Phosphorylation at Ser 727 Increases STAT3 Interaction with PKCε Regulating Neuron–Glia Crosstalk via IL-6-Mediated Hyperalgesia In Vivo and In Vitro

Methods A rat hyperalgesia model was induced using an intraplantar injection of Freund's complete adjuvant (FCA) or an intrathecal injection of IL-6. Mechanical allodynia was evaluated using von Frey filament tests after intrathecal injections of T-5224 (c-Fos/AP-1 inhibitor), minocycline (Mino, a specific microglia inhibitor), L-2-aminoadipic acid (LAA, an astroglial toxin), PKCε inhibitor peptide, APTSTAT3-9R (STAT3 inhibitor), or anti-IL-6 antibody. The c-Fos, GFAP, Iba-1, PKCε, STAT3, pSTAT3Tyr705 and pSTAT3Ser727, and IL-6 expression at the spinal cord level was assessed by Western blot analysis. The interactive effects of PKCε and STAT3 were determined using immunofluorescence staining and immunoprecipitation in vivo and in vitro. Interleukin-6 promoter activity was examined using luciferase assays. Results T-5224, Mino, and LAA attenuated FCA- or IL-6-mediated inflammatory pain, with a decrease in c-Fos, GFAP, Iba-1, PKCε, and IL-6 expression. PKCε inhibitor peptide and APTSTAT3-9R reversed FCA-induced nociceptive behavior, while decreasing pSTAT3Ser727, IL-6, c-Fos, GFAP, and Iba-1 expression and PKCε and STAT3 coexpression. Interleukin-6 promoter activity increased in the presence of PKCε and STAT3. The interaction with PKCε increased on phosphorylating STAT3 at Ser727 but not at Tyr705. Conclusion STAT3 phosphorylation at Ser 727 and the interaction with PKCε contribute to hyperalgesia via the IL-6-mediated signaling pathway, thus regulating neuron–glia crosstalk during inflammatory pain.


Introduction
Inflammatory mediators play important roles in pain development by interfering with nociceptive cellular signal trans-duction and transmission. The proinflammatory cytokine interleukin-6 (IL-6) is secreted by astrocytes and microglia in the central nervous system [1,2]. It might play an important role in the development and maintenance of hyperalgesia in various pain models [3]. The upregulation of IL-6 in the spinal cord causes mechanical hyperalgesia in rats and is associated with the nociceptive sensory process [4][5][6]. The dysregulation of IL-6 results in the production and release of several inflammatory mediators that may activate neurocytes and trigger neuropathic pain [3,7]. The blockade of IL-6 signaling leads to substantial clinical improvement in inflammatory arthritis [8].
Protein kinase C (PKC) is an important family of intracellular signaling enzymes involved in central sensitization and pain transmission [9]. The PKC epsilon (PKCε) isoform is associated with the initiation of hyperalgesia by regulating nociceptor excitability [10,11]. Signal transducer and activator of transcription 3 (STAT3) modulates gene expression involved in promoting inflammatory pain. Notably, PKCε interacts with STAT3 [12], which targets and activates the IL-6 gene to increase IL-6 production [13,14]. However, whether the interaction between PKCε and STAT3 affects IL-6-mediated neuron-glia activation and hyperalgesia remains unknown.
In this study, we sought to explore the interaction between PKCε, STAT3, and IL-6 during inflammationinduced hyperalgesia. First, we investigated whether PKCε, STAT3, and IL-6 directly participated in neurocyte (including neurons, astrocytes, and microglia) activation during inflammatory pain in rats. Then, we aimed to determine the roles of PKCε and STAT3 in IL-6-induced hyperalgesia and neuron-glia crosstalk [15,16] in vitro. Finally, we examined the interactions between PKCε and STAT3 as well as their effects on IL-6 promoter activity in vivo and in vitro.

Animals.
A local experimental animal committee approved the experimental protocol, which was implemented according to the guidelines of the Institutional Animal Care and Use Committee (number 2019-004). Adult (6-to 8-week-old) male Sprague-Dawley rats weighing 200-220 g (n = 130) were acclimated for 1 week under a 12 h light/dark cycle at 22°C ± 2°C and 55% ± 5% relative humidity and received food and water ad libitum.

Inflammatory Pain and IL-6-Induced Hyperalgesia
Model. Freund's complete adjuvant (FCA; Sigma-Aldrich Corp.) consisting of heat-killed Mycobacterium tuberculosis (1 mg/mL) in paraffin oil (150 μL) was injected into the plantar area of the right hind paws of the rats under 1%-3% isoflurane anesthesia delivered with oxygen at a flow rate of 1 L/min. The left hind paw was not injected. Physical signs (e.g., redness and swelling) and pain behavior were monitored for 24 h after FCA injection. A model of hyperalgesia was established using naïve rats given an intrathecal injection of IL-6 (20 ng/50 μL) through the L5-L6 lumbar interspace identified by the tail-flick reflex 10 min before drug administration under isoflurane anesthesia.

Experiments
In Vivo. Figure 1 shows the experimental protocol in vivo.
The paw withdrawal mechanical threshold (PWMT) was measured daily after the intraplantar injection of FCA ( Figure 1). All agents described earlier were injected intrathecally on days 4 and 6 after the intraplantar injection of FCA. The spinal cords were harvested on day 7. In the IL-6-induced hyperalgesia model, the pain thresholds after chemical administration were assessed as the PWMT up to 120 min after IL-6 administration.
2.5. Von Frey Filament Tests. We measured PWMT using von Frey filament tests (Stoelting Co., IL, USA) using the up-and-down method [18]. The rats were habituated to a wire mesh platform for at least 1 h/d for 3 days before starting experiments. All groups (n = 6 each) were tested daily 2 Mediators of Inflammation before drug application to determine baseline levels. Briefly, the positive and negative data were tabulated as follows: X = withdrawal and 0 = no withdrawal. The 50% response threshold was interpolated using the following formula: 50 %g threshold = ð10 ðxf +kδÞ Þ/10,000, where xf is the last value (in log units) of the von Frey filament test, k is the tabular value for positive/negative responses, and δ is the mean difference (in log units) between stimuli [18]. The PWMT was defined as the means of six animals before and after chemical injections. Areas under receiver operator characteristic curves (AUC) were calculated to determine the effects of the injected chemicals.
2.6. Western Blot Analysis. The rats were euthanized using isoflurane anesthesia after the aforementioned tests. The lumbar segments (L3-L5) of whole spinal cords (n = 6 /group) were homogenized and centrifuged at 16,000 g and 4°C for 15 min. Equal amounts of protein (50 μg) in supernatants were denatured, resolved by sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE, 5% stacking gel, and 10%-12% separating gel), and then electrotransferred onto polyvinylidene difluoride (PVDF) membranes (Millipore, Massachusetts, USA). Nonspecific protein binding on the membranes was blocked with 5% nonfat milk in tri-buffered saline and Tween 20 (TBST) for 2 h at room temperature and then incubated with the primary antibodies: rabbit anti-c-Fos, anti-Iba-1, anti-PKCε, anti-STAT3 (1 : 1000; Affinity Biosciences, OH, USA), anti-IL-6 (1 : 1000; Abcam, MA, USA), anti-GFAP, and anti-GAPDH (1 : 1000; Cell Signaling Technology, MA, USA). The blots were washed with TBST and then probed with secondary HRP goat anti-rabbit IgG (1 : 10,000) in 5% nonfat milk in TBST for 1 h at room temperature. The proteins of interest on the blots were visualized using an ECL reagent (Affinity Biosciences) and photographed using an X-ray film. The protein band intensity was quantified using the ImageJ software (National Institutes of Health, MD, USA).

Immunofluorescence Staining.
After transcardial perfusion 7 days after FCA administration with 4% paraformaldehyde, the spinal cords were removed from the rats, postfixed, and dehydrated. Transverse frozen sections (10 μm) prepared from OCT-embedded tissues were incubated overnight with rabbit polyclonal anti-PKCε (Abcam) and mouse monoclonal anti-STAT3 (Cell Signaling Technology). The proteins in the tissues were visualized using a DMIL LED scanning microscope (Leica Microsystems GmbH, Wetzlar, Germany). Primary or secondary antibodies were omitted to ensure staining specificity. The data from three to four sections per rat (n = 6/group) were analyzed.    Mediators of Inflammation 2.8. Immunoprecipitation Assays. Naïve rats were euthanized under isoflurane anesthesia. The lumbar segments L3-L5 from whole spinal cords were ultrasonicated in ice-cold RIPA buffer (Beyotime, Shanghai, China). The supernatant after low-speed centrifugation was diluted with lysis buffer and adjusted to 2 mg/mL. The cell suspensions passed through agarose resin served as controls. The sonicates were incubated overnight at 4°C with PKCε and IgG primary antibody (5 μL) to form immune complexes that were captured with fresh elution buffer and then incubated at 95°C for 6-8 min in RIPA buffer. The mixture was centrifuged, and then, endogenous PKCε/STAT3 immune complexes in the supernatant were analyzed by Western blot analysis.
Extracts of transfected cells were quantified using BCA assay kits (Beyotime). Supernatants were diluted with lysis buffer and adjusted to 2 mg/mL of protein; then, 500 μL portions were incubated overnight with 20 μL of anti-Flag magnetic beads (Beyotime) at 4°C. Immune complexes with magnetic beads were washed three times with lysis buffer, precipitated, then eluted from the beads by boiling with 30 μL of SDS-PAGE loading buffer for 6-8 min. The eluate was centrifuged; then, proteins in the supernatant were analyzed by western blotting.
2.11. Statistical Analysis. All data were presented as means ± standard deviation (SD) and analyzed by one-way analysis of variance (ANOVA) followed by the Bonferroni tests using SPSS 19.0 (IBM Corp., NY, USA). The graphs were constructed using the GraphPad Prism 7.0 software (GraphPad Software Inc., CA, USA). The statistical significance was set at P < 0:05.

Mediators of Inflammation
When these three conditions were applied to rats with IL-6-induced hyperalgesia, only anti-IL-6 improved the hypersensitivity and response to pain in the right paw (Figure 4(d)). The AUC of the PKCε inhibitor peptide, APTSTAT3-9R, and anti-IL-6 antibody was 0:64 ± 0:08, 0:64 ± 0:08, and 0:95 ± 0:03 in the right hind paw under IL-6-mediated hyperalgesia, respectively (Figure 4(e)). The analgesic effect of anti-IL-6 was similar in the left hind paw (Figure 4(f)). Endogenous PKCε/STAT3 immunocomplexes in spinal cord tissues were assessed (Figure 7(a)). After incubating HEK293 cells with lipopolysaccharide (LPS), STAT3 increased IL-6 promoter activity, which was also enhanced in the presence of PKCε (Figure 7(b)). The STAT3 Y727D phosphomimetic mutant had more affinity for PKCε, whereas the other mutants generated results similar to those of wild-type STAT3, indicating that the phosphorylation at Ser727 increased the ability of STAT3 to bind to PKCε (Figures 7(c) and 7(d)).

Discussion
Our results showed that FCA-induced inflammatory pain and IL-6-induced hyperalgesia were alleviated by inhibiting neurocyte activation or by anti-IL-6 therapy, indicating that IL-6 participated in the maintenance of inflammationinduced nociception. This study was novel in demonstrating that T-5224, Mino, and LAA inhibited FCA-induced inflammatory pain and IL-6-induced hyperalgesia, despite previous findings of the therapeutic effects of Mino against chronic bone cancer pain and chronic pain [19,20].

Mediators of Inflammation
The present study used an inflammatory pain model created by unilateral injections of FCA or intrathecal injections of IL-6 that induced hyperalgesia in rats. The mechanical paw withdrawal threshold was significantly reduced for up to 7 days by FCA and up to 60 min by IL-6. An intraplantar injection of FCA induced central sensitization and increased the levels of pain mediators, including IL-6, peripherally and centrally, whereas the intrathecal injection of IL-6 likely resulted in transient central sensitization due to a local increase in pain mediators per se [6,21].
We found that T-5224, Mino, and LAA decreased the expression levels of c-Fos, GFAP, Iba-1, PKCε, and IL-6 but did not alter STAT3 levels during the FCA-induced inflammatory process (Figure 3). Our results indicated that the inhibition of neurocyte activation reduced IL-6mediated pain sensitivity, which was in line with previous findings of the crosstalk between activated neurocytes and IL-6-induced pain [3,6,16,22,23]. However, the analgesic effects of T-5224, Mino, and LAA were associated with other cytokines such as IL-1β and TNF-α [24], suggesting that activated neurocytes comprised a control mechanism of inflammatory pain [25].
We further examined the roles of PKCε, STAT3, and IL-6 in FCA-induced inflammatory pain and IL-6-induced hyperalgesia. Anti-IL-6 immediately alleviated inflammatory pain for an extended period and reversed the hyperalgesic effects of IL-6 ( Figure 4), suggesting that IL-6 was a potent pain mediator. In contrast, the PKCε inhibitor peptide and  Figure 3: Western blot analysis of the proteins related to activated neurocytes during inflammatory pain. Examples (a) and mean values (bg) of c-Fos, GFAP, Iba-1, PKCε, STAT3, and IL-6 proteins in the spinal cord. Intrathecally injected T-5224 (c-Fos/AP-1 inhibitor, 500 μg/ 50 μL), minocycline (Mino, 100 μg/50 μL), and L-2-aminoadipic acid (LAA, 1 mg/50 μL) reduced c-Fos, GFAP, Iba-1, PKCε, and IL-6 protein levels in FCA-treated rats, whereas STAT3 expression did not change between groups (P > 0:05). Data were normalized against GAPDH and are expressed as ratios (%) of control. Data are shown as means ± SD (n = 4-5). * P < 0:05, * * P < 0:01; * * * P < 0:001, oneway ANOVA with Bonferroni tests. 8 Mediators of Inflammation 9 Mediators of Inflammation the STAT3 inhibitor did not exert immediate analgesic effects but elicited a mild, sustained analgesic effect for 5-7 days at the end of the experimental period. This phenomenon was accompanied by a decrease in the pSTAT3 Ser727 level but not in the levels of pSTAT3 Tyr705 , IL-6, c-Fos, GFAP, and Iba-1 in the spinal cord, indicating that the upstream effectors PKCε and STAT3 regulated the formation of IL-6, thus mediating the activation of neurocytes (Figures 5 and 8). These findings were comparable with the association of PKCε and STAT3 levels with the IL-6 level and neurocytes [26][27][28][29][30]. PKCε and STAT3 regulated the production of IL-6, and pSTAT3 Ser727 formed complexes with PKCε and enhanced STAT3 localization to the IL-6 promoter, thus increasing IL-6 expression [31]. The decrease in the levels of pSTAT3 and IL-6 in cells incubated with APTSTAT3-9R and anti-IL-6 suggested that STAT3 was involved in a negative feedback loop in the IL-6-induced signaling pathway per se.
PKCε and STAT3 interacted under physiological and pathological conditions [11,31]. Our immunofluorescence results also revealed the coexpression of PKCε and STAT3 in cells in the spinal cord (Figure 6), suggesting that PKCε together with STAT3 contributed to IL-6 production and the activation of neurons, astrocytes, and microglia during inflammation. Alone, PKCε did not affect the activity of the IL-6 promoter; but IL-6 promoter activity was increased more in the presence of both PKCε and STAT3 than that in the presence of STAT3 alone under LPS stimulation (Figure 7), suggesting that PKCε increased the ability of STAT3 to bind to the IL-6 promoter. The immunoprecipitation results in vitro showed that the pSTAT3 Ser727 but not pSTAT3 Tyr705 affected the interactions between PKCε and STAT3 (Figure 7), suggesting that pSTAT3 Ser727 regulated the formation of PKCε/STAT3 complexes, thus influencing IL-6-mediated inflammatory pain.
Given the analgesic effect of T-5224, Mino, and LAA, they may serve as the potential therapeutic agents for inflammatory pain-related disease. This needs further discussion to understand the potential pharmacological characteristics of these compounds. T-5224 was first designed to inhibit the arthritis upstream of inflammatory cytokine and matrix metalloproteinase action [32]. It can be used in human articular chondrocytes, resulting in the inhibition of transactivation of downstream matrix metalloproteinases and inflammatory cytokines (including IL-6, IL-1β, and TNF-α) and effectively preventing cartilage destruction and osteophyte formation in an osteoarthritisinduced mouse model [33]. T-5224 was also found to be

11
Mediators of Inflammation used in acute myeloid leukemia [34], mast cell [35], and triple negative breast cancer [36], which might serve as a synergistic therapeutic strategy for the clinical diseases; however, T-5224 is still in the preclinical stage. Hence, the data related to pharmacokinetics are lacking; more research and evidence are needed in the future. Mino is a semisynthetic tetracycline antibiotic with anti-inflammatory properties, which is used to treat multiple inflammatory diseases and could be safely applied in the clinical setting, such as Parkinson's disease [37] and neurodegenerative and psychiatric diseases [38] as well as the cerebral ische-mia [39]. It is generally well tolerated, and skin-related complaints, nausea, and dizziness are the most common patient-reported side effects [40]. LAA, a selective astrocytic toxin, has been demonstrated to exert some regulatory effects on tibia fracture [41], myotubes [42], and retina [43], thus contributing to the fracture-induced nociceptive, cell autophagy in myotubes, and retinal vascular responses. It has not yet been clinically applied due to its unusual astroglial toxin, which may trigger locomotor network damage. How to reduce the toxicity of LAA to central and peripheral nerves to the minimum is worth exploring [44]. Values were normalized against GAPDH and are expressed as ratios (%) of control values. Data are shown as means ± SD (n = 4-5). * P < 0:05, * * P < 0:01, * * * P < 0:001; one-way ANOVA followed by Bonferroni tests. 12 Mediators of Inflammation  Data are shown as means ± SD (n = 6). * P < 0:05, * * P < 0:01, * * * P < 0:001; one-way ANOVA followed by Bonferroni tests.

Mediators of Inflammation
This study had some limitations. It focused only on IL-6-induced pain, and thus, its clinical relevance is debatable. However, early and delayed IL-6 elevation is associated with chronic neuropathic pain [45]. Interleukin-6 plays a key role in the chronic inflammation associated with rheumatoid arthritis (RA), and blocking IL-6 signaling is an important strategy in treating RA-associated dis-eases clinically [46]. Moreover, targeting IL-6 might be an option for treating other chronic inflammatory diseases [47]. Therefore, an in-depth understanding of how IL-6 induces cellular signaling that causes pain, and the development of new analgesic strategies associated with IL-6, have theoretical and clinical significance for pain management. Interleukin-6 promoter activity increased by STAT3 was further enhanced by PKCε and STAT3, indicating that PKCε improved the ability of STAT3 to bind to IL-6 promoter. Data are shown as means ± SD (n = 6). * P < 0:05, * * P < 0:01, * * * P < 0:001; one-way ANOVA followed by Bonferroni tests. (c, d) HKE293 cells were transfected with GFP, GFP-PKCε, Flag, Flag-STAT3, and phosphomimetic and dephosphomimetic mutants of STAT3, and then, immunoprecipitants were assayed. Protein complexes were detected using an anti-GFP antibody (c), and then, relative PKCε binding to STAT3 was quantified (d). Data are presented as means ± SD (n = 3). * P < 0:05, * * P < 0:01, * * * P < 0:001; one-way ANOVA followed by Bonferroni tests.
In this study, we found that the phosphorylation at Ser727 increased STAT3 interaction with PKCε. This increased IL-6 promoter activity and upregulated IL-6 expression, thus enhancing neuron-glia activation during the development of inflammatory pain. In addition, the PKCε inhibitor peptide and STAT3 inhibitor (APTSTAT3-9R) attenuated FCA-induced nociceptive behavior via IL-6 downregulation ( Figure 8).

Conclusions
In summary, pSTAT3 Ser727 interaction with PKCε contributes to FCA-induced inflammatory pain and IL-6mediated hyperalgesia via IL-6-modulating crosstalk among neurons, astrocytes, and microglia and their activation. The translational value of our findings warrants further investigation.

Data Availability
The data used to support the findings of this study are available from the corresponding upon request.

Conflicts of Interest
The authors have no competing interests to declare.