Dynamic changes in the microRNA expression profile reveal multiple regulatory mechanisms in the spinal nerve ligation model of neuropathic pain

Abstract

Neuropathic pain resulting from nerve lesions or dysfunction represents one of the most challenging neurological diseases to treat. A better understanding of the molecular mechanisms responsible for causing these maladaptive responses can help develop novel therapeutic strategies and biomarkers for neuropathic pain. We performed a miRNA expression profiling study of dorsal root ganglion (DRG) tissue from rats four weeks post spinal nerve ligation (SNL), a model of neuropathic pain. TaqMan low density arrays identified 63 miRNAs whose level of expression was significantly altered following SNL surgery. Of these, 59 were downregulated and the ipsilateral L4 DRG, not the injured L5 DRG, showed the most significant downregulation suggesting that miRNA changes in the uninjured afferents may underlie the development and maintenance of neuropathic pain. TargetScan was used to predict mRNA targets for these miRNAs and it was found that the transcripts with multiple predicted target sites belong to neurologically important pathways. By employing different bioinformatic approaches we identified neurite remodeling as a significantly regulated biological pathway, and some of these predictions were confirmed by siRNA knockdown for genes that regulate neurite growth in differentiated Neuro2A cells. In vitro validation for predicted target sites in the 3'-UTR of voltage-gated sodium channel Scn11a, alpha 2/delta1 subunit of voltage-dependent Ca-channel, and purinergic receptor P2rx ligand-gated ion channel 4 using luciferase reporter assays showed that identified miRNAs modulated gene expression significantly. Our results suggest the potential for miRNAs to play a direct role in neuropathic pain.

Figure 1. Profiling of miRNAs in DRG.

a. Schematic representation of spinal L5 nerve ligation adapted from Decosterd and Woolf (Pain, 2000). Individual L4 and L5 DRG tissue samples were collected from the ipsi- and contralateral sides of 5 sham (no ligature) surgery and 5 SNL surgery rats four weeks after SNL for analysis. b. Pair-wise correlation of raw C _T values among DRG samples. Heat map of sample-to-sample (Pearson) correlation of raw C _T values from 248 detectable miRNAs in 40 DRG samples from the 8 experimental groups. Column and row labels are symmetric. c. Distribution of normalized C _T values (Δ C _T) in 40 samples. Each line represents C _T distributions from one particular sample. The color represents the experimental group. L4 and L5 ipsilateral DRGs in the SNL cohort (blue and black, respectively) are clearly separated from each other and from the rest of the groups which appear to form an aggregate ‘group’ that contains all sham samples and all SNL contralateral samples. The majority of data points (C _T ranging from 29 to 39) in the blue curves are “shifted” to the right of the rest of the DRG sample distributions, indicating a general down-regulation (hence higher C _T) of miRNA expression in L4 ipsilateral DRGs of the SNL cohort.

Figure 2. Luciferase reporter gene assay to assess the effect of miRNA on gene function.

The effect of miRNA on gene regulation was monitored by co-transfecting 100 nM miRNA and the luciferase reporter gene. The 3′-UTR of Scn11a (Nav1.9), Cacna2d1and P2rx4 were cloned downstream of the luciferase open reading frame. In addition to miRNAs that could bind to the corresponding UTR, negative control pre-miRNA was transfected with the corresponding luciferase reporter plasmid for each gene analyzed. The luciferase activity was measured 24 hrs after transfection and the ratio of firefly∶renilla luciferase is expressed as percentage of control ± standard deviation shown. A statistically significant difference from control miRNA was calculated using one-way ANOVA and p-value of <0.003 are represented by * in the figure.

Figure 3. Distribution of predicted 63-set target sites in individual transcripts.

The TargetScan algorithm was used for miRNA target prediction and there are hundreds of transcripts containing multiple predicted 63-set target sites. We increased the stringency for the predicted target sites of the 63 miRNAs by applying a context score filter above the 60th percentile. This reduced the number to 10,751 predicted target sites in 4899 rat genes.

Figure 4. Ingenuity Pathway Core Analysis functions.

Analysis of gene transcripts which had three-or-more (Panel A), four-or-more (Panel B) and five-or-more (Panel C), predicted 63-set target sites and a TargetScan context score above the 60^th percentile shown. The listed functions under the graphics are from the category called “Nervous System Development and Function.” The functions are sorted by p-value and only rows having 10 or more molecules are shown. Controls (Panel D) of randomly chosen gene sets of comparable sizes to the five-, four-, and three-or-more gene sets were also run through Ingenuity Pathway Core Analysis and appear in this Analysis Comparison. Only the top three broad categories are shown and they are sorted by the five-or-more order of functions. Overlap (Panel E) between the L4 ipsi- vs. contralateral mRNA transcripts and predicted mRNA targets of 63-set miRNA (60^th percentile) were run through Ingenuity Pathway Core Analysis. The top six functional categories are shown and the sub-categories under “Nervous System Development and Function” are listed, sorted by p-value and only rows having 10 or more molecules are shown.

Figure 5. Functional categories perturbed in the SNL L4 DRGs.

Functional processes related to Nervous System Development and Function with a minimum of 10 genes are indicated. The bars represent the negative (log p-value) and the orange data points indicate the number of genes for the given functional category. P-values indicate the probability that the observed association between the described functional categories and the SNL pain gene set is due to chance. The ratio indicates the percentage of genes in these functional categories that are found in the SNL pain gene set. The ratio thus indicates the strength of the association, and the p-value measures the statistical significance. The horizontal line indicates a p-value of 0.05.