The addition of “omics” to certain molecular term reflects a comprehensive assessment of a set of molecules. Integrated analysis of multiple NGS-based datasets, which highly rely on high-throughput sequencing, has already revolutionized the medical research [8]. Chronic pain is a critical health issue globally, affecting millions of people. However, there exists difficulty for directly obtaining certain tissues directly from patients clinically. Thus, we believed that understanding the molecular networks of chronic pain animal models through a comprehensive approach is an alternative. Herein, we provided an integrated analysis based molecular profile, which focuses on the response of rat DRG tissues to CFA injection, at epigenetic, transcription and post-transcriptional regulation level. Overall, we identified 418 differentially expressed mRNAs, 120 differentially expressed miRNAs and recognized more than 2,500 DMRs in CFA-treated groups. Through gene set enrichment analysis, we validated some of the previously reported CFA-response related signaling pathways that are also highly related to CFA induced inflammation in our dataset as well, such as, NF-κB signaling pathways [36] and IFN signaling pathways [38]. Moreover, we also identified many new genes/pathways that are potentially highly involved in this pain response model, including Reg family genes (Reg3a, Reg3b) and the AP-1 transcription related pathways. It is interesting that after we adjusted the cell heterogeneity of the methylation sequence data by using CHALM [41], a recent invented software for analyzing methylome, we identified 6832 significant DMRs, suggesting the CFA’s impact on rat’s methylome is underestimated by the traditional analysis method. Besides inflammation related pathways or terms, the CHALM identified DMRs are also enriched to heart contractions and heart rate, indicating that methylome re-wire is involved in the chronic pain linked heart malfunctions or diseases (Supplementary Table 10). Finally, Based on our multi-omics profiling of the CFA-induced chronic pain model, we selected top 10 differentially expressed genes, miRNAs and differentially methylated regions as the multi-omics signature for our chronic pain model. CFA treatment group and control group can be clearly separated by these signatures (Fig. 6a-c) (Supplementary Table 7–9), and is consistent to the idea that miRNA and DNA methylation regulate mRNA transcription.
Pancreatitis-associated proteins, which are from Reg families, have been previously linked to modulation of spinal sensory pathways in pathological pain states [9]. Our results demonstrated that Reg3b and Reg3a were significantly upregulated for 12-fold and 6-fold, respectively, in CFA-treated groups, indicating that the Reg family genes may play a crucial role in regulating inflammatory pain response in DRG tissues. AP-1 transcription factor is composed of dimerization of a bZIP (basic region leucine zipper) domain via the Fos and Jun subunits. AP-1 regulation network was previously linked to chronic pain response [3]. However, it was never recognized as a central pathway in pain response before. In our dataset, AP-1 network is considered as the regulation central hub. Not only the CFA-induced miRNA and mRNA interactions are highly enriched for AP-1 networks, but also the CFA-induced DMR motifs are enriched for the AP-1 transcription binding activities. Therefore, we hypothesized that inhibiting certain AP-1 network genes, such as Egr1, which is recognized both as a pleiotropic inflammatory trans-activator [25] and a chronic pain contributor [11], could have a chance to alleviate chronic pain. Interestingly, most of the differentially expressed miRNAs (104/120) were downregulated after CFA treatment. The global downregulation of miRNA itself is an interesting phenomenon, which was missed in the previous array-based study [17]. Notably, in these downregulated miRNAs, many of them are targeting AP-1 network genes. Thus, we speculate adding back of these miRNAs could inhibit the overactivation of AP-1 network, thus alleviate pain. Overall, we believe AP-1 network plays a central role in regulating inflammatory pain responses through a Methylation-transcription-posttranscription regulation axis.
In conclusion, this study provides a comprehensive transcriptomic profile of the CFA-induced inflammatory pain rat model via an approach that target DNA methylation, gene expression as well as post-transcriptional regulation. Our study has certain limitations. In the first place, although we included 10 pairs of rats for the CFA treatment, we only have a single time point post CFA injection, which is 24 hours. Future studies should focus on the 48hours, 72hours or even 7 days post treatment, as CFA could induce a chronic inflammation as well. In addition, although we demonstrated that AP-1 is likely to work as a regulation hub for CFA-induced inflammation response potentially, we did not include the biochemistry assays to further investigate the alteration of the AP-1 signal regulation. Future studies could focus on the effect of individual molecule or gene of AP-1 network on regulating chronic pain.