Findings

SHANK proteins are master scaffolding proteins of the postsynaptic density (PSD) of glutamatergic synapses and are critical determinants of glutamate transmission and synaptic spine dynamics [1]. Loss of one functional copy of SHANK3 accounts for about 0.5% of the cases of autism spectrum disorder (ASD) and/or developmental delay [2], and there is likely a wider role for SHANK3 and glutamate signaling abnormalities in ASD and related neurodevelopmental disorders [3, 4]. Targeted disruption of the full-length form of Shank3 (sometimes called Shank3a) in mice leads to deficits in hippocampal AMPA signaling, long-term potentiation (LTP), and motor performance [5–7], likely reflecting delayed synaptic development as shown by the reduced AMPA signaling [5] and decreased levels of PSD-95 (unpublished results). IGF-1, which enters the central nervous system (CNS) through an interaction with lipoprotein-related receptor 1 (LRP1) [8], has multiple effects on neuronal and synaptic development and function, including effects on neurogenesis and synaptogenesis [9]. IGF-1 treatment also enhances the PSD as measured both by PSD length and by levels of PSD-95 [10, 11]. Recombinant human IGF-1 has substantial human safety data and is approved for use in children, making IGF-1 an attractive compound for evaluation in neurodevelopmental disorders.

To investigate whether IGF-1 could reverse deficits in a preclinical model of SHANK3-haploinsufficiency, we made use of a mouse with hemizygous loss of full-length Shank3 due to targeted disruption of the ankyrin repeat domain (ARD) [5]. This isoform has been directly implicated in ASD, language delay, and intellectual disability (ID), as there exist disruptive de novo point mutations in ARD in patients with ASD and ID [12, 13]. In all studies, we compared heterozygous mice with wild-type littermates using heterozygote × heterozygote mating. Consistent with previous results from our group [5], LTP induced by high-frequency stimulation was reduced in the heterozygous mice compared to wild-type littermates in the current experiments (Figures 1a and 2a) (for example, in Figure 2a, repeated measures ANOVA was used for analysis of the last five time points, F(1,6) = 33.71, P = 0.001).

Figure 1
figure 1

(1–3)IGF-1 reverses deficits in LTP and basal synaptic properties in Shank3 -deficient mice. Wild-type (WT) and heterozygous (Het) mice were treated with saline or (1–3)IGF-1 for 2 weeks before testing (injections began at postnatal day (PND) 13 to 15 and animals were analyzed immediately after the last injection). Methods for all experiments were as described previously [5], with 3 to 4 mice per group, and 1 to 2 slices per animal. (a) Hippocampal LTP was induced with high-frequency stimulation. Inset: Representative excitatory postsynaptic potential traces at 90 min after LTP induction from saline-injected (1) and (1–3)IGF-1-injected (2) heterozygous mice (scale bar: 0.5 mV, 10 ms). (b) Input–output curves comparing field excitatory postsynaptic potential (EPSP) slopes (mV/ms) as a function of stimulation intensity (mA). EPSP: excitatory postsynaptic potential; Het: heterozygous; LTP: long-term potentiation; PND: postnatal day; WT: wild-type.

Full size image
Figure 2
figure 2

IGF-1 reverses deficits in LTP, AMPA signaling, and motor function in Shank3 -deficient mice. Wild-type (WT) and heterozygous (Het) mice were treated with saline or recombinant human IGF-1 (rhIGF-1) for 2 weeks (beginning at PND 13 to 15) before testing and analyzed immediately after the last injection. Methods for all experiments were as described previously [5, 7], with 4 to 9 mice per group. (a) Hippocampal LTP was induced with high-frequency stimulation. Inset: Representative excitatory postsynaptic potential traces at 90 min after LTP induction from saline-injected (1) and rhIGF-1-injected (2) heterozygous mice (scale bar: 0.5 mV, 10 ms). (b) Slices were incubated in the presence of the N-Methyl-D-aspartate (NMDA) antagonist R-2-amino-5-phosphonopentanoate (APV) to expose AMPA receptor signaling. (c) Mice were tested for motor performance and motor learning by measuring latencies to fall off a rotating rod over three trials. Het: heterozygous; LTP: long-term potentiation; NMDA: N-Methyl-D-aspartate; rhIGF-1: recombinant human IGF-1; WT: wild-type.

Full size image

We first tested an active peptide derivative of IGF-1, (1–3)IGF-1, which has been shown to cross the blood–brain barrier and rescue Rett syndrome symptoms in Mecp2-deficient mice [11]. We observed that intraperitoneal injections at 10 μg/g/day for 2 weeks restored normal hippocampal LTP in Shank3 heterozygous mice but had no effect on wild-type mice (repeated measures ANOVA was used to analyze the last five time points, F(3,11) = 6.07, P = 0.011). In post hoc analyses, vehicle-treated heterozygous mice were significantly different from wild-type mice (P = 0.004), while (1–3)IGF-1 treated heterozygous mice were not (P = 0.66). Furthermore, peptide treatment reversed deficits in the mean slope of the input/output (I/O) function (Figure 1b) (one-way ANOVA, F(3,19) = 4.25, P = 0.02). Vehicle-treated heterozygous mice were significantly different from vehicle-treated wild-type mice (P = 0.001), while (1–3)IGF-1 treated heterozygous mice were not different from vehicle-treated wild-type (P = 0.89), and there were no significant differences between vehicle-treated wild-type mice and wild-type mice treated with IGF-1 (P = 0.812), so further studies used just three conditions.

We next administered full-length IGF-1, like that used in children with short stature due to primary IGF-1 deficiency, by intraperitoneal injection at 240 μg/kg/day, starting at PND 13 to 15 and continuing for 2 weeks (Figure 2a). This dose, chosen because it represents the maximum dose according to the current FDA label for IGF-1, was effective in rescuing deficits in LTP (repeated measures ANOVA was used to analyze the last five time points, comparing heterozygous mice with and without IGF-1, F(1,6)=28.04, P=0.002). In contrast, lower dose IGF-1 (120 μg/kg/day for 2 weeks) was associated with more modest reversal of deficits in LTP (for the last five time points: F(1,6)=2.62, P=0.012), showing a dose–response effect and providing preclinical dosing information.

Specific deficits in the glutamate AMPA receptor component of neural signaling [5] were also reversed by a 2-week treatment of 240 μg/kg/day full-length IGF-1 (Figure 2b). The mean slope of the I/O function was 0.50 ± 0.14 for wild-type, 0.34 ± 0.06 for Shank3 heterozygotes and 0.61 ± 0.059 for IGF-1 injected heterozygotes (one-way ANOVA, F(2,9) = 8.62, P = 0.008). In post hoc analyses, vehicle-treated heterozygous mice were significantly different from vehicle-treated wild-type mice (P = 0.039), while IGF-1-treated heterozygous mice were not different from vehicle-treated wild-type mice (P = 0.12).

Patients with SHANK3-haploinsufficiency frequently present with hypotonia and motor deficits of variable severity, and we have observed subtle motor deficits in Shank3-heterozygous mice [5, 7]. After treating male heterozygous mice with 240 μg/kg/day for 2 weeks, we observed enhanced motor performance following treatment (Figure 2c) (F(2,20) = 3.98, P = 0.03).

Our results provide preclinical evidence for a beneficial role for IGF-1 in SHANK3-haploinsufficiency. Moreover, as there is emerging evidence that the SHANK3 pathway and the postsynaptic density, which it helps sculpt, play a role in many neurodevelopmental disorders, as evidenced by large-scale genetic, proteomic, and gene expression studies [3, 4, 14], therapies for SHANK3 deficiency and synaptic development represent important targets that could have a widespread positive impact for neurodevelopmental disorders. The beneficial effects of IGF-1 in models of Rett syndrome [11, 15] are consistent with this hypothesis.

There are some limitations to the current study. We, and others working with similar Shank3-deficient mice, see only limited behavioral abnormalities, with none except for rotarod deficits at the ages where we carried out the IGF-1 treatments and electrophysiological studies. For this reason, the phenotypes we measure are somewhat limited. In addition, a mechanistic understanding of the neuronal effects of IGF-1 has eluded the neuroscience community and we cannot precisely explain how IGF-1 reverses the deficits observed. We do hope, however, that our findings, together with those on IGF-1 in Rett syndrome models, may help spur further research on the action of IGF-1 in the CNS. We did not see any effect produced by the (1–3)IGF-1 peptide on control animals but we did not test the effects of full-length IGF-1 on wild-type mice. There could be enhanced LTP or rotarod performance in control animals following treatment with full-length IGF-1. Many drugs have effects on both healthy and non-healthy individuals and there is hence no a priori reason to assume that IGF-1 has no effect on control animals. In fact, given the positive effects of IGF-1 in Rett syndrome models it is likely that IGF-1 has a general effect on CNS function, which might also be observed in controls.

In summary, our results show that IGF-1, approved for use in children, can lead to functional improvements in a mouse model of ASD and developmental delay, representing an important preclinical step towards novel therapeutics. Clinical trials of IGF-1 in SHANK3-deficient individuals and in ASD are now underway (ClinicalTrials.gov Identifier NCT01525901).