Isoflurane Disrupts Postsynaptic Density-95 Protein Interactions Causing Neuronal Synapse Loss and Cognitive Impairment in Juvenile Mice via Canonical NO-mediated Protein Kinase-G Signaling

Abstract

Background: Inhalational anesthetics are known to disrupt PDZ2 domain-mediated protein-protein interactions of the postsynaptic density (PSD)-95 protein. The aim of this study is to investigate the underlying mechanisms in response to early isoflurane exposure on synaptic PSD-95 PDZ2 domain disruption that altered spine densities and cognitive function. The authors hypothesized that activation of protein kinase-G by the components of nitric oxide (NO) signaling pathway constitutes a mechanism that prevents loss of early dendritic spines and synapse in neurons and cognitive impairment in mice in response to disruption of PDZ2 domain of the PSD-95 protein.

Methods: Postnatal day 7 mice were exposed to 1.5% isoflurane for 4 h or injected with 8 mg/kg active PSD-95 wild-type PDZ2 peptide or soluble guanylyl cyclase activator YC-1 along with their respective controls. Primary neurons at 7 days in vitro were exposed to isoflurane or PSD-95 wild-type PDZ2 peptide for 4 h. Coimmunoprecipitation, spine density, synapses, cyclic guanosine monophosphate-dependent protein kinase activity, and novel object recognition memory were assessed.

Results: Exposure of isoflurane or PSD-95 wild-type PDZ2 peptide relative to controls causes the following. First, there is a decrease in PSD-95 coimmunoprecipitate relative to N-methyl-d-aspartate receptor subunits NR2A and NR2B precipitate (mean ± SD [in percentage of control]: isoflurane, 54.73 ± 16.52, P = 0.001; and PSD-95 wild-type PDZ2 peptide, 51.32 ± 12.93, P = 0.001). Second, there is a loss in spine density (mean ± SD [spine density per 10 µm]: control, 5.28 ± 0.56 vs. isoflurane, 2.23 ± 0.67, P < 0.0001; and PSD-95 mutant PDZ2 peptide, 4.74 ± 0.94 vs. PSD-95 wild-type PDZ2 peptide, 1.47 ± 0.87, P < 0.001) and a decrease in synaptic puncta (mean ± SD [in percentage of control]: isoflurane, 41.1 ± 14.38, P = 0.001; and PSD-95 wild-type PDZ2 peptide, 50.49 ± 14.31, P < 0.001). NO donor or cyclic guanosine monophosphate analog prevents the spines and synapse loss and decline in the cyclic guanosine monophosphate-dependent protein kinase activity, but this prevention was blocked by soluble guanylyl cyclase or protein kinase-G inhibitors in primary neurons. Third, there were deficits in object recognition at 5 weeks (mean ± SD [recognition index]: male, control, 64.08 ± 10.57 vs. isoflurane, 48.49 ± 13.41, P = 0.001, n = 60; and female, control, 67.13 ± 11.17 vs. isoflurane, 53.76 ± 6.64, P = 0.003, n = 58). Isoflurane-induced impairment in recognition memory was preventable by the introduction of YC-1.

Conclusions: Activation of soluble guanylyl cyclase or protein kinase-G prevents isoflurane or PSD-95 wild-type PDZ2 peptide-induced loss of dendritic spines and synapse. Prevention of recognition memory with YC-1, a NO-independent activator of guanylyl cyclase, supports a role for the soluble guanylyl cyclase mediated protein kinase-G signaling in countering the effects of isoflurane-induced cognitive impairment.

Fig.1.

Disruption of NMDA receptor NR2A/2B and PSD-95PDZ2 protein interactions by isoflurane and PSD-95 wild-type PDZ2 peptide in the hippocampus of neonatal mouse brain. At postnatal day 7, mouse brains were harvested immediately after exposure to control conditions (50% oxygen: control), 1.5% isoflurane in 50% O₂ (isoflurane), 8 mg/kg inactive PSD-95 mutant PDZ2 peptide referred as (mutant PDZ2 peptide) or 8 mg/kg active PSD-95 wild-type PDZ2 peptide referred as (wild-type PDZ2 peptide), in figures. (A) Representative Western blot shows co-immunoprecipitation of PSD-95 and NMDAR2A&B subunits in control, isoflurane, PSD-95 mutant PDZ2 peptide and PSD-95 wild-type PDZ2 peptide treated groups. The panel represents equal amounts of the NMDAR2A&B immunoprecipitated from mouse brain hippocampus with NMDAR2A&B antibody. The amount of sample loaded for the input was 10% of that for the immunoprecipitation. (B) Representative densitometric analysis of PSD-95 co-immunoprecipitate relative to NMDAR2A&B precipitate. Data represent mean ± SD, n=3 independent experiments. **P < 0.01 compared to respective control. Data were analyzed using one-way ANOVA with Bonferroni’s multiple comparisons test. IP, immunoprecipitate; IB, immunoblot.

Fig.2.

Effect of isoflurane and PSD-95 wild-type PDZ2 peptide on neuronal cell viability and dendritic spines. Primary hippocampal neurons were cultured until 7 days in vitro. (A) Representative fluorescent photomicrograph (40X) with neuronal-specific marker MAP-2. DAPI counterstain was used to identify nuclei. Scale bar =5 µm. (B, C) Neuronal cells were exposed for 4 h to control conditions of 25% O₂ and 5% CO₂ that was balanced with N₂ or to 1.5% isoflurane in a gas of 25% O₂ and 5% CO₂ that was balanced with N₂. Other hippocampal neurons were exposed to 1 µM inactive PSD-95 mutant PDZ2 peptide or 1 µM active PSD-95 wild-type PDZ2 peptide for 4 h. Cell viability was assessed with the MTT assay. Viability was not significantly different in the experimental, control and untreated groups. Data represent mean ± SD, n=3 independent culture experiments and were analyzed by one-way ANOVA with Tukey’s multiple comparisons test. (D, E) Neuronal cells were co-immunostained with MAP-2 and drebrin. Representative confocal photomicrographs (63x) show dendritic segments and spines after Isoflurane (D) or PSD-95 wild-type PDZ2 peptide (E) exposures. Lower panels show (2x) zoomed image of the boxed areas in the upper images. Arrows indicate the dendritic spines. Scale bar=20 µm upper panel and 5=µm lower panel. (F, G) Summary plots represent total spine density per 10 µm dendrite length after isoflurane (F) or PSD-95 wild-type PDZ2 peptide (G) exposure in comparison with respective controls. (H, I) Summary plots represent mean spine length after isoflurane (H) or PSD-95 wild-type PDZ2 peptide (I) exposure in comparison with respective controls. Data were analyzed by independent two-tailed unpaired t-test. Data represent mean ± SD, n=5 independent experiments for spine density and spine length. *P < 0.05, *** P < 0.001, **** P <0.0001, compared with respective control.

Fig.3.

Soluble guanylate cyclase inhibitor ODQ blocked the ability of NO donor DETA NONOate to prevent isoflurane and PSD-95 wild-type PDZ2 peptide induced dendritic spine loss. (A, B) Primary neuronal cells were pre-treated with the DETA NONOate or the sGC inhibitor ODQ for 4 h, neuronal cells were exposed to the conditions shown and were co-immunostained for MAP-2/drebrin and total spine density was calculated. Representative photomicrographs (63x) show dendritic spines in the hippocampal neurons after respective treatments, scale bar = 10 µm. Arrows indicate the dendritic spines. (C, D) Summary plots represent spine density per 10 µm dendrite length. Data were analyzed using one-way ANOVA with Bonferroni’s multiple comparisons test. Data represent mean ± SD, n=4 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 compared to respective control or as indicated.

Fig.4.

Protein kinase-G inhibitor KT5823 blocked the ability of 8-Br-cGMP to prevent isoflurane- and PSD-95 wild-type PDZ2 peptide induced dendritic spine loss and decline in cyclic GMP-dependent protein kinase activity. (A, B) Primary neuronal cells were pre-treated with 8-Br-cGMP or KT5823 at the time of gas or peptides exposure and then co-immunostained for MAP-2/drebrin and total spine density was calculated. Representative photomicrographs (63x) show dendritic spines in the hippocampal neurons after respective treatments, scale bar=10 µm. Arrows indicate the dendritic spines. (C, D) Summary plots represent spine density per 10 µm dendrite length. Data were analyzed using one-way ANOVA with Bonferroni’s multiple comparisons test. Data represent mean ± SD, n=4 independent experiments. *P < 0.05, ***P < 0.001, ****P < 0.0001 compared with respective control or as indicated. Hippocampal neurons were exposed to isoflurane or PSD-95 wild-type PDZ2 peptide in the presence/absence of 8-Br-cGMP or KT5823, and cyclic GMP-dependent kinase activity was assayed. (E, F) Graphs represent the relative cyclic GMP-dependent protein kinase activity levels after respective treatments. Data were analyzed using one-way ANOVA with Bonferroni’s multiple comparisons test. Data represent mean ± SD, n=3 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 compared with respective controls or as indicated.

Fig.5

-Br-cGMP and YC-1 prevent isoflurane induced loss in synapse, blocked in presence of KT5823. Primary neuronal cells were pre-treated with 8-Br-cGMP or YC-1 or KT5823 at the time of gas exposure. (A) Representative fluorescent images of primary neurons co-immunostained at 14 days in vitro for pre- and post-synaptic markers synaptophysin (green), and PSD-95 (red), respectively, and MAP-2 (magenta). Arrows in the inset indicate the co-localized puncta of synaptophysin and PSD-95. Scale bar = 20μm (B) Co-localization between synaptophysin and PSD-95 was analyzed for quantification of synapses in neuronal cells. Graph represents mean ± SD [in percentage], n = 4 independent experiments. Data were analyzed using one-way ANOVA with Bonferroni’s multiple comparisons test. *p < 0.05 and **p < 0.01 compared with respective controls or as indicated.

Fig.6.

PSD-95 wild-type PDZ2 peptide exposure causes synapse loss which is mitigated in the presence of 8-Br-cGMP and YC-1. Primary neuronal cells were pre-treated with 8-Br-cGMP or YC-1 or KT5823 at the time of peptides exposure. (A) Representative fluorescent images of primary neurons co-immunostained at 14 days in vitro for pre- and post-synaptic markers synaptophysin (green) and PSD-95 (red), respectively, and MAP2 (magenta). Arrows in the inset indicate the co-localized puncta of synaptophysin and PSD-95. Scale bars = 20μm (B) Co-localization between synaptophysin and PSD-95 was analyzed for quantification of synapses in neuronal cells. Graph represents mean ± SD [in percentage], n = 4 independent experiments. Data were analyzed using one-way ANOVA with Bonferroni’s multiple comparisons test. *p < 0.05 and ***p < 0.001 compared with respective controls or as indicated.

Fig.7.

Neonatal exposure to isoflurane impairs recognition memory at 5 weeks in both male and female mice. At 1 week, mice were exposed to oxygen (control) or isoflurane and were injected with YC-1 or Vehicle (A, C) Plots representing time mice spent investigating novel or known objects among experimental groups. Data are plotted as; mean [seconds] ± SD for novel versus known ; male ; oxygen control +vehicle (N=15), P=0.001; oxygen control +YC-1 (N=15), P=0.001; isoflurane +YC-1 (N=15) , P=0.014; isoflurane + vehicle (N=15), P = 0.940, Female; oxygen control +vehicle (N=14), P<0.0001; oxygen control +YC-1(N=15), P=0.007; isoflurane +YC-1(N=15), P=0.009 ; isoflurane + vehicle (N=14), P = 0.091 Data were analyzed with paired-t-test. (B, D) Plot showing recognition index as % time investigating novel object/total time investigating both objects × 100; Data were analyzed using one-way ANOVA with Bonferroni’s multiple comparisons test. Mean recognition index ± SD, in male; oxygen control+vehicle versus isoflurane+vehicle, P=0.001, oxygen control+vehicle versus oxygen control+YC-1, P>0.999, oxygen control+vehicle versus isoflurane+YC-1, P=0.486 and in female; oxygen control+vehicle versus isoflurane+vehicle P=0.003, oxygen control+vehicle versus oxygen control+YC-1, P=0.667, oxygen control+vehicle versus isoflurane+YC-1, P=0.100 *P < 0.050, **P < 0.01, ****P < 0.0001 compared with respective controls or as indicated.

Fig.8.

Schematic representation illustrates the dissociation of NMDAR-PSD95-nNOS interaction induced by isoflurane or PSD-95 wild-type PDZ2 peptide. # To avoid complexity, not all PSD-95 family members are shown. Basal conditions: The NMDA receptor associates with neuronal nitric oxide synthase (nNOS) a downstream signaling molecule through PSD-95, linked through its first and second PDZ domain, PSD-95 forms a ternary complex by binding to NMDAR NR2 subunit and PDZ domain in nNOS. NMDAR-PSD-95/93-nNOS complexes helps Ca2+ influx through NMDAR on the postsynaptic membrane to activate nNOS. The NO generated is required for neuronal plasticity. It acts predominantly via cGMP and protein kinase G (PKG) for activation of kinase pathways. NO-independent pathways may also be activated by NMDAR. NO exerts significant control over gene expression.