Correlational analysis for identifying genes whose regulation contributes to chronic neuropathic pain

Abstract

Background: Nerve injury-triggered hyperexcitability in primary sensory neurons is considered a major source of chronic neuropathic pain. The hyperexcitability, in turn, is thought to be related to transcriptional switching in afferent cell somata. Analysis using expression microarrays has revealed that many genes are regulated in the dorsal root ganglion (DRG) following axotomy. But which contribute to pain phenotype versus other nerve injury-evoked processes such as nerve regeneration? Using the L5 spinal nerve ligation model of neuropathy we examined differential changes in gene expression in the L5 (and L4) DRGs in five mouse strains with contrasting susceptibility to neuropathic pain. We sought genes for which the degree of regulation correlates with strain-specific pain phenotype.

Results: In an initial experiment six candidate genes previously identified as important in pain physiology were selected for in situ hybridization to DRG sections. Among these, regulation of the Na+ channel alpha subunit Scn11a correlated with levels of spontaneous pain behavior, and regulation of the cool receptor Trpm8 correlated with heat hypersensibility. In a larger scale experiment, mRNA extracted from individual mouse DRGs was processed on Affymetrix whole-genome expression microarrays. Overall, 2552 +/- 477 transcripts were significantly regulated in the axotomized L5DRG 3 days postoperatively. However, in only a small fraction of these was the degree of regulation correlated with pain behavior across strains. Very few genes in the "uninjured" L4DRG showed altered expression (24 +/- 28).

Conclusion: Correlational analysis based on in situ hybridization provided evidence that differential regulation of Scn11a and Trpm8 contributes to across-strain variability in pain phenotype. This does not, of course, constitute evidence that the others are unrelated to pain. Correlational analysis based on microarray data yielded a larger "look-up table" of genes whose regulation likely contributes to pain variability. While this list is enriched in genes of potential importance for pain physiology, and is relatively free of the bias inherent in the candidate gene approach, additional steps are required to clarify which transcripts on the list are in fact of functional importance.

Figure 1

L5 spinal nerve ligation (SNL) alters the expression of mRNAs in the ipsilateral L5DRG. A, Dark-field micrographs display hybridization signals for each of the six transcripts studied: Scn10a, Scn11a, P2rx3, Trpv1, Trpa1 and Trpm8. DRG sections originate from naïve B6 mice (left column) and B6 mice 3 days after SNL surgery (right column). The scale bar refers to all images: 100 μm. B, Column heights (mean ± SEM, n = 4–6) indicate the proportion of DRG neuronal profiles that were positive for expression of each of the six mRNA types in each of the five strains examined. White columns refer to naïve mice; black columns refer to SNL operated mice. Statistical differences between naive and nerve injured mice are indicated with asterisks (Mann-Whitney U test) * p < 0.05; ** p < 0.01.

Figure 2

Galanin is up-regulated and Scn11a is down-regulated following spinal nerve ligation (SNL). Gene expression, plotted in relation to that of the stable housekeeping gene Gapdh, was assessed using TaqMan analysis 3 days following L5 SNL. Expression profiles for both genes are shown for individual mice of the A, C3H strain (labelled C3H_a-o) and B, C58 strain (labelled C58_a-r). Arrows in B, indicate two outlier mice in which expression levels of both Galanin and Scn11a correspond to sham operated animals, suggesting that spinal nerve injury had not been carried out adequately.

Figure 3

Regulation of expression of the cytokine receptor Ccr2 after L5 SNL correlated positively with the degree of tactile hypersensitivity. Each symbol represents one mouse of the AKR, CBA or B6 strain. Expression data, for individual mice, are plotted as: A, the raw signal intensity values associated with Ccr2 gene expression, and B, the log ratio of the signal intensity of the individual mouse divided by the average signal intensity of all sham operated mice of the same strain. Values for tactile hypersensitivity of the individual mice were normalized as indicated in the Methods, where increasing values indicate greater sensitivity. r = Pearson correlation coefficient, p = statistical significance of r, q =false discovery rate (FDR) coefficient associated with Ccr2 in the microarray analysis.

Figure 4

Regulation of expression of the TTX-R Na⁺ channel alpha subunit Scn11a after L5 SNL correlated negatively with the degree of tactile hypersensitivity. Each symbol represents one mouse of the AKR, CBA or B6 strain. Expression data, for individual mice, are plotted as: A, the raw signal intensity values associated with Scn11a gene expression, and B, the log ratio of the signal intensity of the individual mouse divided by the average signal intensity of all sham operated mice of the same strain. The value for tactile hypersensitivity for these mice was normalized as indicated in the Methods, where increasing values indicate greater sensitivity. r = Pearson correlation coefficient, p = statistical significance of r, q = false discovery rate (FDR) coefficient associated with Scn11a in the microarray analysis.