miR‐26a‐5p alleviates CFA‐induced chronic inflammatory hyperalgesia through Wnt5a/CaMKII/NFAT signaling in mice

Abstract Background Inflammation often leads to the occurrence of chronic pain, and many miRNAs have been shown to play a key role in the development of inflammatory pain. However, whether miR‐26a‐5p relieves pain induced by inflammation and its possible mechanism are still unclear. Methods The complete Freund's adjuvant (CFA)‐induced inflammatory pain mouse model was employed. Intrathecal or subcutaneous injection of miR‐26a‐5p agomir was performed after modeling to study its antinociceptive effect and the comparison of different administration methods. Bioinformatics analysis of miRNAs was performed to study the downstream mechanisms of miR‐26a‐5p. HE staining, RT‐qPCR, Western blotting, and immunofluorescence were used for further validation. Results A single intrathecal and subcutaneous injection of miR‐26a‐5p both reversed mechanical hypersensitivity and thermal latency in the left hind paw of mice with CFA‐induced inflammatory pain. HE staining and immunofluorescence studies found that both administrations of miR‐26a‐5p alleviated inflammation in the periphery and spinal cord. Bioinformatics analysis and dual‐luciferase reporter gene analysis identified Wnt5a as a direct downstream target gene of miR‐26a‐5p. Wnt5a was mainly expressed in neurons and microglia in the spinal cord of mice with inflammatory pain. Intrathecal injection of miR‐26a‐5p could significantly reduce the expression level of Wnt5a and inhibit the downstream molecules of noncanonical Wnt signaling Camk2/NFAT, inhibiting the release of spinal cord inflammatory factors and alleviating the activation of microglia. In addition, miR‐26a‐5p could also inhibit lipopolysaccharide (LPS)‐stimulated BV2 cell inflammation in vitro through a noncanonical Wnt signaling pathway. Conclusions miR‐26a‐5p is a promising therapy for CFA‐induced inflammatory pain. Both intrathecal and subcutaneous injections provide relief for inflammatory pain. miR‐26a‐5p regulated noncanonical Wnt signaling to be involved in analgesia partly through antineuroinflammation, suggesting a pain‐alleviating effect via noncanonical Wnt signaling pathway in the CFA‐induced inflammatory pain model in vivo.


| BACKG ROU N D
Inflammation is a vital physiological response to noxious stimuli (e.g., trauma, infection, and surgery), which facilitates to repair the impairment and restore the homeostasis. [1][2][3] While the inflammation maintains 6 weeks or longer, for example, dermatitis, arthritis, and colitis, inflammation would exhibit its evil nature resulted in persistent damage. Inflammatory factors, for example, IL-1β, TNFα, and prostaglandin, produced during chronic inflammation could sensitize peripheral nerve resulted in suffering chronic pain. 4 Chronic inflammatory pain has become one of the major public health problems worldwide, seriously affecting the quality of patients' life and causing a huge economic burden. 5,6 A recent population-based survey in Europe found that 25%-35% of adults report chronic inflammatory pain. 5,7,8 Another study showed that 79% of patients with chronic inflammatory pain still had some degree of chronic pain 4 years later. 9 Therefore, the clinical management of chronic inflammatory pain remains a great challenge. 10 The drugs currently used to treat chronic inflammatory pain mainly include non-steroidal anti-inflammatory drugs (NSAIDs), opioids, and other adjuvant drugs, but most of them have many side effects and the treatment efficacy is not satisfied. 11 There is an urgent need to redefine and improve the pain management strategy (e.g., cellular therapy and nuclear acid medicine) to relief inflammatory pain.
Inflammatory pain is closely related with immune response and is normally mimicked in animals through exerting pro-inflammatory stimuli, such as tissue incision, complete Freund's adjuvant (CFA), and lipopolysaccharides (LPS). 4 During the acute phase, pain is usually proportional to the inflammation. While there is the development into the chronic phase, the persistent inflammation triggers cytokines and irritant substances to regulate the peripheral and central mechanisms of pain through the neuroimmune mechanism, which would ultimately result in hyperalgesia, allodynia, and idiopathic pain. 12 During inflammation, immune cells release mediators that act on the peripheral nerve terminals of nociceptive neurons. 13 Action potentials are transmitted through the dorsal root ganglion (DRG) to the spinal cord, where they are eventually relayed to the brain and processed as pain. In the dorsal horn of the spinal cord, neuroimmune interactions contribute to the central mechanisms of pain. 14 Neuroinflammation act as a key mechanism in the pathogenesis of chronic inflammatory pain, 15 and in inflammatory pain states, neuroinflammation involves activation of microglia, release of pro-inflammatory mediators, and overexpression of pain-related signaling molecules. Primary nociceptive afferent nerves in the DRG release glutamate, ATP, and chemokines from their central terminals. T cells, microglia, and astrocytes also produce pro-inflammatory cytokines and growth factors that act presynaptically and postsynaptically to stimulate nerve endings to increase nerve conduction and mediate central pain sensitization. Immune cells release mediators directly sensed by nociceptor terminals to modulate neuronal excitation and pain transduction. IL-1β, tumor necrosis factor (TNF), nerve growth factor (NGF), and prostaglandin E2 (PGE2), etc. bind to cognate receptors expressed at nociceptor terminals to mediate neuronal firing. 12 The overall result of these immune-mediated pathways in nociceptors is a lowering of the threshold for responses to mechanical or thermal stimuli, leading to increased pain sensitivity.
Therefore, targeting these signaling molecules and/or cells may be a potential therapeutic strategy for chronic inflammatory pain. 10 miRNAs are short and endogenous noncoding RNAs which regulate mRNA expression through inhibiting mRNA expression and/or reducing mRNA stability. 16,17 In the past decades, miR-NAs not only have been developed as biomarkers for diagnosis, prognosis prediction, but also have been considered as a potential and attractive target for developing new therapies of different diseases, including cancers, cardiovascular diseases, and diabetes. 18,19 Many previous studies have shown that miRNAs have different roles during the pathophysiological process of inflammatory pain. 20 26 In our previous study, we found that miR-26a-5p in small extracellular vesicles (sEVs) derived from human placental mesenchymal stem cells (hPMSCs) could alleviate peripheral nerve injury-induced neuropathic pain by decreasing spinal neuroinflammation via Wnt5a, a noncanonical Wnt signaling pathway. 27 Additionally, miR-26a-5p also exhibits its antiinflammatory potential in other disease models. miR-26a-5p/IL-6 axis alleviates sepsisinduced acute kidney injury by inhibiting renal inflammation. 28 Overexpression of miR-26a-5p could alleviate acute lung injury by targeting TLR4 to reduce cellular inflammation. 29 Another previously study also reported that miR-26a-5p alleviates neuropathic pain in a rat model of CCI through decreasing the release of the inflammatory factors IL-1β, IL-6, and TNFα by targeting MAPK6. 30 However, whether miR-26a-5p could alleviate inflammatory pain through its antiinflammatory effect and its potential target molecules is still need to be further explored.
Therefore, in the current study, we examined the analgesic effect of miR-26a-5p in CFA-induced inflammatory pain mice model.

K E Y W O R D S
inflammatory pain, miR-26a-5p, neuroinflammation, Wnt5a In addition, we compared the analgesic effects of two different miR-26a-5p administration methods (intrathecal and subcutaneous).
Finally, we demonstrate that miR-26a-5p and its target gene Wnt5a play an important role in allodynia by reducing microglial activation in the spinal dorsal horn and suppressing inflammation through non-

| Intrathecal injections
For intrathecal injection, we use the method described previously. 32 Mice were anesthetized with isoflurane. The thumb and middle finger of the left hand firmly hold the paralumbar region of the iliac crest, and the index finger is placed on the tip of the spinous process of the sixth lumbar vertebra (L6), the highest point of the vertebral body. All intrathecal injections were delivered in a total volume of 10 μL. The needle is inserted into the fifth intervertebral space (L5-L6), resulting in a sudden lateral movement of the tail. The needle is held in place for 10 s and then slowly withdrawn to avoid spillage.

| Behavioral test
All mice were acclimated to the test environment for 30 min before testing for baseline nociceptive thresholds. The mechanic threshold of mice was measured by using the von Frey monofilament (Semmes Weinstein) "up-down" method. 33 Mice were placed in glass boxes on grid iron racks (9 × 25 × 25 cm). At the beginning of the test, von Frey filament was applied to the plantar surface of the hind paw. Bend and hold the filament for 3 s. The researchers were blinded to the treatment group during behavioral testing. The threshold force required to retract the hind paw each time was measured and averaged (with a minimum interval of 20 min between measurements). For the Hargreaves test, infrared heat was applied to the plantar surface of the hind paw using a Hargreaves device (Ugo Basile) and the latency to exit the paw (thermal latency) was measured, using 20 s as the cutoff latency.
Thermal tests were repeated three times at 20-min intervals, and the mean value was calculated.

| Bioinformatic analysis of microRNA target genes
The following four databases were used to predict the putative target genes of miR-26a-5p: microT, 34 miRanda, PicTar, 35 and TargetScan. 36 Venn diagrams were used to screen the intersection of predicted target genes from four databases. DAVID Bioinformatics Resources 6.82 was used to analyze bioinformatics data for Gene Ontology enrichment of common target genes predicted by the four databases. 37 Statistical analysis and visualization were carried out in R version 3.6.3. R packages involved cluster Profiler package (for enrichment analysis and visualization) and ggplot2 package (for visualization).

| Dual-luciferase reporter assay
The luciferase assay was performed with a dual-luciferase reporter system (Promega). Briefly, wild-type (WT) or mutant (MUT) Wnt5a 3′-UTR reporter constructs were co-transfected into BV2 cells with the miR-26a-5p-mimic or negative control (NC), using lipofectamine 2000 (Invitrogen, USA), followed by the analysis of luciferase activities, in which Renilla was applied as a normalized control. The

| RNA Extraction and quantitative real-time polymerase chain reaction (qRT-PCR)
Total RNA isolated from BV2 cells, spinal cord tissue, and subcutaneous tissue of the hind paw using TRIzol reagent (Thermo Fisher Scientific) was used for the subsequent qPCR verification. In brief, total RNA samples were used for cDNA library preparation using a PrimeScript RT reagent Kit (TaKaRa). mRNA expression was determined by quantitative real-time polymerase chain reaction (qRT-PCR) using a SYBR Premix Ex TaqTM II (Tli RNaseH Plus) kit (TaKaRa).
qRT-PCR was performed on the ABI QuantStudio six flex (Applied Biosystems, United States). The PCR reaction was performed as follows: cycling conditions began with an initial DNA denaturation step at 95°C for 20 s, followed by 40 cycles at 94°C for 15 s, 56°C for 30 s, and 72°C for 25 s. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was the normalization for the quantification of gene expression, using the ΔΔCt method. The primers for mouse genes were synthesized by RiboBio (GuangZhou, China), and sequences can be found in Table 1.

| ELISA for the determination of cytokine levels
A cell suspension was prepared by inoculating BV2 microglial cells in the logarithmic growth phase into a 6-well culture plate. After BV2 cells were transfected with miR-26a-5p agomir for 48 h and treated with or without LPS (1 μg/mL) and Foxy5 (1 μg/mL) for 24 h, the supernatant was collected. Similarly, spinal cord L4-5 of mice were lysed after different treatments, and cytokine levels were determined. The concentrations of IL-1β, TNDα, TGFβ, and IL-10 levels were measured by ELISA according to the manufacturer's instructions (Proteintech). Optical density (OD) was measured at 450 nm using a microplate reader (Thermo Scientific).

| H&E staining
Twenty-four hours after intrathecal or subcutaneous injection, mouse hind paw subcutaneous tissues were taken for experiment.  Shapiro-Wilk normality test indicated that the data have normal distribution; therefore, comparisons were done using parametric tests.

| Statistical analysis
Differences were statistically significant when p value was <0.05.

| miR-26a-5p alleviates complete Freund's adjuvant-induced inflammatory pain and suppresses inflammation by a single intrathecal or subcutaneous injection
First, we adopted a mouse model of inflammatory pain by subcutaneously injecting 20 μL of CFA into the hind paw of mice. miR-26a-5p agomir was then injected to mice for 8 h after modeling by two methods of injection: intrathecal injection in subarachnoid space and subcutaneous injection in mice hind paw ( Figure 1A).
The results showed that, on the first day after modeling, both intrathecal injection and subcutaneous injection of miR-26a-5p agomir had a significant antinociceptive effect compared with the CFA modeling group ( Figure 1B,C). However, subcutaneous injection of miR-26a-5p agomir in the mice hind paw showed apparently weak analgesia than intrathecal injection via mechanic threshold and thermal latency test. Moreover, although the antinociceptive effect of subcutaneous injection was still existed on day 3 compared with the CFA model group, there was no significant statistical difference. Compared with subcutaneous injection, intrathecal injection still maintained a significant increase of mechanical threshold and thermal latency at the 7th day, although the analgesic effect has a marked decline from the 3rd day. These results suggested that miR-26a-5p agomir can produce a decent analgesia in a short term after a single intrathecal or subcutaneous injection, while intrathecal injection could produce a more potent and long-lasting analgesia effect than subcutaneous injection. We took the hind paw tissue of the mice one day after intrathecal or subcutaneous injection of miR-26a-5p agomir ( Figure 1A) to examine the changes of inflammationrelated cytokines after different administration methods. It can be clearly seen that both intrathecal and local subcutaneous injection of miR-26a-5p agomir and the levels of inflammatory cytokines IL-1β, IL-6, and TNFα in mice hind paws were significantly downregulated ( Figure 1D), while antiinflammatory cytokines IL-10 and TGFβ were significantly upregulated ( Figure 1E). In addition, intrathecal injection of miR-26a-5p agomir appears to have a more pronounced increase in antiinflammatory cytokines compared with local subcutaneous injection ( Figure 1E). These results suggest that miR-26a-5p can reduce CFA-induced inflammatory pain and suppress inflammation in mice paw tissue by a single intrathecal or subcutaneous injection.

| miR-26a-5p can significantly alleviate complete Freund's adjuvant-induced peripheral inflammation and neuroinflammation at the spinal cord level
Next, we performed HE staining on the paw tissue. The images of HE staining exhibited that the subcutaneous tissue of the hind paw has obvious inflammatory infiltration and tissue structure damage after modeling (

| miR-26a-5p inhibits the upregulation of Wnt5a/CaMKII/NFAT pathway after CFA-induced inflammatory pain
It has been previously reported that the noncanonical Wnt signaling pathway is partially mediated by Wnt5a and causes an elevation of downstream CaMKII expression, ultimately affecting the expression of the transcription factor NFAT. 42-45 Therefore, we examined the protein expression of Wnt5a, CaMKII, and NFAT at the spinal cord of different time points after CFA-induced inflammatory pain ( Figure 4A,B). It can be clearly seen that the expression of Wnt5a, CaMKII, and NFAT increased significantly on the first day and continued to the fifth day or even longer, and the changes of Wnt5a, CaMKII, and NFAT were basically follow the same style. Next, we also used different miR-26a-5p agomir administration methods to observe the expression changes of noncanonical Wnt signaling molecules Wnt5a, CaMKII, and NFAT. mRNA levels of Wnt5a, CaMKII, and NFAT were significantly decreased regardless of intrathecal or local subcutaneous injection ( Figure 4C-E). Consistent with the previous comparison of the two administration methods (Figure 1; Figure 2), intrathecal injection has a more significant downregulation effect on noncanonical Wnt signaling molecules. We then used Foxy5, a mimetic peptide of Wnt5a, to reverse the effect of miR-26a-5p and elucidate the relationship between Wnt5a, CaMKII, and NFAT. After intrathecal injection of miR-26a-5p agomir, the protein expression of Wnt5a, CaMKII, and NFAT decreased significantly( Figure 4F,G).
Then, we injected 10 μL Foxy5 intrathecally, and the expression of Wnt5a, CaMKII, and NFAT were reversed to varying degrees, and the three proteins both showed significant increase. After that, we injected miR-26a-5p agomir again intrathecally, Foxy5-induced upregulation of Wnt5a, CaMKII, and NFAT was rescued again. In conclusion, miR-26a-5p significantly inhibited the expression of Wnt5a, CaMKII, and NFAT in noncanonical Wnt signaling pathways after CFA-induced inflammatory pain.

| miR-26a-5p suppresses LPS-induced inflammation in BV2 cells in vitro through noncanonical Wnt signaling
Next, we used BV2 cells to verify the mechanism in vitro. with LPS for 24 h, the levels of IL-1β and TNFα in the supernatant were significantly upregulated, while the LPS/miR group was significantly decreased compared with the LPS/NC group. This decline was reversed 24 h after Foxy5 stimulation of BV2 cells ( Figure 5H).
In addition, we also found that the expression of antiinflammatory factors TGFβ and IL-10 was significantly increased in LPS/ miR group compared with LPS/NC group, and this effect was also reversed by Foxy5( Figure 5I). In summary, miR-26a-5p suppresses LPS-induced inflammation in BV2 cells in vitro through noncanonical Wnt signaling.

| DISCUSS ION
In current study, we found that miR-26a-5p had a significant alleviation effect on CFA-induced inflammatory pain. Our results found that both intrathecal and subcutaneous local injection of miR-26a-5p agomir could alleviate inflammatory pain via antiinflammation in spinal cord and hind paw. Intrathecal injection could produce longer pain relief period, which lasting to at least 7 days after miR-26a-5p injection. Moreover, both intrathecal and local application miR-26a-5p can significantly reduce the inflammatory factors level, alleviate inflammatory infiltration, and reverse the inflammation-induced subcutaneous tissue lesion of the mice hind paw. Furthermore, we found that two administration methods could alleviate activation of macrophages and microglia in subcutaneous tissue of paw and dorsal horn of the spinal cord, respectively. We also found that miR-26a-5p could reverse the M1/M2 proportion after LPS-stimulated BV2 cells.
Then, our in vitro and in vivo study suggested that miR-26a-5p could target Wnt5a to decrease the IL-1β, TNFα, and IL-6 through regulating CamK II and NAFT. Taken together, our study confirmed that miR-26a-5p could relief inflammation through intrathecal or subcutaneous local injection. We also demonstrated that the analgesia of intrathecal injection miR-26a-5p is produced by targeting Wnt5a, a noncanonical Wnt signaling pathway, to regulate the inflammation of spinal cord and hindpaw (Figure 7).
Myeloid cells, especially microglia and macrophages, is the pivotal cells during the pathophysiologic process of inflammatory pain, which is closely related with many potential mechanisms, including activation of ion channels, synaptic plasticity, and central sensitization, in the inflammatory pain onset, maintenance, and lasting. 46 Our results found that the activated cell number of microglia in spinal cord and macrophages in hindpaw have been significantly decreased by miR-26a-5p injection, which is consistent with previous reports. 30 In the current study, we adopted two different methods to inject miR-26a-5p, intrathecal injection and subcutaneous injection in hindpaw, which not only aim to compare the efficacy of different injection methods, but try to screen a better method which may help to avoid the potential complications of intrathecal injection such as bleeding, infection, 47 for future developing optimal therapy.
Although subcutaneous injection could reduce activation of spinal microglia, which may result from local injection of miR-26a-5p inhibited the inflammatory stimuli, the subcutaneous injection could only provide a one-day pain-relieving period. Thus, we selected the intrathecal injection in the following parts to investigate the potential mechanism of miR-26a-5p.
Spinal microglia are considered to be a significant central system resident immune cell that plays a crucial role in the inflammatory response. 48 The pathogenesis of inflammatory pain is known to be closely related to the release of pro-inflammatory mediators and activation of microglia. 14 It has been previously reported that microglia are key initiators of the development of inflammatory pain. 49 Activated microglia trigger the release of pro-inflammatory cytokines, such as IL-1β, IL-6, and TNFα, 50 which can further activate microglia. 51 In our cytes. 57 In neurological diseases, miR-26a-5p could target DYRK1A to inhibits Tau phosphorylation and Aβ accumulation, which ameliorated the cognitive dysfunction, in Alzheimer's disease mice. 58 More importantly, miR-26a-5p presented favorable antiinflammatory effect in several disease model. In cerebral ischemic/reperfusion brain injury, miR-26a-5p could target CDK6 to induce microglial apoptosis 59 and also could target NRSF to reduce neuroinflammation, 60 which both alleviated the brain injury. miR-26a-5p also regulates inflammation through other target genes. For example, miR-26a-5p alleviated acute lung injury by targeting TLR4 to suppress inflammation and apoptosis, 29 and it decreased inflammation in diabetic nephropathy by targeting CHAC1/NF-κB pathway. 61 In our study, we also found that miR-26a-5p significantly downregulated the inflammatory cyto- are preserved in BV2 cells. 63 However, although immortalized cells are easy to replicate and maintain in culture, their validity as sufficient substitutes for primary microglia is somewhat controversial. 64 Functionally, microglial cell lines share similarities with primary microglia, but differences exist in secretion and gene expression upon LPS stimulation. 65,66 In addition, although both BV2 cells and primary microglia express Iba1, a microglia marker, BV2 cells showed much less induction of some pro-inflammatory genes and much lower levels of cytokine secretion by LPS compared with primary microglia. 65,66 Therefore, the BV2 cell line has certain limitations as an in vitro model of microglia.
F I G U R E 7 miR-26a-5p alleviates CFA-induced chronic inflammatory hyperalgesia through Wnt5a/CaMKII/NFAT signaling in Mice. miR-26a-5p regulated Wnt5a/CaMKII/NFAT, a noncanonical Wnt signaling, involved in analgesia partly through antineuroinflammation, suggesting a pain-alleviating effect via noncanonical Wnt signaling pathway in CFA-induced inflammatory pain model in vivo and vitro.  77 In the current study, we adopted a similar LPS-induced myeloid cell lines to mimic the inflammatory situation as Pereia et. al 44 and then we also found that a deceased whole protein level of CaMKII is correlated with the pro-inflammatory cytokines. NFAT is initially present in nuclear extracts of activated T cells and is expressed in a variety of cells. 78,79 In the nervous system, NFAT is expressed in neurons and glial cells.
Neuronal NFAT isoforms are involved in the regulation of neuronal survival, apoptosis, and axonal outgrowth during development or injury. 80 NFAT is also expressed in primary microglia and leads to a pro-inflammatory response. 79,81 Previous studies have shown that loss of NFAT alters the expression of several genes, including Itgam, Tnf, Il-1b, and c-Myc. NFAT directly mediates the upregulation of TNFα and IL-1β. 82 Our study also found that after the noncanonical Wnt signaling pathway was inhibited, the levels of IL-1β, IL-6, and TNFα in the spinal cord decreased and at the same time attenuated the neuroinflammation induced by microglia. Furthermore, in the current study, we focused on the status and function of microglia during the inflammatory pain, but neurons also may involve in the analgesia of miR-26a-5p by targeting Wnt5a, which is shown in Figure 3E and still needs further study to investigate the related mechanism of Wnt5a in neurons and neuron-microglia interaction under inflammatory pain background. Taken together, Wnt5a and the noncanonical Wnt signaling pathway act as a vital role in inflammatory pain, which is a promising target for developing miR-26a-5p-related biomedicine.
In the current study, we also explored the administration route of miRNA. Our results showed that intrathecal injection has better antiinflammatory and biological effects than local subcutaneous administration. However, intrathecal drug delivery is often associated with complications such as bleeding, infection, and even death in severe cases. 47 Subcutaneous administration in the hind paw is less invasive and has fewer side effects for administration than intrathecal injection. When options are available, patients tend to opt for a less invasive approach to treatment. Therefore, how to improve the effectiveness of subcutaneous administration is a direction for our future research. We can see that subcutaneous administration of miR-26a-5p agomir in the hind paw can also inhibit the activation of microglia in the dorsal horn of the spinal cord, and intrathecal injection of miR-26a-5p agomir can also reduce the subcutaneous inflammation in the hind paw. The way in which the mutual regulation of central-peripheral inflammation is carried out is also worthy of our further study.

| CON CLUS IONS
In conclusion, we reported that miR-26a-5p is a promising candidate for CFA-induced inflammatory pain. We also compared the analgesic effect of intrathecal and subcutaneous injection of miR-26a-5p for inflammatory pain. Both administration methods could reduce the subcutaneous tissue and spinal cord inflammation. However, intrathecal injection produces longer analgesic period than local administration. Therefore, it still needs further research to develop an optimal administrate regimen for local injection, which may facilitate to avoid the adverse events result from intrathecal injection.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data and materials supporting the conclusions of this study are available from the corresponding author on reasonable request.