less robust habituation, they spent relatively equal amounts of time investigating the social smell in each of the three trials per odor (f, linear regression of the slopes of data in e). (g) In the three-chambered social interaction assay, both cKO mice and controls displayed a preference for a caged, novel mouse (ovariectomized wild-type female), versus an empty cage, but APC cKOs spent significantly less time interacting with (sniffing) the novel mouse, suggesting reduced social interest (**P < 0.01, Sidak-Bonferroni corrected t-test). (h) APC cKO mice and control littermates made a similar number of entries to both side chambers, suggesting no difference in exploration. n = 11 cKOs and 7 controls.

Figure 4

Increased levels of β-catenin and canonical Wnt target gene expression in adenomatous polyposis coli (APC) conditional knock-out (cKO) forebrain neurons. Representative immunoblot and quantification (a) of hippocampal, cortical and striatal lysates show twofold increased levels of β-catenin in APC cKO mice. Signals were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as a loading control. *P < 0.05, ***P < 0.001 n = 5 cKO; 4 control (ctl) littermates. (b) Epifluorescence micrographs show increased β-catenin nuclear immunostaining in forebrain neurons of APC cKO mice, compared with control littermates, processed in parallel, consistent with enhanced canonical Wnt signaling. Neuronal nuclei were identified by their characteristic size and shape, and by 4′-6-diamidino-2-phenylindole (DAPI) staining. Lower panels: higher magnification views of increased nuclear β-catenin immunostaining in APC cKO cortical layer 5 pyramidal neurons (characteristic triangular shape of soma detectable because of increased cytoplasmic β-cat levels as well) relative to control littermate layer 5 neurons (red, β-catenin staining; blue, DAPI nuclear staining). (c) Histogram showing APC cKOs display increased transcript levels of the Wnt target genes: dkk1, Sp5, neurog1 and syn2 in the hippocampus (*P < 0.05, **P < 0.01, Student’s t-test; n = 5 APC cKO, 6 ctl littermate mice). (d) Immunoprecipitation of N-cadherin from the hippocampus showing increased association with β-catenin (n = 2 APC cKO, Ctl). (e) Quantitative immunoblot and histogram showing unchanged levels of N-cadherin and a significant decrease in Presenilin 1 levels (*P < 0.05, Student’s t-test, n = 5 APC cKO, 4 Ctl).

Figure 5

Adenomatous polyposis coli (APC) loss leads to increased synaptic spine density, greater frequency of miniature excitatory postsynaptic currents (mEPSCs), and modestly enhanced synaptic plasticity in pyramidal neurons. (a) Representative images showing increased spines on the apical dendrite of APC conditional knock-out (cKO) cortical layer 5 pyramidal neurons. (Left panels) Golgi-Cox stained bright-field images; (right panels) Imaris reconstructions of confocal stacks of apical dendrite (red) and spines (blue) from APC cKO-Thy-1-YFP and littermate control layer 5 neurons. (b) Histogram shows increased spine density (number of spines per unit length) on the apical dendrite of APC cKO cortical layer 5 pyramidal neurons (175.1 ± 21.7% of control (ctl) littermate levels) and hippocampal CA1 pyramidal neurons in the striatum radiatum (160.7 ± 6.9% of ctl levels; *P < 0.05, ***P < 0.001, Student’s t-test; n = 10 neurons each region, 3 ctl and 3 cKO). (c) Quantification of spine shapes from Golgi-stained hippocampal apical dendrites quantified as a percentage of total spines analyzed (***P < 0.001, Student’s t-test, n = 10 APC cKO, 10 Ctl neurons, average of 10 side plane view spines/neuron). (d) Representative traces and histogram of frequency of AMPAR-mediated mEPSCs measured by whole-cell recordings from CA1 neurons of control and APC cKO brain slices. APC cKO CA1 neurons show increased AMPAR mEPSC frequency (cKO = 0.8835 ± 0.22 Hz; ctl = 0.208 ± 0.049 Hz). (e) APC cKO mice show modestly enhanced LTP induced at SC-CA1 synapses by delivering five trains of theta burst stimulation in hippocampal slices. *P = 0.0227, repeated measure ANOVA. n = 11 cKOs and 10 ctls. AMPAR, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor.