Neuropathic pain (NP) is poorly managed and causes by a disease or lesion of the somatosensory nervous system (Cohen and Mao, 2014), accompanied by metabolic diseases, mechanical trauma or tumor invasion; it concerns various pathophysiological changes within the peripheral and/or central nervous system (Costigan et al., 2009; Zhang Y. H et al., 2021). A distinctive feature of peripheral neuropathic pain is mechanical allodynia, which is triggered by light touch (Costigan et al., 2009; Liu et al., 2018). A number of rodent NP models that have been established to shape exploration efforts in the pathophysiological mechanisms of NP in the nervous system include the sciatic chronic constriction injury (CCI), spared nerve injury (SNI), spinal nerve ligation (SNL), sciatic nerve transection, diabetic neuropathy and drug-induced neuropathy models. Although human perception of pain is subjective that it cannot be completely duplicated in animal models, tactile allodynia in rodent models is regarded to be a corresponding pattern for neuropathic mechanical hypersensitivity in patients. These models commonly provide a clear picture of cause-effect relationship or particular biomarkers in NP.

In recent years, several studies have comprehensively discussed epigenetic modifications have dominated our understanding about activation and suppression of various gene expressions in the persistent and development of NP models (Penas and Navarro, 2018; Guo et al., 2019; An et al., 2021; Zheng et al., 2021). Unfortunately, in-depth mechanisms of gene transcriptome alterations in NP pathogenesis are not yet fully understood. Our previous study (Guo et al., 2019) have investigated the biological function of microRNAs (miRNAs) in frequently used NP rat model like CCI, SNI and SNL. Although miRNAs are non-coding RNAs containing only 19–25 nucleotides and are not directly involved in peptide synthesis, they markedly control biological processes by affecting mRNA stability, as well as protein translation (Lopez-Gonzalez et al., 2017; Song et al., 2020). Nervous tissues, such as dorsal root ganglia (DRG), spinal cord and prefrontal cortex (PFC), are of considerable concern in terms of the therapeutic potential of miRNAs in NP (Tang et al., 2021). A number of miRNAs, such as miR-183 cluster, miR-155, miR-145, miR-203 and miR-200b/429, are detected in nervous tissues and are involved in NP by affecting neuronal excitability, neuroinflammation, or neuronal plasticity (Hamada et al., 2014; Peng et al., 2017; Shi et al., 2018; Zhang Y et al., 2020). In addition, the dysregulation of certain miRNAs mediates downstream molecular mediator of pain, which contain transient receptor potential receptor channels, purinergic receptors, and voltage-gated sodium and calcium channels (Yan et al., 2020a; Yeh et al., 2020; Li et al., 2021). Although numerous studies have attempted to provide novel mechanistic insights into the role of miRNAs roles in NP, the entire pain pathway still remains to be elucidated. Hence, we performed comprehensive literature search on the microRNA-related molecular mechanisms of NP, and bioinformatics analysis of their transcriptional regulation. This study may shed light on the enigmatic pathophysiology of NP.

We identified 6,653 records from online databases. After duplicate records were removed, 5,288 items remained. Two authors assessed the title and abstract and found that 201 articles were pushed on full-text review. At last, 133 literatures met the eligibility criteria and were included (excluded articles are listed in Supplementary Table S1). Among these included studies, 23 were about miRNA profiles (Supplementary Tables S2, S3), 113 were about miRNA experimental verification and three (Dong et al., 2014; Lu et al., 2017; Liu L et al., 2020) were both about miRNA profile and experimental verification. Among the 113 miRNA experimental verification articles, 96 are about NP surgical models (Supplementary Table S4), 6 are drug-induced NP models (Supplementary Table S5), and 11 are disease-induced NP models (Supplementary Table S6). The flowchart of the study selection procedure is shown in Figure 1.

In this study, miRNAs in nervous tissues of NP rodent models were comprehensively analyzed using 23 miRNA profile articles and 113 miRNA experimental verification articles. Compared with the sham-operated group, a total of 91 miRNAs verified by 113 experimental articles were differentially expressed in the NP group. The potential functional specificity of the differentially expressed miRNAs for pain was determined by GO and KEGG pathway analysis. The results suggest that miRNAs served a vital role in NP development and potentially novel strategies for NP management.

NP pathogenesis is complex and remains poorly understood. Regions of DRGs, SDH and anterior cingulate cortex (ACC) form the somatosensory pathway from primary sensory neurons to afferent nerve to central nervous system (CNS) in NP (Guo et al., 2019). DRGs contain most of the body’s sensory neurons, and transmitted sensory messages from receptors, such as thermoreceptors and nociceptors, are active participants in the signalling process (Pope et al., 2013). SDH is the relay station of nociceptive stimuli to process somatosensory information and is characterized as an unambiguous laminar structure with a quantity of excitatory and inhibitory interneurons (Peirs and Seal, 2016). ACC is a dominant cortical area implicated in diverse neurological processes, such as nociception, cognition and emotion and plays a critical role in emotional/aversive component of pain (Tsuda et al., 2017).

Numerous circuits and molecules were identified as potential biomarkers and regulators in NP. The renowned mediators contain the transient receptor potential (TRP) channels (Lu et al., 2017; Miao et al., 2019; Duan et al., 2020; Zhang and Chen, 2021), voltage-gated sodium channels (Na_Vs) (Chen et al., 2014; Lin et al., 2014; Shao et al., 2016; Su et al., 2017; Cai et al., 2018; Li L et al., 2019; Yang et al., 2019; Yan et al., 2020b; Ye et al., 2020; Sun et al., 2021), and voltage-gated calcium channels (CaVs) (Favereaux et al., 2011; Gandla et al., 2017; Peng et al., 2017). In DRG and SDH, TRP channels act as transducers for selectively activating sensory neurons and conveying diverse kinds of stimuli involving chemical, mechanical, and thermal stimuli. TRP ankyrin 1 (TRPA1) is one of branches of TRP channels contributing to NP. In NP surgical models and drug-induced NP models, TRPA1 is overexpressed and targeted by miR-155 (Miao et al., 2019; Duan et al., 2020), miR-141-5p (Zhang and Chen, 2021), mmu-miR-449a (Lu et al., 2017). Ectopic action potential generation is considered to be one of causes of NP formulation. Na_Vs intrinsically contribute to ectopic activity generation (Bannister et al., 2020) and are typically encoded by different subunits, Na_V1.1-Na_V1.9. DRG neurons contain the most varied kind of Na_V subtypes. Among them, Na_V1.3 (Chen et al., 2014; Ye et al., 2020), Na_V1.6 (Li L et al., 2019), Na_V1.7 (Sun et al., 2021) and Na_V1.8 (Yan et al., 2020b) are overexpressed in NP and are targeted by miR-96, miR-183, miR-182 and miR-30b. In addition, Sun and colleagues (Sun et al., 2021) demonstrated that miR-96 knockout mice showed thermal and mechanical allodynia. This allodynia was alleviated by intraperitoneal or intrathecal injection of Na_V1.7 or Na_V1.8 blockers. SNI-induced suppressed expression of miR-96 in SDH showed negative correlation to overexpression of Na_V1.7-Na_V1.9. This dysregulation promotes the NP process, and can be attenuated by intrathecal injection of corresponding Na_Vs blockers. Thus, alternation in the expression of Na_Vs in nerve injury-induced NP might be efficacious for pain relief. Furthermore, increased expression and function of Ca_Vs also caused NP by increasing transmitter release (Davies et al., 2007). Ca_Vs are grouped into two classes, namely, high-voltage activated type and low-voltage activated type (also known as T-type) (Catterall et al., 2005). Apart from T-type Ca_Vs, Ca_Vs have three types of subunits, namely, the α1 channel-forming subunit, the intracellular β subunit and the α2δ auxiliary subunit. The α2δ auxiliary subunit is composed of two disulfide-bonded polypeptides (α2 and δ) (Dolphin, 2018). Furthermore, Ca_V1.2 that encodes the α1 subunit of Ca_Vs targeted by miR-103 is upregulated in SNL rats (Favereaux et al., 2011); Ca_V2.3, which targeted by miR-34c-5p, is downregulated in bone metastatic pain mice (Gandla et al., 2017). Both mRNA and protein levels of the α2δ-1 subunit and α2δ-2 subunit of Ca_V are upregulated in SNI mice and are correlated with mechanical allodynia (Peng et al., 2017). The miR-183 cluster (miR-183/96/182) controls over 80% of NP-related genes and scales mechanical allodynia by modulating the α2δ-1 subunit and α2δ-2 subunit (Peng et al., 2017). Thus, post-injury alteration in α2δ expression level, bound with upstream miRNAs dysregulations, provides convincing evidence for potential biomarkers and regulators in NP and therapeutic possibilities of Ca_Vs.

Several studies indicated that epigenetic changes, translational modification and post-translational control influence NP development and management. Epigenetic changes are considered to involve histone acetylation and DNA methylation. As such, gene expression is regulated but does not change the coding sequence. DNA methylation likely reduces transcription efficiency, whilst histone acetylation in DNA presents the active transcriptional region (Jenuwein and Allis, 2001). Methyl-CpG-binding protein 2 (MeCP2), a protein with affinity for methylated DNA and repressing transcription from methylated gene promoter, is vital for neuronal proliferation and embryonic development. Manners et al. (Manners et al., 2016) found that SNI caused the redistribution of MeCP2 to methyl-CpG binding domain. Enriched MeCP2 can bind to the miR-126 locus in NP condition and restrain miR-126 transcription. Repressed miR-126 expression contributes to the up-regulation of its two target genes Dnmt1 and Vegfa in SNI mice. DNA methyltransferase 3A (Dnmt3a) was increased in injured DRG. Xu et al. (Xu et al., 2017) indicated that SNL-induced miR-143 downregulation is a negative regulator in Dnmt3a expression in DRG. Furthermore, histone deacetylases (HDACs) were implied in the mechanisms of transcriptional regulation, cell cycle progression, neuron degeneration and regulation of neuronal plasticity. Overexpression of HDAC4 alleviated the effects of miR-206-3p on NP (Wen et al., 2019). Overall, miRNA-regulated DNA methylation and histone acetylation may be a potential target for NP therapeutic management. The interactions between transcriptional factors (TFs) and mitogen-activated protein kinases (MAPKs) are also involved in NP. MAPK6 up-regulation in SDH of CCI rats was deemed to be a direct downstream target gene of miR-26a-5p (Zhu et al., 2018). Suppressed miR-125a-3p contributed to p38 MAPK up-regulation in rat trigeminal ganglions at different time points with prosopalgia (Dong et al., 2014). Inhibition of miR-155 reduces NP during chemotherapeutic bortezomib via downstream signals p38-MAPK (Duan et al., 2020). Additionally, in diabetes mellitus rats, miR-133a-3p antagomir administration lightened diabetic NP and down-regulated p38 MAPK (p-p38) phosphorylation (Chang et al., 2020). Hence, pain-related miRNAs and proteins may reveal critical insights into how neurons process incoming stimuli inputs and attempt to create an effective method for low risk of inflammation and NP management.

PPI network analysis may help predicted some target genes to broaden the treatment options for NP in the near future. Mras connected with Pik3ca displayed highest connective score with combined score of 0.997 predicted by PPI network. Mras is a member of the Ras family of small GTPases that impacts mouse embryonic stem cell plasticity and neurogenesis (Mathieu et al., 2013). Pik3ca is usually described as an oncogenic gene and related to PI3K-AKT signaling pathway (Koundouros et al., 2020), which is in line with Figure 9. Another pair with high combined score of 0.996 is Creb1 and EP300. Creb1 encodes a transcription factor. This protein activity is regulated by protein kinase A (Maekawa et al., 2008) and also related to PI3K-AKT signaling pathway (Rodon et al., 2014). Creb1 was recognized as one of most enriched genes in brain and found to regulate a quite number of downstream genes in NP process (Yan et al., 2019). EP300 is identified as a transcriptional co-activator and is important in the processes of cell proliferation and differentiation (Son et al., 2019). Knockdown of miR-30a-3p in L4-6 spinal dorsal toot increased EP300 level and induced NP (Tan et al., 2020). At present, there is no study targeting Mras and Pik3ca in neuropathic pain. Using Pik3ca inhibitors in NP model may be an initial attempt to ameliorate pain in the near future.

Bioinformatics insights into the potential neurobiological mechanisms correlative of NP were investigated in a comprehensive and systematic literature search framework. This framework provided guidance for literature validation of microRNA regulatory networks inferred from real experimental data. An extensive range of miRNAs was obtained to explore the relationship among their downstream genes via network and pathway-based analysis. Based on this analysis, it is helpful to form integrative research strategies from NP mechanisms to treatment. Some limitations still exist in this study. Firstly, a few microRNAs revealed contradictory results in NP process, which may be caused by differences in tissue collection timing, modeling methods and detected tissue region. Secondly, the GRADE system to evaluate level of evidence and Funnel plots for publication assessment are lacking. Although the outcomes deduced from bioinformatic analysis, to some extent, may exist heterogeneity from different study designs, network analysis combined pathway analysis can be more robust for possible false positive results.

In summary, we investigated miRNA expression and their target gene roles in NP. Bioinformatics analysis elucidates that several miRNAs (miR-200b-3p, miR-96, miR-182, miR-183, miR-30b, miR-155 and miR-145) and its target genes are involved in known relevant pathways of NP. Targets on voltage-gated sodium channels may be harnessed for pain relief. As knowledge of molecular, genetic and epigenetic mechanisms about pain specificity and neuron plasticity accrues from rodent models, further delineation of signal processing and modulation in neuronal ensembles is key to achieving therapeutic success in future studies.