Hypersensitivity of Prelimbic Cortex Neurons Contributes to Aggravated Nociceptive Responses in Rats With Experience of Chronic Inflammatory Pain

Previous experience of chronic pain causes enhanced responses to upcoming noxious events in both humans and animals, but the underlying mechanisms remain unclear. In the present study, we found that rats with complete Freund’s adjuvant (CFA)-induced chronic inflammatory pain experience exhibited aggravated pain responses to later formalin test. Enhanced neuronal activation upon formalin assaults and increased phosphorylated cAMP-response element binding protein (CREB) were observed in the prelimbic cortex (PL) of rats with chronic inflammatory pain experience, and inhibiting PL neuronal activities reversed the aggravated pain. Inflammatory pain experience induced persistent p38 mitogen-activated protein kinase (MAPK; p38) but not extracellular regulated protein kinase (ERK) or c-Jun N-terminal kinase (JNK) hyperphosphorylation in the PL. Inhibiting the p38 phosphorylation in PL reversed the aggravated nociceptive responses to formalin test and down-regulated enhanced phosphorylated CREB in the PL. Chemogenetics identified PL–periaqueductal gray (PAG) but not PL–nucleus accumbens (NAc) as a key pathway in inducing the aggravated formalin pain. Our results demonstrate that persistent hyperphosphorylation of p38 in the PL underlies aggravated nociceptive responses in rats with chronic inflammatory pain experience.


INTRODUCTION
Chronic pain is one of the most prevalent clinical situations. The experience of chronic pain affects physiological states of the individual, even after the pain has perceptually recovered. A subject with chronic pain experience frequently shows enhanced responses to following noxious events, reflected in lower pain thresholds and increased pain ratings (Bachiocco et al., 1993;Lidow, 2002;Ren et al., 2004;Hermann et al., 2006;Wegner et al., 2015). In animal studies, enhanced formalinevoked pain behaviors are also observed in adult rats with chronic inflammatory pain experience (Li et al., 2012). However, mechanisms underlying deteriorated pain responses following previous pain experience are not well illustrated.
Based on the findings above, we hypothesize that persistent functional alterations in the PL facilitate nociceptive responses to subsequent noxious exposure, even after the perceptual recovery of chronic pain. To test this hypothesis, we examined molecular changes of the PL in rats with experience of CFA-induced chronic inflammatory pain, and observed sustained PL hyperreactivity which mediated the vulnerability to formalin pain test in these rats.

Animals
Adult male Sprague-Dawley rats (230-250 g at the beginning of experiments) were provided by the Department of Laboratory Animal Sciences, Peking University Health Science Center (Beijing, China). All animals were housed in standard cages with a 12-h alternating light/dark cycle and food and water available ad libitum. All experimental procedures were approved by the Animal Care and Use Committee of our University, according to the guidelines of the International Association for the Study of Pain. By the end of the experiment, euthanasia was performed with 1% pentobarbital sodium (1 ml/100 g, i.p.).

Establishment of CFA-Induced Inflammatory Pain Model of Rats
Following our previous protocol (Yue et al., 2017;, the rat was anesthetized with isoflurane. The plantar surface of left hindpaw was cleaned by 75% ethanol before a total of 100 µL CFA was injected intraplantarly. For control, an equal volume of normal saline was injected.

Measurement of Thermal and Mechanical Pain Thresholds
The rat was handled for 10 min, and adapted in a plexiglas box for 30 min per day for three consecutive days before the first measurement. Thermal or mechanical pain thresholds were measured as previously described (Chaplan et al., 1994;Zhang et al., 2016) while the rat was calm and awake. Paw withdrawal latencies (PWLs) to thermal stimuli were measured by a focused radiant heat (40 W of power) applied to either hindpaw (Hargreaves Method, IITC 390). PWLs were recorded three times and averaged as the thermal pain threshold. A cut-off value of 30 s was set to avoid any possible tissue injuries.
Fifty percent paw withdrawal thresholds (50% PWTs) to mechanical stimuli were measured by von Frey hairs (0.41-15.1 g; North Coast, Gilroy, CA, USA). The von Frey hair was applied to the central plantar surface of either hindpaw. The 50% PWTs were calculated by the ''up and down'' method as described by Chaplan et al. and in our lab (Chaplan et al., 1994;Zhang et al., 2016). Thermal hyperalgesia and mechanical allodynia were measured 1 day before and 1, 3, 7, 14, 21 and 28 days after CFA injection.

Hot Plate Test
The rat was handled for 10 min and adapted in the hot plate for 10 min per day for three consecutive days before the test. Rats were placed individually onto the center of the hot plate (49 • C) and the latency of the first sign of hind paw licking or jumping to avoid heating pain was recorded as an index of the nociceptive threshold (Luo et al., 2004;Yu et al., 2008). A cut-off value of 30 s was set to avoid any possible tissue injuries. The hot plate was cleaned by 75% ethanol between tests.

Open Field Test
The rat was placed in a 100 × 100 × 50 cm box exposed to 50 lux illumination, with its activities videotaped for 10 min . Time spent (C.Time) and distance traveled (C.Dis) in the central area (60 × 60 cm), and total distance traveled (T.Dis) in the field were measured using the SMART software (v2.5.21, Panlab, Harvard Apparatus). The box was cleaned by 75% ethanol between tests.

Elevated Plus-Maze Test
The elevated plus-maze test was carried out on the next day of the open field test Zhang M. et al., 2017). The maze was placed 50 cm above the floor in a 30 lux illuminated room and consisted of two open arms and two closed arms (48 × 8 cm, and 40 cm wall height for the closed arms). The rat was placed onto the center area, heading toward the same open arm, and videotaped in the following 10 min. Time spent (O.Time) and numbers of entries (O.Entries) into open arms and total arm entries (T.Entries) were analyzed. The maze was cleaned by 75% ethanol between tests.

Formalin Test
Formalin test was performed 30 days after the CFA injection. Rats were handled for 10 min and adapted in a plexiglas chamber for 20 min per day for 3 days before test. The rat received an injection of 100 µL of 5% standard formalin solution into the plantar surface of right hindpaw (the opposite hindpaw of CFA/saline injection), with its behavior videotaped in the following 60 min. Time spent on licking and lifting the formalin injected paw were counted, and the formalin pain score was calculated as previously described: (time lifting + 2 × time licking)/total time Zhang et al., 2014). The chamber was cleaned by 75% ethanol between tests.
Brains were removed 45 min after formalin injection for Western blotting and immunostaining, or 90 min after formalin injection for c-Fos immunostaining.
To test anxiety-like behaviors of rats after the formalin injection, open field and elevated plus-maze tests were performed 1 and 2 days after formalin injection, respectively.

Cannula Implantation and Drug Microinjection
The rat was anesthetized deeply with 1% sodium pentobarbital (0.5 ml/100 g, i.p.) and positioned in a stereotaxic frame (RWD, Shenzhen, China). A guide cannula (O.D. 0.48 mm/I.D. 0.34 mm, C.C 1.2 mm, RWD, Shenzhen, China) was implanted 1.5 mm above PL [anterior-posterior (AP) + 2.9 mm; medial lateral (ML) ± 0.6 mm from Bregma; dorsal-ventral (DV) −2.5 mm from brain surface] . Four skull screws were used for securing the guide cannula to the skull surface with dental acrylic. The matching cap (0.5 mm below the guide cannula, RWD, Shenzhen, China) was inserted into the guide cannula. All animals were given at least 1 week for recovery from surgery before further experiments. The injection needle (1.5 mm below the guide cannula, RWD, Shenzhen, China) was used for microinjection with a polyethylene catheter connecting a micro-syringe. GABA A R agonist muscimol (1 µg/µl, 0.5 µl/side, Tocris Bioscience), p38 inhibitor SB203580 (1 µg/µl, 0.5 µl/side, Sigma-Aldrich) or vehicle (artificial cerebrospinal fluid, aCSF, 0.5 µl/side) was injected into PL of either side over 2 min. The injection needle was held on for at least 2 min to allow drug diffusion. The behavioral tests were performed 30 min after drug/vehicle injection. Rats with incorrect site of the guide cannula were excluded from analysis.

Stereotaxic Microinjection of Adeno-Associated Virus (AAV) Vectors Into PL
AAV5-CaMKIIα-hM4D(Gi)-mCherry and AAV5-CaMKIIα-mCherry viruses were packaged and purchased from the University of North Carolina Vector Core Facilities . AAV virus solution was microinjected into PL (AP +2.7/3.2 mm; ML ± 0.6 mm from Bregma; DV −2.5 mm from brain surface) with 0.5 µl/hole, 2 holes/side, at a speed of 0.1 µl/min after being anesthetized with 1% sodium pentobarbital (0.5 ml/100 g, i.p.). The needle was kept on the site for 3 min to allow for virus diffusion and gradually withdrawn over 1 min to prevent possible leakage from the needle track. The behavioral tests were performed 6 weeks after virus injection.

Statistical Analysis
Data were presented as means ± SEM. Unpaired or paired two-tailed t tests and one-way analysis of variance (ANOVA) with Bonferroni post hoc tests were used for the comparison Frontiers in Molecular Neuroscience | www.frontiersin.org of two groups. Comparisons of two groups with different time points were performed using two-way ANOVA or ANOVA with repeated measures and Bonferroni post hoc test. The differences were calculated with software GraphPad Prism 5.0 and statistical significance was defined as p < 0.05.
Frontiers in Molecular Neuroscience | www.frontiersin.org FIGURE 3 | Inhibiting PL reverses aggravated formalin pain in rats with chronic pain experience. Inhibiting PL activities by muscimol relieved the aggravated phase II formalin pain in CFA (the middle column), but not saline (the right column) group. No significant differences in phase I of formalin pain were observed among groups (the left column). n = 8 in each group. In the middle column: * * * p < 0.001, CFA+aCSF vs. CFA+Muscimol. * * p < 0.01, CFA+aCSF vs. Saline+aCSF, one-way ANOVA. In the right column: * p < 0.05, * * p < 0.01, * * * p < 0.001, CFA+aCSF vs. CFA+Muscimol, ANOVA with repeated measures and Bonferroni post hoc test.

Enhanced Activation of PL Neurons Upon Formalin Assaults in Rats With Chronic Pain Experience
To examine whether PL participated in the observed aggravated nociceptive responses, we first performed c-Fos protein mapping after formalin injection. After saline injection, PL showed little c-Fos protein expression in rats with or without chronic pain experience. By contrast, formalin injection induced significant c-Fos expression in PL, especially in rats with chronic pain experience (F (3,85) = 159.5, p < 0.001; Figures 2A,B). The majority (93.70 ± 1.53%) of c-Fos positive neurons in PL co-labeled with EAAC1, a marker of excitatory glutamatergic neurons ( Figure 2C).
Meanwhile, we observed significantly up-regulated phosphorylation of CREB, a marker of neuronal activation, in PL of rats receiving formalin injection with chronic pain experience. However, formalin alone was not sufficient to elevate the content of p-CREB in the PL (F (3,26) = 15.45, p < 0.001; Figure 2D). These findings indicate stronger activation of PL by formalin pain in rats with chronic pain experience than in those without.

Inhibiting PL Reverses Aggravated Formalin Pain in Rats With Chronic Pain Experience
To examine whether PL contributed to aggravated nociceptive responses, we injected muscimol, a GABA A receptor agonist, into PL before the formalin test. Muscimol significantly relieved formalin-induced pain behaviors in rats with chronic pain experience, but not in those without (phase I: F (3,33) = 1.35, p > 0.05; phase II: F (3,33) = 11.62, p < 0.001; right: group effect: F (3,330) = 11.32, p < 0.001; time effect: F (11,330) = 55.07, p < 0.001; interaction: F (33,330) = 2.47, p < 0.001; Figure 3). These results indicate that PL contributes to aggravated nociceptive FIGURE 4 | Persistent hyperphosphorylation of p38 accompanies aggravated formalin pain in rats with chronic pain experience. (A) Increased PL p-p38 after formalin injection in rats with chronic pain experience. Representative Western blots of p-p38, p38 and β-tubulin were shown above the corresponding histogram. n = 7 in each group. * p < 0.05, CFA+Formalin vs. Saline+Formalin, t test. (B) p-p38 in the PL increased in rats with chronic pain experience, with or without formalin injection. Formalin injection alone did not obviously influence the expression of p-p38 in PL. Representative Western blots of p-p38, p38 and β-tubulin were shown above the corresponding histogram. n = 8 in each group. * * * p < 0.001, CFA vs. Saline, * p < 0.05, CFA+Formalin vs. Saline+Formalin, one-way ANOVA.
formalin pain responses in rats with chronic inflammatory pain experience.
These results suggest that persistent hyperphosphorylation of p38 in the PL contributes to the hypersensitized formalin pain in rats with chronic pain experience.
These findings indicate that PL-mediated aggravated formalin pain in rats with chronic pain experience is achieved through the PL-PAG pathway, but not the PL-NAc core pathway.

PL Contributes to Aggravated Nociceptive Formalin Pain Responses in Rats With Chronic Inflammatory Pain Experience
Prior pain experience alters future responses to painful stimuli in both human and animals (Bachiocco et al., 1993;Lidow, 2002;Ren et al., 2004;Hermann et al., 2006;Li et al., 2012;Wegner et al., 2015). In the present study, we identified PL as a key region in regulating the aggravated formalin-induced pain in rats after the recovery from CFA-induced chronic inflammatory pain. Rats SB203580/aCSF micro-injection and formalin test were performed 28 d after CFA/saline injection. The cannula implantation surgery was performed 2 weeks before behavior tests. (B,C) SB203580 micro-injection into PL inhibited the hyperphosphorylation of p38 in rats with chronic pain experience, without (B) or with formalin injection (C). Representative Western blots of p-p38, p38 and β-tubulin were shown above the corresponding histogram. n = 6 in each group. * * * p < 0.001, CFA+aCSF vs. CFA+SB203580. * * p < 0.01, CFA+aCSF+Formalin vs. CFA+SB203580+Formalin, t test. (D) SB203580 micro-injection into PL inhibited the hyperphosphorylation of cAMP-response element binding protein (CREB) after formalin injection in rats with chronic pain experience. Representative Western blots of p-CREB, CREB and β-tubulin were shown above the corresponding histogram. n = 6 in each group. * p < 0.05, CFA+aCSF+Formalin vs. CFA+SB203580+Formalin, t test. (E) Inhibiting the phosphorylation of p38 in PL by SB203580 reversed the aggravated phases II formalin pain in the CFA group, but not the saline group (the middle column). Detailed pain scores shown in every 5 min (the right column). n = 6 in each group. In the middle column: * * * p < 0.001, Saline + aCSF/CFA+SB203580 vs. CFA+aCSF, one-way ANOVA. In the right column: * p < 0.05, * * * p < 0.001, CFA+aCSF vs. CFA+SB203580, ANOVA with repeated measures and Bonferroni post hoc test.
with chronic pain experience showed elevated pain scores in the second, but not the first, phase of formalin pain. The second phase of formalin pain reflects the development of inflammation and central sensitization, and indicates the involvement of the central nervous system in the aggravated pain (Shibata et al., 1989;Vaccarino and Melzack, 1992).
Structural and functional changes in several brain regions persist long after pain recovery. For example, reduced gray matter densities in chronic back pain patients are observed in the middle cingulate gyrus, thalamus and prefrontal cortex (Ivo et al., 2013), and an increase of gray matter density is found in PAG, thalamus and cerebellum months Compared to the control, inhibiting PL-PAG pathway relieved the aggravated phases II formalin pain in the CFA group (the middle column). Detailed pain scores were shown in every 5 min (the right column). n = 8 in each group. In the middle column: * * * p < 0.001, hM4Di+aCSF vs. hM4Di+CNO, one-way ANOVA. In the right column: * p < 0.05, * * p < 0.01, * * * p < 0.001, hM4Di+aCSF vs. hM4Di+CNO, ANOVA with repeated measures and Bonferroni post hoc test.
after whiplash injury and after the headache subsided in most patients with posttraumatic headache (Obermann et al., 2009). Chronic back pain also alters the human brain chemistry: reduction of N-acetyl aspartate and glucose is demonstrated in the dorsolateral prefrontal cortex, but not cingulate, sensorimotor and other brain regions (Grachev et al., 2000). In animals, long-term neuropathic pain decreases frontal cortex volumes several months after nerve injury (Seminowicz et al., 2009), and increases basal dendrites of neurons and spine densities in the PL (Metz et al., 2009). These structural and functional changes are accompanied by altered synaptic plasticity and neuronal excitability (Metz et al., 2009;Baliki et al., 2014). Taken together, PL is a brain region particularly vulnerable to chronic pain, regardless of pain recovery or not (Grachev et al., 2000;Apkarian et al., 2004). Some evidences show that pain-induced functional and structural abnormalities in the PFC are at least partially reversible by effective pain treatment (Rodriguez-Raecke et al., 2009;Seminowicz et al., 2011). However, we demonstrated persistent hypersensitivity of PL neurons in rats with chronic pain experience to following pain assaults.
However, we do not consider PL to be the only brain region that contributes to the aggravated nociceptive responses in rats with experience of chronic inflammatory pain. Other areas such as anterior cingulate cortex (ACC) and infralimbic cortex (IL), insular cortex and thalamus are densely connected with PL and undergo plastic changes in chronic pain (Obermann et al., 2009;Ivo et al., 2013;Lin, 2014;Zhuo, 2016;Yue et al., 2017).

Persistent Hyperphosphorylation of p38 Contributes to Aggravated Formalin Pain in Rats With Chronic Pain Experience
p38 is one of the major MAPK members crucial for generating pain hypersensitivity through transcription-dependent and -independent means (Ji and Woolf, 2001;Crown et al., 2006;Wynne, 2006;Toyoda et al., 2007;Ji et al., 2009). In both peripheral and central nervous system, such as spinal cord and ACC, p38 is activated following pain stimulations, and the activation lasts for more than 3 weeks after chronic constriction injury or spinal nerve ligation (Toyoda et al., 2007;Crown et al., 2008;Cao et al., 2014). Inhibiting p38 in spinal cord or ACC alleviates both subacute and chronic pain (Kumar et al., 2003;Crown et al., 2008;Cao et al., 2014). In the present study, p-p38 + neurons are mostly glutamatergic neurons in PL. Under persistent noxious stimulation, hyperphosphorylated p38 can be invoked to activate the downstream signal pathway, including p-CREB and c-Fos, and the activation of nociceptive neurons further sensitizes pain. Inhibiting the phosphorylation of p38 relieves the aggravated formalin pain in rats with chronic pain experience and down-regulates the phosphorylation of its downstream effector CREB and the neuronal activity. The persistent hyperphosphorylation of p38 in PL in rats with chronic pain experience does not directly contribute to pain responses unless a strong stimulus such as formalin injection is performed. These findings indicate that some endogenous inhibitors overlay sensitized p-p38 to keep the balanced ''recovery'' state.
One candidate of such inhibitors would be layer V pyramidal neurons in the IL, which directly innervate inhibitory interneurons in the PL (Saffari et al., 2016). In addition to pain (Baliki et al., 2012;Wang et al., 2015;Wu et al., 2016), PL also regulates other long-term cognitive behaviors including fear conditioning, working memory, drug addiction and chronic stress (Gisquet-Verrier and Delatour, 2006;Lasseter et al., 2010;Negrón-Oyarzo et al., 2014;Fitzgerald et al., 2015;Moench and Wellman, 2015;Seo et al., 2017). PL neuronal activities maintain freezing behaviors in fear conditioning, whereas extinction induces a novel inhibitory learning in the IL, which antagonize PL pyramidal neuronal activities through local interneurons (Ji and Neugebauer, 2012). These facts lead to an intriguing proposal that the recovery from chronic pain represents pain extinction, instead of simple oblivion of the pain memory .

PL-PAG Pathway Mediates Aggravated Formalin Pain in Rats With Chronic Pain Experience
PL has direct projections to several nuclei, including ACC, NAc, amygdala and PAG. Previous studies have revealed roles of PL-NAc and PL-PAG pathways in regulating chronic pain (Kucyi et al., 2013;Yu et al., 2014;Lee et al., 2015). Activation of the PL to NAc core circuit inhibits persistent neuropathic pain and relieves the affective symptoms associated with chronic pain (Lee et al., 2015). Functional and structural connectivity between PAG and mPFC relates to individual differences in attention to pain (Kucyi et al., 2013). PAG is a critical component of the descending pain modulatory system, and usually exerts an inhibitory effect on nociceptive transmission (Umana et al., 2017). Evidence shows that glutamatergic projections from mPFC act to inhibit PAG function (Franklin et al., 2017). Pharmacogenetic inhibition of the projections from PL to PAG therefore disinhibits PAG, and in turn relieves the aggravated formalin-induced pain. The finding that paw lifting behaviors in response to formalin injection, indicating peripheral reflexes, increase in rat with chronic pain experience also supports the involvement of the descending pathway.
Different from the role of PAG in perceptual part of pain, NAc is more involved in pain affection, such as depression-like behaviors in the chronic neuropathic pain state (Goffer et al., 2013;LeBlanc et al., 2015;Navratilova et al., 2015;Kaneko et al., 2017). The present study shows that PL-PAG, but not PL-NAc, pathway regulates the aggravated formalin pain in rats with chronic pain experience, which reflects mostly aggravated pain sensation.

PL as a Potential Therapeutic Target
Patients with chronic pain have less PFC deactivation than controls during cognition or visual attention tasks (Baliki et al., 2008;Seminowicz et al., 2011), in keeping with our results that the PL is more strongly activated during the following noxious assault in rats with chronic pain experience. Previous work has identified the PL as a target for pain treatment. Repetitive transcranial magnetic stimulation, deep brain stimulation, cognitive-behavioral therapy or other therapies can lead abnormal function of prefrontal cortex to stable or normal state (Fierro et al., 2010;Seminowicz et al., 2013;Boccard et al., 2015), thus relieves the aggravated nociceptive responses in patients with chronic pain experience.
In conclusion, our present study demonstrates that persistent hyperphosphorylation of p38 in the PL underlies aggravated nociceptive responses in rats with chronic inflammatory pain experience, and indicates inhibiting the activity of PL as a promising option for treating hyperalgesia in patients with chronic pain experience.

AUTHOR CONTRIBUTIONS
X-CF, MY and YW designed experiments and wrote the manuscript; X-CF, SF, F-YL and SC performed the experiments and analyzed the data; YW and MY supervised the experiments.
FIGURE S5 | Chronic pain experience and formalin injection do not affect the phosphorylation of ERK1/2 and JNK in PL. (A) Similar levels of p-ERK1/2 in PL in rats with chronic pain experience. Formalin injection did not affect the phosphorylation of ERK1/2 in PL in rats with or without chronic pain experience. Representative Western blots of p-ERK1/2, ERK1/2 and β-tubulin were shown above the corresponding histogram. n = 8 in each group, one-way ANOVA.
(B) Lack of significant changes of p-JNK in PL in rats with chronic pain experience. Formalin injection did not affect the phosphorylation of JNK in PL in rats with or without chronic pain experience. Representative Western blots of p-JNK, JNK and GAPDH were shown above the corresponding histogram. n = 8 in each group, one-way ANOVA.