In vitro zinc supplementation alters synaptic deficits caused by autism spectrum disorder-associated Shank2 point mutations in hippocampal neurons

Shank2 is a postsynaptic density (PSD) protein that is located at excitatory glutamatergic synapses [5, 9]. Shank2 plays a major role in synapse development, where it interacts directly with other PSD proteins to regulate surface glutamate receptors and the actin cytoskeleton [5, 14]. SHANK2 mutations have been identified in ASD patients, with SHANK2-single nucleotide variants (SNVs) detected in the SH3, PDZ, and proline-rich Shank2 domains [4, 11, 15]. Shank2 proteins have been shown to be zinc-sensitive: For example, Shank2 increases in hippocampal neurons supplemented with zinc [8], while chronic zinc deficiency induces loss of Shank2/3 and increased incidence of ASD-behaviours [8]. Supplementation of dietary zinc in Shank3^−/− mice also induces Shank2 recruitment to cortico-striatal synapses and reverses ASD-associated behaviours [6]. Transiently increasing synaptic zinc availability using clioquinol in vivo restored glutamatergic hippocampal synaptic transmission in Shank2^−/− mice through NMDA receptor activation [12]. However, zinc regulation of synaptic transmission in neurons expressing Shank2 ASD-associated SNVs has not been investigated. Here, we explore the potential of in vitro zinc supplementation in preventing synaptic deficits in hippocampal neurons expressing ASD-associated Shank2 SNVs.

Primary hippocampal dissociated cultures were prepared from male and female postnatal day zero (P0) Wistar rats as described previously [1, 2, 3, 7]. At days in vitro (DIV) 9, hippocampal cultures were transfected via calcium phosphate precipitation [7, 10, 13] with one of three ASD-associated SNVs occurring in highly conserved sites: (1) Shank2-S557N SNV (Serine to Asparagine amino acid change), identified at nucleotide position C70322303T in exon 13 of ASD patients and healthy controls, and is encoded in the Shank2 SH3 domain; (2) Shank2-V717F (Valine to Phenylalanine mutation), an SNV that occurs at nucleotide position C70026597A in exon 17, identified only in ASD patients, and is encoded in the Shank2 PDZ domain; and (3) Shank2-L1722P (Leucine to Proline mutation), presented at nucleotide position A69997007G in exon 25, identified only in ASD patients, and encoded in the Shank2 proline-rich region [11]. The Shank2-wildtype and Shank2 SNV genes (S557N, V717F, and L1722P) were encoded into a 4.7 kb pEGFP-C1 vector including a neomycin resistance cassette, SV40 early promoter, neomycin/kanamycin resistance gene of Tn5, and polyadenylation signals from the Herpes simplex virus thymidine kinase (HSV TK) gene (B.D. Bioscience Clontech; gifted by Professor Craig C. Garner). Hippocampal cultures were treated with 10 µM ZnCl₂ on DIV9 immediately after transfection for 5–7 days. A minimum of 3 independent culture preparations and transfections were examined for all data sets. Cultures were paraformaldehyde-fixed at DIV16 for immuno-staining for pre- and postsynaptic proteins using the following antibodies: anti-VGLuT1 (1:500, Neuromab N28/9, 75-066), rabbit anti-Homer (1:500, Santa Cruz Biotechnologies H-342), donkey anti-mouse IgG Alexa 594 (1:500, Molecular Probes A21203), and donkey anti-rabbit IgG Alexa 647 (1:500, Molecular Probes A31573). Z-stack images were acquired at 0.2 µm z-intervals using the Zeiss Axio Imager M2 Fluorescence Microscope at 63 × magnification with a 1.4NA oil immersion objective lens. Images were analysed using ImageJ Biophotonics (NIH, USA) to measure synaptic density defined by number of co-localised VGluT1 and Homer puncta per 10 µm length of dendrite Arons et al. [1, 2].

All Shank2 SNVs showed similar cellular expression patterns in the dendritic and spine compartments in our cultured hippocampal neurons (Fig. 1A; intensity ratios dendrite/spine: Shank2 WT 0.671 ± 0.009; S557N 0.669 ± 0.016; V717F 0.677 ± 0.014; L1722P 0.685 ± 0.018). However, the L1722P SNV displayed a significantly expression ratio in the somatic/dendritic and somatic/spine expression ratios (intensity ratios soma/dendrite: Shank2 WT 1.549 ± 0.106; S557N 1.455 ± 0.067; V717F 1.604 ± 0.082; L1722P 2.770 ± 0.170, p < 0.0001; intensity ratios soma/spine: Shank2 WT 1.058 ± 0.078; S557N 0.983 ± 0.063; V717F 1.123 ± 0.063; L1722P 2.168 ± 0.183, p < 0.0001), reflecting its high localisation in the soma (Fig. 1A), and a potential difference in the subcellular trafficking of this SNV. Our immunocytochemical analysis showed Shank2-wildtype induced a significant increase in synaptic density compared to control EGFP-transfected neurons (p-value = 0.021; Fig. 1B; Table 1). In contrast, ASD-associated Shank2 SNVs failed to induce this significant increase in synapse density suggesting that these Shank2 SNVs are loss-of-function mutations (S557N: p-value = 0.029; V717F: p-value = 0.031; L1722P: p-value = 0.0003). This is consistent with data previously reported by Leblond et al., [11] showing decreased synaptic density in ASD-Shank2 transfected neurons compared to wildtype Shank2. Interestingly, the observed Shank2 SNV-induced reduction in synaptic density did not occur in the presence of chronic zinc supplementation, suggesting that chronic zinc supplementation can rescue SNV-associated deficits in their ability to promote synapse formation and/or maintenance (Fig. 1B, Table 1). This zinc prevention of synaptic density deficits was observed for all Shank2 SNVs, i.e. S557N, V717F, and L1722P, with the greatest increase in synapse density in L1722P expressing neurons (Fig. 1B, Table 1). Chronic zinc supplementation also increased synapse density in EGFP-transfected neurons expressing endogenous levels of Shank2, however, it did not further increase synapse density in neurons overexpressing the wildtype form of Shank2.

To examine whether these structural synaptic changes translated to functional changes in synaptic transmission, whole-cell patch-clamp recordings were conducted between DIV 14–16 to measure miniaturised excitatory postsynaptic currents (mEPSCs). Cells were perfused with aCSF (in mM: 119 NaCl, 2.5 KCl, 1 Na₂HPO₄, 1.3 MgSO₄, 26.2 NaHCO₃, 11 D-( +)-glucose, 2.5 CaCl₂) containing 1 µM Tetrodotoxin (TTX) and 100 µM picrotoxin. Internal solution consisted of (in mM): 120 K gluconate, 40 HEPES, 5 MgCl₂, 2 Na₂ATP, 0.3 NaGTP, pH 7.2 with KOH, 298 mOsm, and neurons were voltage-clamped at -65 mV. Recordings with series resistance variation greater than 20% were discarded. MiniAnalysis (Synaptosoft version 6.0.7) software was used to analyse the frequency and amplitude of spontaneous AMPAR-mediated mEPSCs. Zinc-dependent effects on mEPSC frequency were measured as alterations in the time intervals between events (inter-event interval, IEI; plotted as cumulative probability graphs, where a rightward-shifted IEI represents a decrease in mEPSC frequency and leftward-shifted IEI represents an increase in mEPSC frequency).

In control media (Fig. 1C), Shank2-wildtype expressing neurons displayed a significantly leftward-shifted IEI in comparison to EGFP-transfected neurons, indicating an increase in mEPSC frequency (Fig. 1D, E). Neurons expressing ASD-associated Shank2 SNVs were unable to cause this increase in mEPSC frequency in comparison to Shank2-wildtype (Fig. 1F–H). Varying extents of the loss of function were observed between the ASD-Shank2 SNVs, with Shank-V717F expressing neurons showing higher mEPSC frequencies than EGFP expressing neurons. In contrast, the S557N and L1722P mutations reduced mEPSC frequency even lower than EGFP neurons (expressing endogenous levels of Shank2; Fig. 1F, H). This may reflect domain-specific effects of each mutation on AMPA receptor function, localisation, and/or recruitment. Interestingly, in contrast to control Shank2 and EGFP-expressing neurons, zinc supplementation increased mEPSC frequency across all ASD-associated Shank2 SNVs expressing hippocampal neurons, demonstrating that in vitro zinc supplementation was able to prevent ASD SNV-associated deficits in glutamatergic synaptic transmission. The zinc-induced rescue was most profound in L1722P-expressing neurons, but was also significant in the V717F and S557N-expressing neurons (Fig. 1F–H). Therefore regardless of the domain in which the Shank2 SNV is present, chronic zinc can alter synaptic transmission partially or completely to control levels.

AMPAR mEPSC amplitudes were also observed to be altered differentially by Shank2 SNVs (Fig. 1I, J). Specifically, increased mEPSC amplitudes were only observed in neurons expressing Shank2-S557N, a mutation which may alter SHANK-GRIP interaction in the SH3 domain to prevent AMPAR internalisation. Chronic zinc supplementation significantly decreased mEPSC amplitude in S557N and V717F expressing neurons, as was also observed in control WT-Shank2 and EGFP-expressing neurons. Together our mEPSC data show chronic zinc primarily alters mEPSC frequency in neurons expressing Shank2-SNVs. This suggests the location of zinc action may differ in inducing changes in mEPSC frequency versus amplitude, e.g. potentially postsynaptic versus presynaptic effects, and that how zinc affects mEPSC amplitude is not altered by SNVs in these two sites. We did not observe a significant decrease in mEPSC amplitude in L1722P-expressing neurons; this is likely due to the lower initial amplitude of these currents, but could also reflect this SNV occurs in the Shank2 proline-rich region which is not involved with Shank protein interaction with glutamatergic receptors.

Together our imaging and electrophysiological data reveal that ASD-associated Shank2 SNVs cause structural and functional deficits at excitatory glutamatergic hippocampal synapses. Furthermore, chronic in vitro zinc supplementation induces long-lasting forms of structural and functional plasticity that in the presence of Shank2 SNVs can prevent or lessen the ASD-associated synaptic deficits. The Shank2 variants examined can exhibit differences in the extent they contribute to the synaptic aetiology of ASD. However, although the Shank2-SNVs investigated here occur in different domains of Shank2 thereby having differential effects of synapses and overall ASD aetiology, our data show that neurons expressing these Shank2 SNVs retain their zinc responsiveness. Our imaging data support both a pre- and a postsynaptic effect of zinc supplementation. As Shank3 has a trans-synaptic strengthening effect on both the pre- and postsynapse [2], as well as playing a key role in a zinc-sensitive signalling pathway at glutamatergic synapses [6], we predict zinc supplementation activates and recruits Shank3 to synapses and promotes AMPAR recruitment as well as act trans-synaptically to enhance presynaptic function in Shank2 SNV-expressing hippocampal neurons. These data indicate that altering the neuronal zinc micro-environment can prevent synaptic deficits caused by ASD-associated genetic alterations, and further support zinc supplementation as a therapeutic target in ASD.