Synaptic and cellular changes induced by the schizophrenia susceptibility gene G72 are rescued by N-acetylcysteine treatment

Abstract

Genetic studies have linked the primate-specific gene locus G72 to the development of schizophrenia and bipolar disorder. Transgenic mice carrying the entire gene locus express G72 mRNA in dentate gyrus (DG) and entorhinal cortex, causing altered electrophysiological properties of their connections. These transgenic mice exhibit behavioral alterations related to psychiatric diseases, including cognitive deficits that can be reversed by treatment with N-acetylcysteine, which was also found to be effective in human patients. Here, we show that G72 transgenic mice have larger excitatory synapses with an increased amount of N-methyl-d-aspartate (NMDA) receptors in the molecular layer of DG, compared with wild-type littermates. Furthermore, transgenic animals have lower number of dentate granule cells with a parallel, but an even stronger decrease in the number of excitatory synapses in the molecular layer. Importantly, we also show that treatment with N-acetylcysteine can effectively normalize all these changes in transgenic animals, resulting in a state similar to wild-type mice. Our results show that G72 transcripts induce robust alterations in the glutamatergic system at the synaptic level that can be rescued with N-acetylcysteine treatment.

Figure 1

G72Tg mice have larger synapses, the sizes of which are normalized by N-acetylcysteine (NAC) treatment. (a) Electron micrographs show representative glutamatergic excitatory synapses sampled from the outer two-thirds of the dentate gyrus molecular layer from wild-type (WT, a'), G72 transgenic (G72Tg, a'') and NAC-treated G72 transgenic (G72TgN, a''') animals (synaptic active zone edges are indicated with arrows). Scale bar: 200 nm for all images. (b,c) Dots represent the median values of 3D-reconstructed synaptic areas in individual animals, and whiskers mark 40–60 percentile boundaries. Effects were tested in five littermate pairs (b, their values are connected) and in three littermate pairs (c), and then the G72 effect was reversed by the NAC treatment in G72TgN littermates (c; *P<0.05). Numbers indicate litters. Litter number 1 (L#1): WT: n=32 3D-reconstructed synapses, G72Tg: n=24, P=0.013; L#2: (n=52, 38), P=0.024; L#3: (n=43, 28), P=0.014; L#4: (n=24, 14), P=0.029; L#5: (n=30, 59), P=0.003; L#6: WT (n=62), G72Tg (n=48), G72TgN (n=87), WT versus G72Tg P=0.042, G72Tg versus G72TgN P<0.001; L#7: (n=69, 49, 51), WT versus G72Tg P<0.001, G72Tg versus G72TgN P=0.004; L#8: (n=60, 56, 78), WT versus G72Tg P=0.046, G72Tg versus G72TgN P=0.031.

Figure 2

Correlation between synaptic N-methyl-d-aspartate (NMDA) receptor content and synaptic area is unaffected by genotype and N-acetylcysteine (NAC) treatment. (a) Electron micrographs show representative glutamatergic synapses sampled from the outer two-thirds of the dentate gyrus molecular layer from wild-type (WT, a'), G72 transgenic (G72Tg, a'') and NAC-treated G72 transgenic (G72TgN, a''') animals (synaptic active zone edges are indicated with white bars). Post-embedding immunogold labeling against GluN1 NMDA receptor subunit (arrows). Scale bar: 150 nm for all images. (b) Scatterplots show significant (P<0.01) correlations between synaptic immunogold particle number and synaptic active zone area in WT (Spearman R=0.404, n=94 3D-reconstructed synapses), G72Tg (R=0.438, n=118), G72TgN (R=0.346, n=129) mice, as well as when data were pooled (R=0.385, n=341). Animals used for these measurements were taken from litters L#3, L#4, L#6, L#7 and L#8.

Figure 3

Synaptic cleft width and postsynaptic density (PSD) thickness are unchanged in G72 transgenic animals. (a) Electron micrograph of a glutamatergic synapse sampled for measurement. Scale bar: 150 nm. (b) Area selected for measurements (white box in a). (c) Line graph shows vertically summed optical density values of the synapse shown in b. (d) Synaptic cleft width (P=0.996) and PSD thickness (P=0.711) did not differ between wild type (WT; n=44 synapses from two mice) and G72Tg (n=36 synapses from two mice) mice. The median values and interquartile ranges are shown. Animals used for these measurements were taken from litters L#3 and L#4.

Figure 4

The number of synapses and granule cells decreased in G72Tg mice, and it was normalized by N-acetylcysteine (NAC) treatment. (a) Light micrograph shows a hippocampal section before measurements. White dashed line marks the molecular and granule cell layer border of the dentate gyrus, stripes indicate the outer two-thirds of the molecular layer, from where the samples were collected. (b) Electron micrograph shows an image from a series used for physical disector method. Green/red lines show the stereological counting frame (5 × 5 μm). (c) The number of synapses in the outer two-thirds of the molecular layer is decreased in G72Tg mice. This effect is reversed by NAC treatment in G72TgN mice. Dots represent individual measurements (n=18 in each group), lines represent the mean values (WT: 41.98 × 10⁸, s.d.=±7.77 × 10⁸; G72Tg: 32.38 × 10⁸, s.d.=±1.07 × 10⁹; G72TgN: 41.04 × 10⁸, s.d.=±1.08 × 10⁹). WT versus G72Tg P=0.004, G72Tg versus G72TgN P=0.021. Total number of synapses counted during stereological measurements: WT: 444 synapses, G72Tg: 327 synapses, G72TgN: 417 synapses. (d) Light micrograph shows the dentate gyrus. White dashed line marks the granule cell layer (the sampling area). (e) Light micrograph from optical fractionator measurement. Green/red lines: stereological counting frame (7 × 7 μm). Hematoxylin-labeled nuclei of granule cells (purple) were counted. (f) Number of granule cells is decreased in G72Tg mice, which is reversed by NAC treatment (*, Friedman test P<0.05). Note that the optical fractionator method used in this experiment results in a single final value for every individual animal investigated (dots). Values connected by lines correspond to littermates. Numbers indicate litters. Litter number 6 (L#6): WT: number of cells counted during measurement: n=380, G72Tg (n=349), G72TgN (n=374); L#7: n=369, 335, 370; L#8 n=392, 344, 352; L#9: n=297, 324, 348. Data are given for one brain hemisphere in all cases.