Inhibition of phosphodiesterase‐4 in the spinal dorsal horn ameliorates neuropathic pain via cAMP‐cytokine‐Cx43 signaling in mice

Abstract Background The spinal phosphodiesterase‐4 (PDE4) plays an important role in chronic pain. Inhibition of PDE4, an enzyme catalyzing the hydrolysis of cyclic adenosine monophosphate AMP (cAMP), produces potent antinociceptive activity. However, the antinociceptive mechanism remains largely unknown. Connexin43 (Cx43), a gap junction protein, has been shown to be involved in controlling pain transduction at the spinal level; restoration of Cx43 expression in spinal astrocytes to the normal levels reduces nerve injury‐induced pain. Here, we evaluate the novel mechanisms involving spinal cAMP‐Cx43 signaling by which PDE4 inhibitors produce antinociceptive activity. Methods First, we determined the effect of PDE4 inhibitors rolipram and roflumilast on partial sciatic nerve ligation (PSNL)‐induced mechanical hypersensitivity. Next, we observed the role of cAMP‐Cx43 signaling in the effect of PDE4 inhibitors on PSNL‐induced mechanical hypersensitivity. Results Single or repeated, intraperitoneal or intrathecal administration of rolipram or roflumilast significantly reduced mechanical hypersensitivity in mice following PSNL. In addition, repeated intrathecal treatment with either of PDE4 inhibitors reduced PSNL‐induced downregulation of cAMP and Cx43, and upregulation of proinflammatory cytokines tumor necrosis factor‐α (TNF‐α) and interleukin‐1β. Furthermore, the antinociceptive effects of PDE4 inhibitors were attenuated by the protein kinase A (PKA) inhibitor H89, TNF‐α, or Cx43 antagonist carbenoxolone. Finally, PSNL‐induced upregulation of PDE4B and PDE4D, especially the PDE4B subtype, was reduced by treatment with either of the PDE4 inhibitors. Conclusions The results suggest that the antinociceptive effect of PDE4 inhibitors is contributed by increasing Cx43 expression via cAMP‐PKA‐cytokine signaling in the spinal dorsal horn.


| INTRODUC TI ON
Neuropathic pain, a chronic nociceptive state that worsens the life quality of 7%-10% general population, 1 usually results from injury or diseases in the peripheral or central nervous system (CNS).
Neuropathic pain is one of the most difficult pain syndromes to manage, it is particularly necessary and important to investigate and develop novel analgesics. Neuropathic pain is contributed by the dysregulation of numerous cellular functions at the spinal cord level. 2,3 More specifically, activation of glia throughout the neuron system and the subsequent production of proinflammatory molecules from these cells in the spinal dorsal horn are crucial in the mediation of neuropathic pain. 2 Studies from our laboratory and other groups have demonstrated that proinflammatory cytokines, such as tumor necrosis factorα (TNFα) and interleukin-1β (IL-1β), downregulate the expression of connexin43 (Cx43), a transmembrane protein, in astrocytes. [4][5][6] Cx43 is highly expressed in spinal astrocytes and plays a pivotal role in the formation of gap junctions [7][8][9] and, accordingly, it is initially described as a gap junction α-1 protein (GJA1).
Recently, it has been demonstrated that spinal astrocytic Cx43 plays an important role in nociceptive transduction in the neuropathic pain state. 7,10 Specifically, spinal astrocytic Cx43 is downregulated in the ipsilateral lumbar spinal dorsal horn 7 days after partial sciatic nerve ligation (PSNL) and is restored following treatment of adenovirus vectors expressing Cx43, leading to amelioration of PSNL-induced mechanical hypersensitivity. 7 These results suggest that restoration of decreased spinal astrocytic Cx43 can be a potent therapeutic approach to the treatment of neuropathic pain.
Cyclic nucleotide phosphodiesterases (PDEs) consist of a large family of enzymes that catalyze the hydrolysis of the important second messengers cyclic adenosine monophosphate (cAMP) and/ or cyclic guanosine monophosphate (cGMP). PDEs play an essential role in regulating the intracellular concentrations of the cyclic nucleotides and in controlling their downstream signal transduction. 11,12 PDEs are a complex and diverse superfamily of more than 100 different protein products transcribed from 11 distinct but structurally related gene families (PDE1-11). 13,14 Among the 11 PDEs, PDE4 has been shown to be the major PDE family responsible for cAMP hydrolysis in nerve and immune cells; and inhibition of PDE4 produces antinociceptive and antiinflammatory effects in the CNS. 14,15 Since Cx43 is importantly regulated by cAMP signaling, 16 it is reasonable to believe that PDE4 inhibition may reduce neuropathic pain by modulating the expression of Cx43 in the spinal dorsal horn.
In the current study, we examined the potential role of PDE4 in Cx43 expression and pain-related behavior using PSNL in mice, a model of peripheral neuropathic pain. In addition, we investigated the possible regulatory role of PDE4-mediated cAMP-cytokine signaling in Cx43 expression. The results support a complex interaction between PDE4-cAMP signaling and Cx43 in the mediation of neuropathic pain.

| Partial sciatic nerve ligation
Under anesthesia with sodium pentobarbital (50 mg/kg, i.p.), mice were subjected to the PSNL surgery, in which a tight ligation of approximately one-third to one-half of the diameter of the left sciatic nerve (ipsilateral) was performed with 8-0 silk suture, as described in our previous study. 7 In sham (control) mice, the sciatic nerve was exposed without ligation using the same procedure. The success rate for PSNL operation was approximately 95%. Mice with PSNL that did not show robust mechanical hypersensitivity (hind paw withdrawal threshold >0.16 g) were excluded from the experiment.

| Intrathecal drug administration and testing schedule
Mice with PSNL were intrathecally injected with rolipram, roflumilast, or vehicle (5 µl) 14 days following surgery; mice with sham surgery were treated similarly with a vehicle as to the naive control. Withdrawal thresholds were measured 0. 5

| Hind paw sensitivity to mechanical stimulation
The withdrawal threshold (in grams) of the hind paw to mechanical stimulation was determined using von Frey filaments. 21 Briefly, mice were individually placed in the separate plastic box licking of the hind paw was defined as the withdrawal threshold, which was tested three times in 10-sec intervals, and the mean withdrawal threshold was calculated. Prior to drug treatment, mice were assessed for baseline withdrawal thresholds and then randomized into different treatment groups. All behavioral tests were performed with the observer blinded to the drug treatment.

| Western blotting
The lumbar (L4-L6) segments of the ipsilateral spinal dorsal horn were collected, immediately frozen in liquid nitrogen, and stored at −80°C until use. The spinal tissues were solubilized in ice-cold radioimmunoprecipitation assay buffer with protease inhibitors (100 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton ×-100, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate, and 1 mM phenylmethylsulfonyl fluoride) and phosphatase inhibitor cocktail 2 (Solarbio, Beijing, China)) with sonication. The lysates were centrifuged at 13,000 × g at 4°C for 10 min and the supernatant was added to Laemli's buffer and boiled for 5 min. Equal amounts of protein were separated by 7.5% or 10% sodium dodecyl sulfatepolyacrylamide gel electrophoresis and blotted onto nitrocellulose membranes. Non-specific binding was reduced with blocking buffer

| Enzyme-linked immunoassay assay
The spinal cord tissues were homogenized with ice-cold phosphatebuffered solution containing 1% phenyl methane sulfonyl fluoride.
Lysates were repeatedly thawed and refrozen three times and the supernatants were collected after centrifugation at 5000 ×g for 10 min. The contents of TNFα, IL-1β, IL-6, cAMP, and cGMP in the spinal cord tissues were determined using ELISA kits (Elabscience) following the instructions of the manufacturer; each sample was assessed in duplicates. The colorimetric reaction was conducted and the absorbance at 450 nm was recorded using a multifunctional microplate reader (TECAN). Protein concentrations of samples were determined using enhanced BCA protein assay kits (Solarbio) according to the manufacturer's instructions.

| Statistical analysis
All quantitative data were expressed as the means ± standard errors of the means (SEM). Comparisons of PDE4 protein levels after PSNL surgery were performed using student's t-test (for Figure 1B-E).
Comparisons between treatment groups and the corresponding control groups for mechanical hypersensitivity after PSNL (for Possible interaction between PDE4 inhibitor treatment and withdrawal thresholds following PSNL (for Figures 1A, Figure 2C and F) and between protein kinase A (PKA) or protein kinase G (PKG) inhibitor treatment and withdrawal thresholds following PSNL ( Figure 3D) were analyzed by two-way ANOVA, followed by the Tukey-Kramer method for post hoc comparisons. Differences were considered to be significant when the p-value was less than 0.05.

| PSNL-induced mechanical hypersensitivity and expression of PDE4 isoforms in the spinal dorsal horn
We used PSNL to establish the pain model. The withdrawal thresholds of the ipsilateral hind paw dropped from approximately 1.0 g to nearly 0 g in C57B6/LJ mice beginning 3 days after PSNL surgery; this mechanical hypersensitivity persisted up to 21 days following surgery (F (4, 36) = 35.37, p < 0.0001; Figure 1A). By contrast, no significant changes in withdrawal thresholds were observed following the sham operation.
To determine the involvement of PDE4 subtypes in mechani-

| PSNL-induced mechanical hypersensitivity was attenuated by single or repeated treatment with rolipram or roflumilast
To determine the role of PDE4 in neuropathic pain, we examined the effect of rolipram, a prototypic PDE4 inhibitor, and roflumilast, the first PDE4 inhibitor approved by the Food and Drug Administration for clinic use, 22 on PSNL-induced mechanical hypersensitivity following the schedule shown in Figure 2A.

| PDE4 inhibitors ameliorated PSNL-induced mechanical hypersensitivity via cAMP-PKAcytokines signaling
To determine the role of cAMP signaling in the effect of PDE4 inhibi-  Figure 4A, B), except for PSNL-induced upregulation of IL-6, which was not altered by the PDE4 inhibitors ( Figure 4C).
These data suggest that TNFα in the spinal dorsal horn is involved in the antinociceptive effect of PDE4 inhibitors.

| Effect of PDE4 inhibitors on expression of PDE4 subtypes in the spinal dorsal horn in PSNL mice
To identify the PDE4 subtypes mediated in the mechanical hypersensitivity, we examined the effect of rolipram and roflumilast on expression of specific PDE4s in the spinal dorsal horn in PSNL mice. All PDE4s but PDE4C were increased in the spinal dorsal horn 14 days following PSNL (p < 0.0001 for PDE4A, p < 0.01 for PDE4B and p < 0.001 for PDE4D; Figure 6A

| DISCUSS ION
In the present study, we determined the role of PDE4 in regulating neuropathic pain and its intracellular signaling mechanisms.
In addition, repeated treatment with the PDE4 inhibitors reversed PSNL-induced increases in PDE4 expression in the spinal dorsal horn. The results are supported by the finding that intrathecal administration of rolipram ameliorates bone cancer pain. 15 Consistent with the antinociceptive effect of PDE4 inhibition, knockdown of PDE4B by the intrathecal injection of PDE4B siRNAs attenuates L5 nerve ligation-induced nociceptive activity. 33 These data suggest that PDE4 is involved in the regulation of neuropathic pain; PDE4B may be one of the PDE4 isoforms mediating the antinociceptive effect of PDE4 inhibitors.
Rolipram down-regulates the expression and function of PDE4s. 34 Roflumilast also inhibits all the PDE4 subtypes to a similar
Further studies will be needed to verify this.
As observed in the current study, PSNL-induced decreases in cAMP, but not cGMP, were reversed by treatment with either rolipram or roflumilast. In addition, the antinociceptive effect of ro- showing that rolipram reduces bone cancer pain in rats via its antiinflammatory activity. 15 Since inhibition of PDE4 decreases the levels of cytokines and produces antinociceptive effects, 14,15 it is considered that proinflammatory cytokines contribute to PDE4-mediated pain. Of note, PSNL-induced upregulation of IL-6 was not changed by the PDE4 inhibitors, indicating that the response of proinflammatory cytokines to PDE4 inhibition may differ in different disease models.
Cx43 is an important mediator in neuropathic pain and astrocytic function. 10,43 It has been shown that astrocytic Cx43 expression is significantly decreased in the spinal dorsal horn of PSNL mice. 7 It is known that neural plasticity and neuron-glia interaction have been linked to the spinal machinery underlying the development of neuropathic pain. [43][44][45] Thus, the cell type in which the proposed machinery occurs is very important. Our recent study has demonstrated that the expression of Cx43 is limited in astrocytes, but not neurons or microglia. 7 However, this does not rule out the potential involvement of neurons and microglia, which will be further investigated in our future studies. And we did not test PKA activity in the present study because we have demonstrated that PDE4 inhibitors such as rolipram increase the activity and expression of PKA in the brain of mice. 46 In conclusion, we demonstrated a distinct mechanism of neuropathic pain that is regulated by PDE4-mediated intracellular signaling. PSNL induces expression of PDE4, most likely PDE4B, leading to decreases in cAMP levels and PKA activity. This causes increases in proinflammatory cytokines such as TNFα, and downregulation of Cx43 in the spinal dorsal horn, eventually resulting in neuropathic pain ( Figure 7A). By contrast, inhibition of PDE4, in particular PDE4B, activates cAMP-PKA signaling and suppresses TNFα, leading to increases in Cx43 in the spinal dorsal horn and eventual suppression of neuropathic pain ( Figure 7B). To the best of our knowledge, this is the first demonstration of the role of Cx43 in PDE4-mediated cAMP signaling in the regulation of neuropathic pain. The study provides valued clues on the mechanism whereby PDE4 inhibitors produce antinociceptive activity and will aid in the development of novel antinociceptive agents.

ACK N OWLED G M ENTS
This work was supported by grants from the Natural Science

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflicts of interest.