Lack of nAChR activity depresses cochlear maturation and up-regulates GABA system components: temporal profiling of gene expression in alpha9 null mice

Abstract

Background: It has previously been shown that deletion of chrna9, the gene encoding the alpha9 nicotinic acetylcholine receptor (nAChR) subunit, results in abnormal synaptic terminal structure. Additionally, all nAChR-mediated cochlear activity is lost, as characterized by a failure of the descending efferent system to suppress cochlear responses to sound. In an effort to characterize the molecular mechanisms underlying the structural and functional consequences following loss of alpha9 subunit expression, we performed whole-transcriptome gene expression analyses on cochleae of wild type and alpha9 knockout (alpha9(-/-)) mice during postnatal days spanning critical periods of synapse formation and maturation.

Principal findings: Data revealed that loss of alpha9 receptor subunit expression leads to an up-regulation of genes involved in synaptic transmission and ion channel activity. Unexpectedly, loss of alpha9 receptor subunit expression also resulted in an increased expression of genes encoding GABA receptor subunits and the GABA synthetic enzyme, glutamic acid decarboxylase. These data suggest the existence of a previously unrecognized association between the nicotinic cholinergic and GABAergic systems in the cochlea. Computational analyses have highlighted differential expression of several gene sets upon loss of nicotinic cholinergic activity in the cochlea. Time-series analysis of whole transcriptome patterns, represented as self-organizing maps, revealed a disparate pattern of gene expression between alpha9(-/-) and wild type cochleae at the onset of hearing (P13), with knockout samples resembling immature postnatal ages.

Conclusions: We have taken a systems biology approach to provide insight into molecular programs influenced by the loss of nicotinic receptor-based cholinergic activity in the cochlea and to identify candidate genes that may be involved in nicotinic cholinergic synapse formation, stabilization or function within the inner ear. Additionally, our data indicate a change in the GABAergic system upon loss of alpha9 nicotinic receptor subunit within the cochlea.

Figure 1. Overview of differentially expressed genes in the α9^−/− compared to wild type.

(A) The ages indicate the postnatal (P) days at which cochlear RNA samples were prepared. Efferent innervation in the cochlea peaks around P7 when most of the efferent axons are contacting inner hair cells. Between P7 and P13, the axons are pruned and retargeted to outer hair cells. P13 also marks the initial phase of auditory function stabilization in rodents. Microarray analyses were based on comparative analyses of α9^−/− and wild type at P3, P7, P13 and P60. (B) Comparison of expression levels of all probes between α9^−/− samples and wild type controls at each age. Normalized gene expression data are plotted on a logarithmic scale (log₂), and genes changing in the α9^−/− by at least 1.5 fold and differentially expressed using the limma package in R (FDR q-value <5%) are highlighted for each age. Red dots correspond to significant up-regulation and green to significant down-regulation in the knockout animals compared to wild type controls. Numbers of up- and down-regulated genes are also noted in the plots.

Figure 2. Functional categories of genes significantly over-represented in α9^−/−.

(A) Venn diagram depicting the overlap among differentially expressed genes (DEGs) between P13 and P60. The intersection of differentially expressed genes between wild type and α9^−/− consists of 142 genes. (B) Over-represented classes among these 142 genes are shown (DAVID, BH-adjusted p-value <0.05). (C) Comparative enrichment of functional categories (Ingenuity database; −log₁₀ (enrichment p-value)) between P13 and P60 are indicated; larger values indicate higher significance. The dotted line at 1.3 corresponds to p-values of 0.05 on this scale.

Figure 3. Confirmation of selected genes at the transcript level.

(A) Log₂ fold changes in expression levels of selected transcripts measured by qPCR are plotted alongside their fold changes obtained by microarray analysis. For the ten selected genes, expression values from microarray and qPCR showed similar relative changes (p<0.05, two-tailed t-test). Error bars indicate SEM.

Figure 4. Quantitative protein expression analysis of select GABA receptor subunits and cellular localization of changes to GABA_Aβ2 expression.

(A) Protein expression of GABA_A subunits was assessed by immunoprecipitation (IP) to confirm changes observed at a transcriptional level. 200 µg cochlear lysates from wild type and α9^−/− were subjected to IP with antibodies against GABA_Aα1 and GABA_Aβ2 subunits. The immunoprecipitates were analysed by western blotting with anti-GABA_Aα1 or anti- GABA_Aβ2. A separate western blot was run in parallel using the same lysates used for IP experiments for wild type and α9^−/− samples. A representative western blot including GAPDH loading control is shown. (B) Quantitative western blot analyses for GABA_Aα1 and GABA_Aβ2 are shown. Average pixel values averaged over biological triplicates from IP experiments are indicated for wild type and α9^−/−. GABA_Aα1 protein expression increased by 18% in α9^−/− (p = 0.041) and GABA_Aβ2 expression increased by 58% in α9^−/− (p = 0.029) when compared to wild type. Error bars indicate SEM, and a significant change is denoted with an asterisk (two-tailed t-test). (C) In adult wild type ears, GABA_Aβ2 immunostaining is observed in the inner and outer hair cells. (D) In α9^−/− OHCs, GABA_Aβ2 immunolabeling is particularly dense in the apical regions, as well as at the synaptic pole. The small arrows indicate outer hair cells and the larger arrow in (D) indicates the border between outer hair cells and Deiters' cells. Sp lim, spiral limbus; ToC, tunnel of Corti; OHC, outer hair cell. Scale bar, 25 µm.

Figure 5. Gene set enrichment analysis of α9^−/− and wild type.

(A) GSEA analysis of IGF1_vs_PDGF gene set. The inset depicts the running sum for the IGF1_vs_PDGF gene set at each gene in the rank-ordered list comparing α9^−/− and wild type at P60. The enrichment score is the maximum (up-regulated molecules) or minimum (down-regulated molecules) deviation from zero. The main plot shows the frequency distribution of normalized enrichment scores for all 1,390 gene sets. The arrow points to the bin that includes the IGF1_vs_PDGF gene set. (B) Several significantly differentially expressed gene sets with particular roles in synapse development or function are shown at P3, P13 and P60 (FDR q-value <0.05). Rank indicates the position of the gene set when compared to all others (a total of 1,390). (C) Log₂ fold changes from qPCR validations are shown for four genes that were selected from three gene sets: Capn1 (LIN_WNT_UP), Adam10 and Mbp (LE_MYELIN_DN) and Nedd8 (UBIQUITIN_MEDIATED_PROTEOLYSIS). A significant change is denoted with an asterisk (p<0.05, two-tailed t-test).

Figure 6. Gene set enrichment analysis of bicluster9.

(A) All the genes included in bicluster9 are represented as a heatmap ordered by their ranking in the gene list comparing the expression between wild type and α9^−/− at P60. The top 20 genes that contribute the most to the enrichment score are shown. An asterisk next to the gene name indicates that it is repeated twice because of probe mappings. Italicized genes in bold have previously known functional roles in the nervous system. Red indicates high expression; blue indicates low expression. (B) The running sum enrichment scores for bicluster9 are shown for each age. NES, normalized enrichment score. (C) Leading edge subset analysis for bicluster9 at each age. Normalized microarray signal values for wild type and α9^−/− animals are averaged over all genes in the leading edge subset at each time point. Error bars indicate SEM.

Figure 7. Comparison of dynamic gene expression levels of wild type and α9^−/−.

(A) Hierarchical clustering of centroids indicates similarity between genotypes and across ages. Mosaic patterns are pseudo-colored SOMs of the 650 metagene profiles from developing wild type and α9^−/− cochleae. Each tile contains genes with similar expression patterns and the spatial location of each tile is preserved across maps. Tile colors indicate the expression level of the centroids, with red representing high expression and blue representing low expression. (B) Principal component analysis (PCA) comparing SOM centroids. Each sample is projected onto a three dimensional grid comprised of the first three (strongest) principal components. PCA indicates the differences between wild type and α9^−/− cochlea at P13 and P60.

Figure 8. Summary of over-represented gene categories.

Several functional categories, based on PANTHER annotations and GSEA results, are significantly perturbed upon loss of α9 nicotinic subunit at postnatal time periods P13 and P60. While several processes are common to both ages (orange box), others are preferentially expressed at P13 (green box) or P60 (blue box). Only Wnt/β-catenin pathway (highlighted in red) is identified as over-represented in α9^−/− by both PANTHER and GSEA analyses.