In vitro ischemia triggers a transcriptional response to down-regulate synaptic proteins in hippocampal neurons

Abstract

Transient global cerebral ischemia induces profound changes in the transcriptome of brain cells, which is partially associated with the induction or repression of genes that influence the ischemic response. However, the mechanisms responsible for the selective vulnerability of hippocampal neurons to global ischemia remain to be clarified. To identify molecular changes elicited by ischemic insults, we subjected hippocampal primary cultures to oxygen-glucose deprivation (OGD), an in vitro model for global ischemia that resulted in delayed neuronal death with an excitotoxic component. To investigate changes in the transcriptome of hippocampal neurons submitted to OGD, total RNA was extracted at early (7 h) and delayed (24 h) time points after OGD and used in a whole-genome RNA microarray. We observed that at 7 h after OGD there was a general repression of genes, whereas at 24 h there was a general induction of gene expression. Genes related with functions such as transcription and RNA biosynthesis were highly regulated at both periods of incubation after OGD, confirming that the response to ischemia is a dynamic and coordinated process. Our analysis showed that genes for synaptic proteins, such as those encoding for PICK1, GRIP1, TARPγ3, calsyntenin-2/3, SAPAP2 and SNAP-25, were down-regulated after OGD. Additionally, OGD decreased the mRNA and protein expression levels of the GluA1 AMPA receptor subunit as well as the GluN2A and GluN2B subunits of NMDA receptors, but increased the mRNA expression of the GluN3A subunit, thus altering the composition of ionotropic glutamate receptors in hippocampal neurons. Together, our results present the expression profile elicited by in vitro ischemia in hippocampal neurons, and indicate that OGD activates a transcriptional program leading to down-regulation in the expression of genes coding for synaptic proteins, suggesting that the synaptic proteome may change after ischemia.

Figure 1. OGD induces delayed neuronal death of mature hippocampal neurons in culture.

(A) OGD causes hippocampal neuronal cell death, as determined by analysis of nuclear morphology. After incubation under OGD conditions for 1 h (n = 3), 1 h30 (n = 12) and 2 h (n = 10), cells were returned to the 5% CO₂ incubator for 24 h. Cell viability was then assessed by analysis of the nuclear morphology. Pyknotic nuclei (arrows) were counted as dead cells. The results were expressed as the percentage of dead cells relatively to the total cell number. The right panel depicts nuclear morphology of neurons subjected to control and 2 h OGD. (B) Time-course of OGD-induced neuronal death, as determined by LDH release. Cells were subjected to OGD for 2 h and the LDH release was assessed 0 h (n = 6), 4 h (n = 5), 7 h (n = 3), 14 h (n = 6), 18 h (n = 6) and 24 h (n = 13) after the stimulus. (C–D) OGD induces cleavage of spectrin and the formation of spectrin breakdown products (SBDPs). SBDPs protein levels were analyzed by Western blot 7 h (C) and 24 h (D) after 2 h of OGD. The calpain inhibitor MDL 28170 (50 µM) was added 30 min prior to stimulation and kept during the stimulation and post-stimulation period (24 h). Actin was used as loading control. Bars represent the mean ± SEM of five (C) or ten (D) independent experiments performed in distinct preparations. *p<0.05, ***p<0.001, as determined by the Student's t-test (C) and One-way ANOVA followed by Dunn's Multiple Comparison Test (D; *p<0.05 for the OGD condition compared to control, ##p<0.01 for the OGD+MDL28170 condition compared to OGD).

Figure 2. Inhibition of glutamate receptors protects hippocampal neurons against OGD-induced cell death.

Mature hippocampal neurons were subjected to 2-stimulation periods. Cell viability was assessed 24 h after the stimulus by analysis of the nuclear morphology (A–C) and by determination of the LDH release (D–F). Bars represent the mean ± SEM of 4–9 independent experiments, performed in distinct preparations. Statistical analysis was performed using One-way ANOVA followed by Bonferroni's Multiple Comparison Test: *p<0.05, **p<0.01, ***p<0.001, relative to control; ^# p<0.05, ^## p<0.01, ^### p<0.001 relative to OGD condition. MK-801, selective NMDAR antagonist; GYKI 52466, selective AMPAR antagonist; Naspm, selective CP-AMPAR antagonist.

Figure 3. Summary of gene expression changes at 7 h and 24 h after OGD.

(A) Student's t-test analysis was applied to the microarray data to identify all genes whose expression was significantly different between conditions (p<0.05). Up-regulated and down-regulated genes include those whose expression levels had a fold change ≥2.0. (B and C) Number of up-regulated (B) or down-regulated (C) transcripts at 7 h and 24 h after 2 h OGD. The intersection represents the number of transcripts whose transcription was changed in both recovery periods. The transcripts used in this analysis were considered differentially expressed after using two cut-off criteria: a p-value <0.05 and a fold change of 2.0. The VENNY informatic tool was used to compare the lists of transcripts to obtain the Venn Diagrams.

Figure 4. Time course analysis of OGD-induced differential gene expression within each functional gene category.

The number of genes that were up-regulated (black bars) or down-regulated (white bars) at each period of recovery after the OGD insult is plotted for each functional gene category. Gene ontology analyses were performed using GoMiner and functional groups were selected manually.

Figure 5. Ontology of genes differentially expressed at 7 h and 24 h of incubation after 2 h of OGD.

Gene ontology analyses included genes that had a p-value <0.05 and a fold change of 2.0 and were performed using GoMiner. Classes were selected manually and the number of genes for each class divided by the sum of the total number of genes in the selected classes (586 genes at 7 h and 403 genes at 24 h). Note that some genes are included in more than one class. (A) Ontology of genes up-regulated (upper) and down-regulated (lower) relatively to the control at 7 h of incubation in culture conditioned medium after 2 h OGD. In this analysis, the total numbers of up-regulated and down-regulated genes were 232 and 354, respectively. (B) Ontology of genes up-regulated (upper) and down-regulated (lower) relatively to the control at 24 h of incubation in culture conditioned medium after 2 h OGD. In this analysis, the total numbers of up-regulated and down-regulated genes were 333 and 70, respectively. (C) Ontology of genes up-regulated (upper) and down-regulated (lower) at 7 h and 24 h of incubation in culture conditioned medium after 2 h OGD. In this analysis, the total numbers of up-regulated and down-regulated genes were 60 and 31, respectively.

Figure 6. OGD induces changes in the mRNA levels of transcripts encoding synaptic proteins.

Total RNA was extracted with TriZol 7 µg of total RNA and specific primers for each selected gene. Fold change in mRNA levels was normalized to Gapdh and Actb. Quantitative PCR analysis showed that genes encoding proteins associated with AMPAR trafficking (A) and pre- and post-synaptic compartments (B), as well as subunits of the AMPA and NMDA receptors (C) were mostly down-regulated (with the exception of Sypl2, which had increased expression levels) after OGD, at least at one of the time points analyzed after OGD. Bars represent the mean ± SEM of 5 independent experiments, performed in distinct preparations. *p<0.05, **p<0.01, ***p<0.001, as determined by the Student's t-test on log-transformed data.

Figure 7. OGD increases REST expression in mature hippocampal neurons.

(A) Quantitative PCR analysis showed the OGD insult induced a marked increase in Nrse (Rest) mRNA. Total RNA was extracted with TriZol 7 h and 24 h after the OGD insult. Quantitative PCR analysis was performed using cDNA prepared from 1 µg of total RNA and specific primers for each selected gene. Fold change in mRNA levels was normalized to Gapdh and Actb. Bars represent the mean ± SEM of three independent experiments, performed in distinct preparations. *Significantly different from control (*p<0.05, Student's t-test on log-transformed data). (B) Representative Western blot shows a marked increase in REST protein levels, both in the cytoplasmic and nuclear fractions of hippocampal neurons submitted to OGD followed by 24 h of incubation in culture conditioned medium (n = 3). Actin was used as loading control. (C) Putative RE-1 sequence(s)/REST-binding site(s) in synaptic genes (according to [35]) found to be down-regulated after OGD. (D) Genes with enrichment of REST after ischemia (according to [33]) found to be differently expressed after OGD.

Figure 8. Ischemic insults affect the protein levels of the AMPAR subunits GluA1 and GluA2.

(A) Total protein extracts were prepared 7 h and 24 h after the OGD insult and Western blot analysis were performed. Total GluA1 is decreased after 24 h, whereas GluA2 levels remain unaltered following both periods of recovery. The left panel shows a representative Western blot for total GluA1 and GluA2 present in cell lysates after OGD. The right panel represents the quantification of the Western blots. Bars represent the mean ± SEM of five independent experiments, performed in ditinct preparations. (B) GluA1 and GluA2 surface levels were analyzed after biotinylation of cultured hippocampal neurons, 24 h after the OGD insult. At this time point after OGD GluA1 is removed from the surface, whereas GluA2 surface levels remain unaltered. The left panel shows a representative Western blot for surface GluA1 and GluA2 after the insult. Total actin (from the total extract) was used as loading control. Bars represent the mean ± SEM of 3–4 independent experiments, performed in distinct preparations. *p<0.05, as determined by the Student's t-test.

Figure 9. OGD affects the expression levels of NMDAR subunits.

(A) Total protein extracts were collected 7 h and 24 h after OGD and Western blot analysis was performed. Whereas total GluN1 levels tend to increase after the insult, a decrease in total GluN2A was observed after both periods of recovery while GluN2B decreased at 24 h after the OGD insult. The left panel shows a representative western blot for GluN1, GluN2A and GluN2B levels after OGD. The right panel represents the quantification of the Western blots. Actin was used as loading control. Bars represent the mean ± SEM of 5–9 independent experiments, performed in distinct preparations. (B) OGD induces a marked increase in the mRNA levels of GluN3A but not in GluN3B. Bars represent the mean ± SEM of 3–5 independent experiments, performed in distinct preparations. *p<0.05, as determined by the Student's t-test.