Psychopharmacology (2014) 231:2171–2187 DOI 10.1007/s00213-013-3368-2

ORIGINAL INVESTIGATION

Chronic curcumin treatment normalizes depression-like behaviors in mice with mononeuropathy: involvement of supraspinal system and GABAA

Xin Zhao & Chuang Wang & Jun-Fang Zhang & Li Liu & Ai-Ming Liu & Qing Ma & Wen-Hua Zhou & Ying Xu

Received: 31 July 2013 /Accepted: 18 November 2013 /Published online: 3 December 2013 # Springer-Verlag Berlin Heidelberg 2013

Abstract behaviors, which was abolished by chemical depletion of Rationale Comorbid depression is commonly observed in brain rather than noradrenaline. The paralleled individuals who suffer from neuropathic pain, which necessi- antinociceptive and -like actions of curcumin tates improved treatment. Curcumin, a phenolic compound seem to be pharmacologically segregated, since intrathecal derived from Curcuma longa, possesses both antinociceptive and intracerebroventricular injection of , a non- and antidepressant-like activities in animal studies, suggesting selective 5-HT receptor antagonist, separately counteracted its possible usefulness in treating this comorbidity. the two actions of curcumin. Further, this antidepression was

Objective We investigated the effect of curcumin on abrogated by repeated co-treatment with 5-HT1A receptor depressive-like behaviors in mice with mononeuropathy, and antagonist WAY-100635 and greatly attenuated by acute co- explored the mechanism(s). treatment with GABAA receptor antagonist bicuculline. Methods Chronic constriction injury (CCI) was produced by Conclusion Curcumin can normalize the depressive-like be- loosely ligating the sciatic nerves in mice. The nociceptive haviors of neuropathic mice, which may be independent of the behaviors were examined using Hargreaves test, and the concurrent analgesic action and possibly mediated via the depressive-like behaviors were determined by forced swim supraspinal serotonergic system and downstream GABAA test (FST) and tail suspension test (TST). receptor. Results After CCI injury, the neuropathic mice developed nociceptive and depressive-like behaviors, as shown by ther- Keywords Curcumin . Depression . Neuropathic pain . mal hyperalgesia in Hargreaves test and protracted immobility Serotonin . GABAA time in FST and TST. Chronic treatment of neuropathic mice with curcumin (45 mg/kg, p.o., twice per day for 3 weeks) corrected their exacerbated nociceptive and depressive-like Introduction

Neuropathic pain is commonly a chronic and debilitating * : : < : < : : X. Zhao ( ) C. Wang J. F. Zhang A. M. Liu Q. Ma disease that is significantly associated with mood disorders W.

C21H20O6, the chemical structure shown in Fig. 1a) has been and Xie (1988), with minor modifications (Zhao et al. 2012). shown to possess a variety of pharmacological activities, Briefly, the mice were anesthetized by intraperitoneal (i.p.) including antioxidant, anti-inflammatory, cancer chemopre- injection of pentobarbital sodium (65 mg/kg). The right com- ventive, and neuroprotective actions (Radhakrishna Pillai mon sciatic nerve was exposed at the nerve of the midthigh et al. 2004; Yadav et al. 2005; Reeta et al. 2010). Our previous and, proximal to the sciatic nerve trifurcation, three ligations studies delved into the CNS potentials of curcumin, with focus (4/0 silk threads, with 1 mm spacing) were loosely tied around on its antidepressant-like effect in some animal models (Xu it until a brief twitch was seen in the respective hind limb. The et al. 2005, 2007). Recently, we revealed that curcumin en- surgical area was dusted with streptomycin and the incision genders analgesic activities against neuropathic pain (Zhao was sutured. In sham-operated animals, the sciatic nerve was et al. 2012). The concurrent antinociceptive and exposed but not ligated. antidepressant-like activities exerted by curcumin prompted us to investigate, in the present study, whether it possesses Drugs and pharmacological treatments therapeutic potentials in treating the comorbid depression induced by chronic neuropathic pain. The following drugs were used: β-funaltrexamine (irrevers-

We addressed this subject by examining the profiles indic- ible μ-opioid receptor antagonist), bicuculline (GABAA re- ative of depressive-like behaviors in mice, i.e., immobility ceptor antagonist), curcumin, (injection, 10 mg/ time in forced swim test (FST) and tail suspension test 2 ml), isamoltane (5-HT1B receptor antagonist), methysergide (TST). Moreover, we investigated the mechanisms underlying (non-selective 5-HT receptor antagonist), (5-HT3 the antidepressant-like action by curcumin in the context of receptor antagonist), (5-HT2A/ receptor antago- neuropathic pain. Our results showed that chronic, but not nist), and WAY-100635 (5-HT1A receptor antagonist). acute, curcumin treatment can normalize the pain-related Bicuculline, isamoltane, and methysergide were purchased (hyperalgesia to heat stimuli) and depressive-like behaviors from Tocris Bioscience. Diazepam was purchased from Tian- in mice with mononeuropathy. The pronounced jin Pharmaceutical Co., Ltd. (Tianjin, China). All the Psychopharmacology (2014) 231:2171–2187 2173

Fig. 1 Chemical structure of A curcumin (Cur) and schematic of OCH3 chronic curcumin treatment. a Chemical structure of curcumin. HO b Fifteen days after chronic OH construction injury (CCI) surgery, we started the chronic treatment with either vehicle solution OCH3 (peanut oil) or curcumin (5, 15, O O and 45 mg/kg, p.o., twice a day). The first administration time Curcumin (Cur, molecular weight: 368.38 ) was at 1000 hours and the second administration time was at 1800 B hours. After 3 weeks of curcumin treatment (when the hyperalgesia Surgery (CCI) Treatment (Cur/vehicle) and depressive-like behaviors of neuropathic mice were corrected by curcumin), PCPA, DSP-4, or /GABAergic/ Days opioidergic antagonists were 0 15 20 25 30 35 40 co-administered with vehicle or curcumin Drug intervention PCPC 1st administration 2nd administration DSP-4

Antagonists 7:00 13:00 19:00 1:00 7:00 test remaining drugs were purchased from Sigma-Aldrich. All the curcumin, these antagonists were i.p. injected 30 min before drugs were dissolved or diluted in sterile saline with the behavioral tests (Zhao et al. 2012) on day 36 (i.e., 36 days exception of curcumin and bicuculline, whose vehicle were after CCI surgery). peanut oil and saline containing 1 % DMSO, respectively. The treatment with curcumin (p.o., via gavage, with a Chemical depletion of brain noradrenaline and serotonin volume of 10 ml/kg) began 15 days after the surgical proce- dure (CCI or sham operation), when the neuropathic mice After treatment with curcumin or vehicle for 3 weeks, phar- exhibited marked thermal (heat) hyperalgesia and maceutical manipulations were performed to deplete nor- depressive-like behaviors. For acute treatment, curcumin (NA) or serotonin. For chemical depletion of brain was administered to the sham and CCI-injured mice on day NA, the mice were injected i.p. with 40 mg/kg selective NA 15 (i.e., 15 days after CCI injury), which was followed by neurotoxin DSP-4 as reported previously (Boyce-Rustay et al. behavioral tests. For chronic treatment, the sham and CCI 2008). DSP-4 was dissolved in saline. To ablate brain seroto- mice received two oral administrations (morning and evening) nin, PCPA (an inhibitor of serotonin synthesis), suspended in of curcumin per day for 21 consecutive days (3 weeks, i.e., 0.5 % gum acacia/physiological saline, was administered i.p. from day 15 to day 35 after CCI injury, Fig. 1b). After 3 weeks for five consecutive days at a dose of 300 mg/kg/day (Tanabe of curcumin treatment, the mice were co-administered with et al. 2007;Zhaoetal.2012). one of the 5-HT receptor, GABAergic receptor, or opioid receptor antagonists (Fig. 1b). These antagonists were meth- Intrathecal and intracerebroventricular injection ysergide, WAY-100635, isamoltane, ritanserin, ondansetron, bicuculline, and β-funaltrexamine. The doses of these antag- To localize the monoamine receptors possibly involved, intra- onists were selected on the basis of our previous studies (for thecal (i.t.) or intra-cerebroventricular (i.c.v.) injection of WAY-100635, isamoltane, ritanserin, ondansetron, and β- methysergide or WAY 100635 was undertaken after 3 weeks funaltrexamine, Zhao et al. 2012) and other literature (for of curcumin or vehicle treatment. We did two i.t. or i.c.v. methysergide, Duarte et al. 2008; for bicuculline, De Souza injections on the same day, 30 min before the morning and et al. 2009). For i.p. injection, these drugs were administered the evening treatments with curcumin, respectively. The mice in a volume of 10 ml/kg. For repeated co-administered with were then tested for pain-related and depressive-like behaviors curcumin, these antagonists were injected twice per day, on the following morning, i.e., 39 h following the last admin- 30 min before curcumin administration (Yalcin et al. 2009; istration of curcumin alone. These procedures were based Zhao et al. 2012). For a single co-administration with upon our previously established procedures (Zhao et al. 2012). 2174 Psychopharmacology (2014) 231:2171–2187

Intrathecal injections were performed as described by the average of three separate trials, taken at 5-min intervals to Hylden and Wilcox (1980). Briefly, a 27-gauge needle con- prevent thermal sensitization and behavioral disturbances. A nected to a 50-μl Hamilton syringe was inserted into the sub- cutoff time was set at 22 s to prevent tissue damage. arachnoidal space, between the L5 and L6 vertebrae. The placement of the needle was verified via the induction of a tail FST This procedure was performed according to the method flick movement. A volume of 10 μl was used for i.t. injection. established by Porsolt et al. (1977). Briefly, mice were indi- Under ether anesthesia, intracerebroventricular injection (a vidually placed in clear Plexiglas cylinders (height, 25 cm and volume of 2.5 μl) was undertaken according to the method diameter, 10 cm), which were filled with water (depth, 10 cm described by Haley and McCormick (1957). In short, during and temperature, 24±1 °C), for 6 min. A mouse was judged to anesthesia, mice were grasped firmly by the loose skin behind be immobile when it floated motionless, except for making the head. A 27-gauge needle connected to a 50-μlHamilton only necessary movements to keep its head above the water. syringe was inserted perpendicularly through the skull and no The immobility time was recorded during the last 4 min of the more than 2 mm into the brain of the mouse. The injection site 6-min trial. was 1 mm right to the midpoint on a line drawn through to the anterior base of the ears. Injections were performed into the TST This procedure was based on the method described pre- right ventricle. After behavioral tests, the mice were decapi- viously (Steru et al. 1985). In short, mice were suspended, tated rapidly, and the brains were removed and frozen in 50 cm above the floor, for a 6-min test session using an −70 °C. The brains were sectioned (40 μm thickness) in the adhesive tape affixed 1 cm from the tip of the tail. Immobility coronal plane and the sections were stained by cresyl violet to was defined as the complete cessation of movement while verify the right injection site. The injection success rate in our suspended. The immobility time was recorded with a stop- laboratory was normally more than 95 %. watch during the 6-min test session.

Behavioral tests Rotarod test The apparatus (Model DXP-2, Chinese Acade- my of Medical Science, China) consists of a base platform and Hargreaves test and two behavior despair tests (FST, TST) a rotating rod with a diameter of 3 cm, subdivided into four were employed to examine in mice the pain-related and equal sections. Each mouse was trained to run the rotarod at a depressive-like behaviors, respectively. To rule out the possi- constant speed (30 rpm) for 40 s without falling. The time bility of sedation and motor impairment produced by mice were able to maintain their balance on the rod was curcumin, the locomotor activity test and rotarod test were measured. The cutoff time was 140 s. carried out, respectively. In FST and TST, the animals were tested only once, i.e., the different time points of tests were Locomotor activity test The assessment of locomotor activity done on independent sets of animals and each animal was was performed in mice essentially according to our previous tested only once. In the other behavioral tests (i.e., Hargreaves studies (Zhen et al. 2012). Locomotor activity was mea- test, locomotor activity test, and rotarod test), the animals were sured with an ambulometer with five activity chambers tested consecutively in a time-course fashion. To investigate (JZZ98, Institute of Materia Medica, Chinese Academy of the acute actions of curcumin, the behavior tests were under- Medical Sciences, China). Mice were placed in the center of taken at different time points (e.g., 0.5, 1, 2, 3, and 4 h) after a the chambers, with their paws connected or disconnected single curcumin administration on day 15. In the experimental with the active bars producing random configurations that settings for chronic curcumin treatment, the behavior tests were converted into pulses. The pulses were automatically were done 1 h before the morning curcumin treatment (first recorded for the cumulative total counts of motor activity. curcumin treatment on the day, as shown in Fig. 1b)as Mice were placed in the test chamber 10 min prior to the reported by our previous study (Zhao et al. 2012). recording session and then locomotion counts were recorded for 5 min. Hargreaves test Thermal hyperalgesia was tested according to the Hargreaves procedure (Hargreaves et al. 1988), slightly Measurement of monoamines (metabolites) and monoamine modified by us for mouse, using a Plantar test apparatus (Ugo oxidase activity Basile, Italy). Mice were placed in smaller clear Plexiglas cubicles and allowed to acclimatize for at least 20 min before After 3 weeks curcumin (or vehicle) treatment, the mice testing. A constant intensity (approximately 10 s on average (sham-operated or CCI) were decapitated and the brains for naïve mice) radiant heat source was focused on the right were rapidly removed and dissected on an ice-chilled glass hind paw of mice and the thermal latency was defined as the plate, with the prefrontal cortex and hippocampus obtained time, in seconds, from initial heat source activation to hind according to our previous protocol (Zhen et al. 2012). The paw withdrawal. The thermal latency was determined from tissue samples were weight, and then stored at −80 °C. The Psychopharmacology (2014) 231:2171–2187 2175 contents of monoamines (NA, serotonin, and ) and A Hargreaves Test their metabolites (5-HIAA and DOPAC) were measured as 12 described previously by high-performance liquid chroma- tography with electrochemical detection (Zhao et al. 2012). 10

The monoamine oxidase (MAO) activity was assessed 8 Sham ** according to our previously established protocol (Zhen CCI et al. 2012). 6

Thermal lantency (s) 4 Statistical analysis 0 5 10 20 30 40 Time (days after CCI) All values are presented as the mean±S.E.M. Data were ana- lyzed by multifactor analysis of variance (ANOVA) or one-way B ANOVA. For multifactor ANOVA, for example, the surgery Forced swim test (sham or CCI) and the treatment (vehicle or drug administration) 200 ** ** ** were done as between-group factors. When needed, the time of measurement (time-course data) was done as within-subject factor. The Duncan test was used for post hoc comparisons. 100 For one-way ANOVA, Student–Newman–Keuls test was used for multiple comparisons, followed by Student’stesttoevaluate

Immobility time (s) 0 the difference between two groups at the same time. Differences 1W 2W 4W 6W BL with p <0.05 were considered statistically significant. Time (weeks after CCI) Naïve Sham Results CCI C Tail suspension test ** Effect of peripheral nerve injury (CCI) on pain-related 160 * ** and depressive-like behaviors in mice

The neuropathy induced by CCI produced in mice persistent 80 ipsolateral (but not contralateral paw, data not shown) thermal

hyperalgesia, which lasted for at least 40 days, as compared Immobility time (s) 0 with sham-operated mice (F10,196=5.1,p <0.01;Fig.2a). This 1W 2W 4W 6W BL condition of enduing hyperalgesia caused depressive-like be- Time (weeks after CCI) havior in these CCI mice, as evidenced by prolonged immo- Fig. 2 Exacerbated pain-related and depressive-like behaviors after pe- bility time in FST (F 3,72=7.7, p <0.01; Fig. 2b) and TST ripheral nerve injury (CCI) in mice. a Pain-related behavior was tested by (F 3,68=6.9, p <0.01; Fig. 2c), from 2 to 6 weeks after CCI Hargreaves test. After surgery, the neuropathic (CCI) mice developed injury, but not at earlier time point (Fig. 2b, c). enduring thermal hyperalgesia, which lasted for at least 40 days. b, c Depressive-like behaviors were tested by FST (b) and TST (c). Two weeks after surgery (2W), the CCI mice exhibited more pronounced Effect of chronic curcumin treatment on pain-related depressive-like behaviors (expressed as increased immobility time) and and depressive-like behaviors in neuropathic mice the depressive-like behaviors lasted for at least 4 weeks (from 2W to 6W). Data are expressed as mean±SEM (n =8–12 per group), assessed by multifactor ANOVA followed by Duncan test. BL,baseline;FST, forced After 3 weeks treatment, curcumin markedly attenuated in swim test; TST, tail suspension test CCI mice the pronounced thermal hyperalgesia (F 7,282= 10.6, p <0.01; Fig. 3a) and depressive-like behavior (FST:

F 3,34=22.3, p <0.01, Fig. 3b; TST: F 3,36=25.6, p <0.01, time in FST (Fig. 3b)andTST(Fig.3c), without influencing Fig. 3c) in a dose-dependent manner, with significant differ- the measures in sham-operated mice. ence at doses of 15 and 45 mg/kg (increase in thermal latency by 36.8 and 93.5 %, respectively, in Hargreaves test, Fig. 3a; Effect of acute curcumin treatment on pain-related decrease in immobility time by 23.6 and 37.3 %, respectively, and depressive-like behaviors in neuropathic mice in FST, Fig. 3b; 21.0 and 34.7 %, respectively, in TST, Fig. 3c). High dose curcumin (45 mg/kg), after 3 weeks treatment, As shown in Fig. 4a, acute curcumin treatment, at high exerted recuperating effects on the thermal hyperalgesia dose of 45 mg/kg, did not influence the thermal sen- (Fig. 3a) in CCI mice and reversed their protracted immobility sitivity in Hargreaves test regardless of sham-operated 2176 Psychopharmacology (2014) 231:2171–2187

Fig. 3 Effect of chronic A curcumin (Cur) treatment on Hargreaves Test pain-related and depressive-like 12 behaviors in mice. Chronic Vehicle or Cur (45 mg/kg) CCI Mice 12 curcumin treatment started 10 ** 15 days following CCI procedure Sham + vehicle * 8 when the comorbidity of 8 Sham + Cur neuropathic pain and depression ** CCI + vehicle was developed. a Effect of 6 4 chronic curcumin treatment on CCI + Cur the thermal sensitivity Thermal lantency (s) 0

4 Thermal latency (s) (Hargreaves test) in sham and 0 15 20 25 30 35 40 0 5 15 45 CCI mice. Chronic curcumin Time (days after CCI) Cur (mg/kg) treatment (45 mg/kg, p.o., twice per day) corrected neuropathy- induced thermal hyperalgesia B (left panel for time-course) and Forced swim test the antinociceptive effect of 0 W 1 W curcumin was dose-dependent 240 0 240 (5, 15, and 45 mg/kg, right ** Cur 5 panel). b, c Effect of chronic 120 (mg/kg) 120 curcumin treatment on the 15 measures of depressive-like 45 0 0 behaviors evaluated by FST (b) Sham CCI Sham CCI and TST (c) in sham and CCI mice. Chronic treatment of CCI 2 W 3 W mice with curcumin (5, 15, and 240 240 ** 45 mg/kg, p.o., twice per day) ** * dose- and time-dependently 120 120 decreased the immobility time in Immobility time (s) FST (b) and TST (c). Data are 0 0 expressed as mean±SEM Sham CCI Sham CCI (n =8–12 per group), assessed by multifactor ANOVA followed by Duncan test, and one-way C Tail suspension test ANOVA followed by Student–Newman–Keuls test. 0 W 1 W FST, forced swim test; TST,tail 200 0 200 ** Cur suspension test 5 (mg/kg) 100 15 100 45 0 0 Sham CCI Sham CCI

2 W 3 W 200 200 ** * *

Immobility time (s) 100 100

0 0 Sham CCI Sham CCI or CCI-injured mice. In contrast, acute curcumin treat- Effect of depleting brain 5-HT or NA on the antinociceptive ment (1.5, 5, and 15 mg/kg, p.o.) engendered rapid and antidepressant-like effects of curcumin onset antidepressant-like effects in a dose-dependent manner, as evidenced by a decrease of immobility Our previous studies indicate that monoaminergic system, time in FST (F 3,72=6.1, p <0.01, Fig. 4b)andTST especially 5-HT and NA, is critically responsible for the (F 3,76=8.4, p <0.01, Fig. 4c). This antidepressant-like antinociceptive and antidepressant-like effects of curcumin effect by curcumin seems to be transient and nonspe- (Xu et al. 2005, 2007;Zhaoetal.2012). We therefore ex- cific, since it provided only fleeting amelioration of plored the potential monoamine mechanisms underlying depressive-like behaviors (lasted for no more than 4 h), and curcumin actions in the context of mononeuropathy. We in- this action was expressed in both sham-operated and CCI- vestigated whether the antinociceptive and antidepressant-like injured mice. actions of curcumin would be influenced by chemical Psychopharmacology (2014) 231:2171–2187 2177

Fig. 4 Effect of acute curcumin A (Cur) treatment on pain-related Hargreaves Test and depressive-like behaviors in neuropathic mice. Acute Vehicle or Cur (45 mg/kg) curcumin treatment started 12 15 days following CCI when the comorbidity of neuropathic 10 pain and depression was Sham + vehicle developed. a Effect of acute Sham + Cur curcumin treatment on the 8 ** thermal sensitivity (Hargreaves CCI + vehicle test) in mice. Following baseline 6 CCI + Cur (0 h) assay of thermal sensitivity, mice (sham or CCI) were Thermal lantency (s) 4 administered with curcumin 0 1 2 3 4 (45 mg/kg, p.o.) and then tested Time (h) for thermal sensitivity at 0.5, 1, 1.5,2,3,and4hfollowing curcumin administration. There B are no alterations in thermal Forced swim test sensitivity following single treatment of curcumin. b, c Effect 0 h 1 h 240 240 ** of acute curcumin treatment on ** 0 the measures of depressive-like Cur ** ** 1.5 ** behaviors evaluated by FST (b) 120 (mg/kg) 120 and TST (c) in mice. Mice (sham 5 or CCI) were given vehicle or Cur 15 0 0 (1.5, 5, and 15 mg/kg) and Sham CCI Sham CCI challenged for FST or TST at different time points (0, 1, 2, 4, 2 h 4 h 6 h and 6 h) after Cur administration. 240 240 240

Although acute curcumin Immobility time (s) ** treatment (1.5, 5, and 15 mg/kg, ** 120 120 p.o.) engendered rapid-onset 120 antidepressant-like effect, shown by a decrease of immobility time 0 0 0 in FST (b) and TST (c), this Sham CCI Sham CCI Sham CCI action seems to be transient (no more than 4 h). Data are expressed as mean±SEM (n =8– C 12 per group), assessed by Tail suspension test multifactor ANOVA followed by 0 h 1 h Duncan test or one-way ANOVA 200 – – 200 0 ** followed by Student Newman ** Cur * Keuls test. FST, forced swim test; 1.5 ** (mg/kg) * TST, tail suspension test 100 5 100 15 0 0 Sham CCI Sham CCI

200 2 h 200 4 h 200 6 h * Immobility time (s) 100 ** 100 100

0 0 0 Sham CCI Sham CCI Sham CCI depletion of brain NA or 5-HT. It has been reported that a consecutive PCPA (300 mg/kg, i.p., for five consecutive days) single i.p. injection of DSP-4 or consecutive i.p. injections of administrations, respectively. PCPA reduced the central NA or 5-HT content, respectively As shown in Fig. 5, 6 days after a single administration, (Tanabe et al. 2007; Boyce-Rustay et al. 2008). Here, we also DSP-4 (40 mg/kg, i.p.) did not affect the antihyperalgesic and observed a significant reduction of brain NA or 5-HT content antidepressant-like effects of curcumin. Nevertheless, PCPA (Table 1), after a singular DSP-4 (40 mg/kg, i.p.) or (300 mg/kg, i.p.), after 4–5 days consecutive injections, 2178 Psychopharmacology (2014) 231:2171–2187

Table 1 Effects of DSP-4 (40 mg/kg, a single i.p. injection) Group NA (ng/g) 5-HT (ng/g) and PCPA (i.p., 300 mg/kg per day, for five consecutive days) Hippocampus Frontal cortex Hippocampus Frontal cortex on noradrenaline (NA) and 5-HT levels in the frontal cortex and Sham mice hippocampus of Sham and Vehicle+vehicle 427.8±27.2 504.7±36.5 538.5±22.4 397.8±28.2 CCI mice ** ** Vehicle+DSP-4 143.7±15.4 164.1±24.2 524.3 ±30.6 403.2±31.6 Vehicle+PCPA 405.0±12.9 478.9±18.0 192.9±17.7** 137.5±15.1** Cur45+vehicle 440.3±22.1 526.8±31.3 572.0±31.5 417.7±22.5 Cur45+DSP-4 128.7±26.4** 181.8±36.7** 549.3±27.4 411.4±25.8 Cur45+PCPA 418.6±17.8 493.5±22.3 208.1±23.8** 153.8±18.6** CCI mice Vehicle+vehicle 378.6±24.2 472.5±28.6 389.7±20.2 264.0±24.3 Values were expressed as mean± ## ## SEM of 6–10 mice Vehicle+DSP-4 118.1±32.9 155.1±23.3 395.6±28.8 251.6±22.9 ## ## Cur45 curcumin 45 mg/kg Vehicle+PCPA 371.4±21.0 463.0±26.5 164.4±18.5 122.4±18.2 ** p <0.01 (compared with vehi- Cur45+vehicle 414.9±37.3 511.9±34.4 527.3±34.0 376.1±33.5 cle + vehicle treated sham mice); Cur45+DSP-4 134.0±18.9## 172.4±16.7## 516.9±21.4 365.3±27.0 ## p <0.01 (compared with vehi- Cur45+PCPA 406.7±28.5 491.6±21.6 172.7±16.3## 140.2±21.8## cle + vehicle treated CCI mice) completely abolished both the antinociceptive effect (F 6,224= depressive-like behaviors in sham-operated mice or CCI mice 8.7, p <0.01; Fig. 5a, left panel) and the antidepressant-like treated with vehicle (Fig. 5a, b), which is congruent with other effect (F3,74=7.9, p <0.01; Fig. 5b, right panel) by curcumin, reports (Suzuki et al. 2005; Tanabe et al. 2007) and indicates indicating an essential role for serotonergic system in the that the depletion of 5-HT and NA per se, by PCPA (300 mg/kg, analgesic and antidepressant-like actions of curcumin. Of note, i.p.) and DSP-4 (40 mg/kg, i.p.), respectively, may not influence chemical depletion of brain 5-HT and NA by PCPA and DSP-4, the physiological and pathological profiles of pain-related and respectively, did not change the measures of pain-related and depressive-like behaviors in mice.

Fig. 5 Depletion of 5-HT but not A NA concurrently abolished the Hargreaves Test analgesic and anti-depressant like effects of curcumin (Cur). To Vehicle or Cur (45 mg/kg) Vehicle or Cur (45 mg/kg) deplete NA or 5-HT, DSP-4 12 + PCPA 12 DSP-4 (40 mg/kg, i.p.) or PCPA 10 (300 mg/kg, i.p.) was injected Sham + vehicle 10 once (on day 36) or for five 8 Sham + Cur 8 ** consecutive days (from day 36 CCI + vehicle ** to day 40), respectively. a 6 CCI + Cur 6 Depletion of 5-HT (left panel)but 4 4 Thermal lantency (s) not NA (right panel) abolished 35 37 39 41 Thermal lantency (s) 36 38 40 42 the analgesic action of curcumin Time (days after CCI) on thermal hyperalgesia. b Effect Time (days after CCI) of depletion of 5-HT or NA on the immobility time measured in FST. Depletion of 5-HT by PCPA B Forced swim test abrogated the antidepressant-like effect of curcumin. Data are Sham CCI expressed as mean±SEM 240 240 ** * (n =8–12 per group), assessed by multifactor ANOVA followed by 120 120 Duncan test. FST forced swim 0 test; TST tail suspension test 0 5 Cur 0 Vehicle PCPA (mg/kg) Vehicle PCPA 15 240 45 240 ** ** * *

120 120 Immobility time (s) Immobility time (s)

0 0 Vehicle DSP-4 Vehicle DSP-4 Psychopharmacology (2014) 231:2171–2187 2179

Effect of chronic curcumin treatment on the levels of brain nonselective 5-HT receptor antagonist) to mice chronically monoamines and MAO activity in sham and neuropathic mice treated with curcumin, and evaluated its effect, via three different routes of drug delivery (i.p., i.t., and i.c.v.), on the As shown in Tables 2 and 3, chronic curcumin treatment (5, 15, antidepressant-like and analgesic actions of curcumin. As and 45 mg/kg, p.o., twice per day) dose-dependently increased, shown in Fig. 6, repeated i.p. co-administration of Met in CCI-injured mice, the contents of serotonin in the hippocam- (3 mg/kg) completely abolished both antihyperalgesic and pus (by 5.4 %, 18.7, and 32.7, respectively, Table 2)andfrontal antidepressant-like effects by curcumin (Hargreaves test, cortex (by 3.2, 12.8, and 33.2 %, respectively, Table 3), and F 1,42=31.1, p <0.01, Fig. 6a;FST:F 1,40=54.7, p <0.01, decreased the ratio of 5-HIAA/5-HT (by 6.9, 22.7, and 36.4 %, Fig. 6b), which is consistent with the reversing effect done respectively, in the hippocampus, Table 2; by 3.4, 11.5, and by PCPA. Interestingly, repeated i.t. co-administration of Met 31.0 %, respectively, in the frontal cortex, Table 3). There was a (5 μg/10 μl) specifically blocked the antinociceptive effect of relatively minor increase, after chronic curcumin treatment, in curcumin (F 1,41=34.6, p <0.01, Fig. 6a), without influencing the levels of NA in the hippocampus and frontal cortex of CCI its paralleled antidepressant-like effect. Conversely, repeated mice, with significant difference at the highest dose of 45 mg/ i.c.v. co-administration of Met (10 μg/2.5 μl) only abolished kg (in the hippocampus, Table 2). the antidepressant-like effects of curcumin (F1,38=61.3, p < We also evaluated the effect of chronic curcumin 0.01, Fig. 6b), without affecting its antihyperalgesic effects. treatment on the type A and type B MAO activities in These results suggest the two actions of curcumin should be mouse brain. As shown in Table 4, MAO-A but not regionally discrete. MAO-B activity increased markedly when the mice were exposed to CCI injury, which is congruent with Supraspinal 5-HT1A receptors are essential a recent report by Villarinho et al. (2013). Chronic for the antidepressant-like effect of curcumin in mice treatment of curcumin (5, 15, and 45 mg/kg) dose- with mononeuropathy dependently inhibited the escalated brain MAO-A activity in CCI mice, with a reversing effect at the dose of 45 mg/kg. To identify which subtype of 5-HT receptors mediates the The same curcumin regimen did not affect brain MAO-B antidepressant-like effect of curcumin in neuropathic mice, activity in both sham-operated and CCI-injured mice. we evaluated the effects of several selective 5-HT receptor antagonists on the antidepressant-like effect of curcumin. Effect of i.t. and i.c.v co-administration of methysergide These antagonists were WAY-100635 (1 mg/kg, i.p.), (Met), a nonselective 5-HT receptor antagonist, isamoltane (2.5 mg/kg, i.p.), ritanserin (4 mg/kg, i.p.), and on the antinociceptive and antidepressant-like effects ondansetron (0.5 mg/kg, i.p.). In CCI-injured mice, the of curcumin antidepressant-like effects of curcumin was completely abolished by repeated co-administration of the selective 5-

To determine the regionally discrete serotonergic sites respon- HT1A antagonist WAY-100635 (F 4,185=13.8, p <0.01, sible for the two actions of curcumin, we administrated Met (a Fig. 7a), even though the first acute co-administration exerting

Table 2 Effect of curcumin on monoamine and metabolite levels in the hippocampus of sham and CCI mice

Group Concentration (ng/g)

5-HT 5-HIAA 5-HIAA/5-HT NA Dopamine DOPAC

Sham mice Vehicle 546.9±30.1 153.2±13.9 0.28±0.05 415.4±23.7 105.2±9.7 31.6±5.0 Cur (15 mg/kg) 557.0±21.7 147.9±11.7 0.27±0.04 429.3±18.1 112.0±12.5 29.8±4.7 Cur (45 mg/kg) 572.4±27.5 138.6±15.3 0.24±0.05 436.5±19.8 102.4±10.9 34.1±4.5 CCI mice Vehicle 402.8±22.3## 177.4 ±13.6 0.44±0.04## 381.9±18.4 93.6±11.3 32.2±5.5 Cur (5 mg/kg) 424.5±32.0 173.6±12.2 0.41±0.06 388.7±24.2 98.5±7.8 34.5±4.7 Cur (15 mg/kg) 477.2±19.7** 162.4±14.5 0.34±0.04* 403.6±16.1 91.9±10.0 33.5±5.2 Cur (45 mg/kg) 534.7±25.2** 151.7±9.6* 0.28±0.04** 424.0±17.9* 106.5±12.8 28.6±4.3

Values were expressed as mean values±SEM of six to eight mice # p <0.05 (compared with vehicle treated Sham mice); ## p <0.01, (compared with vehicle treated sham mice); * p <0.05 (compared with vehicle treated CCI mice); ** p <0.01 (compared with vehicle treated CCI mice) 2180 Psychopharmacology (2014) 231:2171–2187

Table 3 Effect of curcumin on monoamine and metabolite levels in the frontal cortex of Sham and CCI mice

Group Concentration (ng/g)

5-HT 5-HIAA 5-HIAA/5-HT NA Dopamine DOPAC

Sham mice Vehicle 382.4±16.4 209.7±15.9 0.55±0.05 527.6±12.4 123.9±8.4 45.2±4.8 Cur (15 mg/kg) 397.6±18.6 201.4±14.0 0.51±0.05 512.3±24.5 125.6±9.1 50.8±5.4 Cur (45 mg/kg) 401.2±11.5 187.5±13.4 0.47±0.04 542.1±15.9 119.7±6.8 42.6±4.5 CCI mice Vehicle 269.5±14.7## 235.8 ±15.7 0.87±0.05## 491.4±24.6 112.8±7.8 48.0±5.6 Cur (5 mg/kg) 278.0±13.9 238.3±9.8 0.84±0.04 509.8±17.2 105.7±8.9 51.4±5.3 Cur (15 mg/kg) 303.8±17.4* 232.7±16.6 0.77±0.06 486.7±9.1 127.3±12.2 43.5±4.2 Cur (45 mg/kg) 358.9±14.6** 217.1±13.4 0.60±0.05** 523.8±15.3 119.5±10.3 45.8±3.8

Values were expressed as mean values±SEM of six to eight mice ## p <0.01 (compared with vehicle treated sham mice); * p <0.05 (compared with vehicle treated CCI mice); ** p <0.01 (compared with vehicle treated CCI mice)

no blocking effect (Fig. 7c). After 4 days of co-treatment, withdrawal (Fig. 7d), implying that this 5-HT1A mechanism WAY-100635 totally abolished the dose-dependent recuperat- may serve as a primary or upstream target. ing effects of curcumin on exacerbated depressive-like behav- To define the regionally discrete roles of the 5- ior (protracted immobility time in FST) in CCI mice treated HT1A receptors in the antidepressant-like effect of with curcumin (F 3,76=20.6, p <0.01; Fig. 7b), without curcumin, we injected WAY-100635 i.t. or i.c.v. in influencing the immobility time (FST) in sham mice or CCI mice chronically treated with curcumin. As shown in mice treated with vehicle. By contrast, the other antagonists Fig. 7f, i.c.v. injection of WAY-100635 (5 μgin for 5-HT receptor subtypes (i.e., isamoltane, ritanserin, and 2.5 μl) abolished the therapeutic effect of curcumin ondansetron) did not modify the antidepressant-like effect of on exacerbated depressive-like behavior (F 1,72=18.7, curcumin (Fig. 7e), supporting the exclusively essential role p <0.01) in CCI mice, without influencing the mea- for 5-HT1A receptor in the antidepressant-like effect of sures of depressive-like behaviors in sham-operated curcumin. Notably, we observed that the modality and time mice. Whereas, i.t. injection of WAY-100635 (1.5 μg course of curcumin–WAY 100635 co-administration (Fig. 7a) in 10 μl) did not affect the immobility time (FST) in were somewhat analogous to that of curcumin treatment sham or CCI mice.

Table 4 Effect of curcumin on type A and type B monoamine oxidase activities in sham and CCI mice

Group Monoamine oxidase-A activity (nmol/30 min/mg protein) Monoamine oxidase-B activity (nmol/30 min/mg protein)

Hippocampus Frontal cortex Hippocampus Frontal cortex

Sham mice Vehicle 110.3±6.5 97.4±5.7 167.8±8.2 143.1±7.3 Cur (15 mg/kg) 113.7±5.6 95.2±6.3 156.4±7.4 147.7±4.3 Cur (45 mg/kg) 105.2±7.5 92.9±6.7 162.7±9.4 141.5±7.2 CCI mice Vehicle 153.9±8.4## 128.0±7.3## 175.3±8.6 150.4±7.6 Cur (5 mg/kg) 146.4±7.7 123.7±5.8 171.2±6.3 154.2±8.1 Cur (15 mg/kg) 129.1±8.0** 112.5±6.4* 177.0±7.8 148. 9±5.5 Cur (45 mg/kg) 115.7±6.8** 99.6±6.0** 168.7±7.1 146.8±6.7

Values were expressed as mean values±SEM of six to eight mice ## p <0.01 (compared with vehicle treated sham mice); * p <0.05 (compared with vehicle treated CCI mice); ** p <0.01 (compared with vehicle treated CCI mice) Psychopharmacology (2014) 231:2171–2187 2181

Fig. 6 Intrathecal (i.t.) and A intracerebroventricular (i.c.v.) Hargreaves Test administration of methysergide (Met), a nonselective 5-HT i.p. injection of i.t. injection of i.c.v. injection of receptor antagonist, separately 15 vehicle or Met vehicle or Met vehicle or Met abrogated the antinociceptive and antidepressant-like effects of ** ** curcumin (Cur). a Effect of co- 10 administrations of Met (i.p., i.t., or i.c.v.) on thermal latency in mice treated with vehicle or Cur. b 5 Effect of co-administrations of Met (i.p., i.t., or i.c.v.) on the

Thermal lantency (s) 0 immobility time measured in FST. Sham CCI Sham CCI Sham CCI Data are expressed as mean± +Cur +Cur +Cur +Cur +Cur +Cur SEM (n =8–12 per group), assessed by multifactor ANOVA followed by Duncan test. FST, Vehicle forced swim test Met (methysergide) B Forced swim test

i.p. injection of i.t. injection of i.c.v. injection of 240 vehicle or Met vehicle or Met vehicle or Met

** **

120

Immobility time (s) 0 Sham CCI Sham CCI Sham CCI +Cur +Cur +Cur +Cur +Cur +Cur

Acute blockade of GABAA receptors, but not mu-opioid dose-dependently attenuated in CCI mice the therapeutic effect receptors, abrogated the antidepressant-like effects of curcumin on thermal hyperalgesia in Hargreaves test of curcumin (F3,152=14.9, p <0.01). Nevertheless, β-FNAatthedoseof 5 mg/kg, which suffices to abolish the antihyperalgesic effect of As opioid receptors have been reported to be implicated in the curcumin, failed to affect the recuperating effect of curcumin on mechanisms underlying antidepressant-induced analgesia exacerbated depressive-like behavior in FST (Fig. 8b,right

(Marchand et al. 2003; Nozaki and Kamei 2006; Benbouzid panel). Conversely, the selective GABAA receptor antagonist et al. 2008), we previously showed that the mu-opioid receptors bicuculline (BIC, 0.3, 1 and 2 mg/kg, i.p.) greatly lessened (but are responsible for the antinociceptive effect of curcumin (Zhao not abolished) in CCI mice the antidepressant-like effect of et al. 2012). Thus, we investigated here whether acute blockade curcumin (F3,144=5.6, p <0.01, Fig. 8b, left panel) in a dose- of mu-opioid receptors by β-funaltrexamine (β-FNA, the irre- related manner, without influencing the antihyperalgesic effects versible mu-opioid receptor antagonist) could influence the of curcumin (Fig. 8a, right panel). Furthermore, neither antidepressant-like effect of curcumin in neuropathic mice. GABAB receptor baclofen nor antagonist phaclofen We also tested the possible involvement of GABAA receptors affected the antidepressant-like effect of curcumin (data not in the curcumin antidepression, inasmuch as GABA system and shown). These results suggest that the GABAA receptors are the GABAA receptors are extensively involved in the patho- critically involved in the therapeutic action of curcumin against physiology of major depression (Brambilla et al. 2003;Shen the comorbid depression induced by neuropathic pain. et al. 2010; Luscher et al. 2011; Vollenweider et al. 2011), as well as the mechanisms underlying the actions of antidepres- Effects of curcumin treatment on motor performance sants (Sanacora et al. 2002;Okamotoetal.2010; Tyagarajan and locomotor activity in CCI-injured mice et al. 2011) including curcumin (Gilhotra and Dhingra 2010). After 3 weeks curcumin treatment, the mice were challenged As shown in Fig. 9, peripheral nerve injury (CCI) did not alter for their acute response to opioid receptor or GABAergic the measures of motor performance (Fig. 9a) and locomotor receptor antagonists on day 36. As shown in Fig. 8a, β-FNA activity (Fig. 9b). Acute treatment of CCI mice with diazepam 2182 Psychopharmacology (2014) 231:2171–2187

Fig. 7 Supraspinal 5-HT1A A B receptor mediates the Vehicle or Cur (45 mg/kg) CCI mice antidepressant-like effect of 200 + i.p. WAY-100635 curcumin in mice with 240 Vehicle neuropathy. a Effect of co- WAY-100635 ** administration of 5-HT1A receptor ** antagonist WAY-100635 (1 mg/ 140 kg, i.p.) on measures of ** 120 depressive-like behavior (immobility time in FST) in mice treated with vehicle or Cur. b Immobility time (s) 80

WAY-100635 co-administration Immobility time (s) 0 (1 mg/kg, i.p., for four 35 37 39 41 0 5 15 45 consecutive days) abolished the Time (days after CCI) Cur (mg/kg) dose-dependent decrease of depressive-like behavior Sham + vehicle Sham + Cur (immobility time in FST) in CCI mice treated with Cur. c Effect of CCI + vehicle CCI + Cur the first acute WAY-100635 co- administration, done after C D Interruption of 3 weeks of Cur or vehicle i.p. WAY-100635 Vehicle treatment treatment (i.e. 36 days after CCI 200 200 or injury), on immobility time in Cur sham or CCI mice. d Effect of Cur treatment (45 mg/kg, p.o.) withdrawal on the depressive-like 140 140 behavior (immobility time in ** FST) in sham and CCI mice. Vehicle or Cur administration was

stopped after 21 days of treatment Immobility time (s) Immobility time (s) 80 80 (i.e., 36 days after CCI injury). e Effect of co-administration of 01234 35 37 39 41 WAY-100635, isamoltane, Time (h) Time (days after CCI) ritanserin, or ondansetron on depressive-like behavior (immobility time in FST) in CCI E F mice treated with Cur. f CCI mice Vehicle 240 Intrathecal (i.t.), but not WAY-100635 intracerebroventricular (i.c.v.), ** ** injection of WAY-100635 240 i.c.v. injection i.t. injection abolished the antidepressant-like effect of curcumin in CCI mice. 120 ** Data are expressed as mean± – SEM (n =8 12 per group), 120 assessed by multifactor ANOVA Immobility time (s) followed by Duncan test. FST 0 forced swim test Curcumin - + + + + +

WAY-100635 - - - - - + Immobility time (s) 0 Isamoltane - - - - + - Sham CCI Sham CCI Ritanserin - - - + - - +Cur +Cur +Cur +Cur Ondansetron - - + - - -

(Dia, 2 mg/kg, i.p., a positive reference drug) significantly test) and protracted immobility time (FST and TST), respec- decreased the staying time in rotarod test (F12,189=21.5, p < tively, should not be explained by possible side effects on 0.01, Fig. 9a) and locomotor activity (F 12,196=51.3, p <0.01, motor performance and locomotor activity. Fig. 9b), clearly showing its actions of sedation and motor inhibition. Nevertheless, neither acute nor chronic curcumin treatment influenced the motor performance (Fig. 9a) and Discussion locomotor activity (Fig. 9b) in neuropathic mice. The paucity of motor impairment and locomotor inhibition by curcumin Curcumin, a naturally occurring monomer, is well known for indicates its antinociceptive and antidepressant-like effects, its use in phytomedicine. In previous studies, we originally expressed by reversal of thermal hyperalgesia (Hargreaves showed the antidepressant-like activity of curcumin in mice Psychopharmacology (2014) 231:2171–2187 2183

Fig. 8 Chronic curcumin (Cur) A treatment requires GABAA Hargreaves Test receptors, but not mu-opioid β-FNA dose 12 0 0.5 1.5 5 12 0 2 BIC dose receptors (MORs), to ameliorate (mg/kg) (mg/kg) depressive-like behavior in neuropathic mice. a The first 10 10

acute co-administration of the ** MORs antagonist β- 8 ** 8 funaltrexamine (β-FNA, left panel)butnottheGABAA 6 6 receptor antagonist bicuculline (BIC, right panel), done after Thermal latency (s) 4 4 21 days of vehicle or Cur 0 30 0 30 0 30 0 30 0 30 0 30 treatment (i.e., day 36 after surgery), dose-dependently Time (min) Time (min) inhibited the therapeutic effect of CCI + vehicle curcumin (45 mg/kg, p.o.) on the pain-related behavior (thermal CCI + Cur hyperalgesia). b The first acute B co-administration of BIC (left Forced swim test panel) but not β-FNA (right 200 0 0.3 1 2 BIC dose 200 0 5 β-FNA dose panel), done after 21 days of (mg/kg) vehicle or Cur treatment, dose- (mg/kg) dependently inhibited the therapeutic effect of curcumin (45 mg/kg, p.o.) on the 140 140 depressive-like behavior (immobility time in FST). Data ** *

are expressed as mean±SEM (n = Immobility time (s) 8–12 per group), assessed by 80 80 multifactor ANOVA followed by 0 30 0 30 0 30 0 30 0 30 0 30 Duncan test. FST forced swim test Time (min) Time (min) and rats (Xu et al. 2005, 2007). As antidepressants are com- rapid-onset thermal hyperalgesia, the depressive-like behav- monly used as analgesics to treat chronic pain, we recently iors in CCI-injured mice, expressed as increased immobility revealed the antinociceptive property of curcumin (Zhao et al. time in FST and TST, developed relatively chronic (about 2012). In this work, we focused on the antidepressant-like 2 weeks), suggesting long-term events of neural plasticity effect of curcumin in the context of neuropathic pain and the may be involved. Paradoxically, although the comorbid de- underlying mechanisms. The data presented here showed that pression is ubiquitously observed in neuropathic patients, the curcumin, when chronically and orally dosed in mice, can results from animal models are not always consistent. Sub- normalize the depressive like behaviors induced by neuro- stantial studies reported the depressive-like behavior in neu- pathic pain, thus potentially providing a therapeutic alternative ropathic animals (Gonçalves et al. 2008;Huetal.2009; Yalcin for the treatment of comorbid depression caused by et al. 2011), but other studies provided controversial evidence chronic pain. It is noteworthy to point out that curcumin (Kontinen et al. 1999;Hasnieetal.2007). Additionally, the suffers from poor oral bioavailability due to low absorption onset of depressive-like behavior in neuropathic animals was into system circulation. An add-on therapy consisting of pip- not always temporally consistent in view of previous reports, erine can reduce this kinetic defect (Anand et al. 2007) and has e.g., 2 weeks following nerve injury reported by Suzuki et al. shown favorable efficacy to reduce neuropathic pain in a (2007), 3–4 weeks by Hu et al. (2009), and 6–8 weeks by recent controlled clinical trial (Di Pierro and Settembre 2013). Yalcin et al. (2011). This inconsistency has not been addressed The CCI model was used, in this study, to induce chronic precisely and may correlate with some factors, such as animal neuropathic pain, inasmuch as it can simulate clinical signs species/strains and age, animal models of neuropathic pain and symptoms in neuropathic patients and is widely used as a and tests used to examine depressive-like behavior. Herein, canonical animal model of neuropathic pain. Similarly, FST we showed the CCI mice exhibited concurrent pain-related and TST are two pharmacologically validated behavioral de- and depressive-like behaviors. This phenotype of altered emo- spair tests in mice and rats (Porsolt et al. 1977). After CCI tional conditions should not be ascribed to the motor dysfunc- injury, the mice exhibited rapid-onset (within 5 days) thermal tion, as there are no alterations of motor coordination (rotarod hyperalgesia in Hargreaves test, which is in agreement with test) and of measures of locomotor activity after peripheral our previous results (Zhao et al. 2012). Compared with the nerve injury. Thus, we may propose here the suitability of CCI 2184 Psychopharmacology (2014) 231:2171–2187

A Rota-rod Test An important mechanistic concern, in this study, is whether Drug administration (Vehicle/Cur/Dia) the antidepressant-like effect of curcumin correlates with its 120 paralleled antinociception. Clinical studies suggest that the therapeutic efficacy of antidepressants on chronic pain is not associated with their antidepressant potentials (Max et al. 60 ** ** 1987;Atkinsonetal.1998). Recently, Hu et al. (2009)com- ** ** 0 pared the antidepressant-like effects of cannabinoid CB2- 0 1 2 34 selective agonist GW405833 with , a clinical Time (h) available antidepressant, in rats with mononeuropathy, and 120 Cur (45 mg/kg) or Vehicle they showed the antidepressant-like actions exerted by the Staying time (s) Staying two agents were distinct in modality and mechanism of ac- 60 tions. In another study by Jesse et al. (2010), the antiallodynic and antidepressant-like effects of , ,

0 and were compared following acute administration 0 15 20 25 30 35 40 in mice with mononeuropathy. All of the three drugs produced Time (days after CCI) antiallodynic effects in neuropathic mice and produced CCI + vehicle antidepressant-like effects in sham mice, but only amitripty- CCI + Cur (45 mg/kg) line exhibited antidepressant-like activity in neuropathic mice. B CCI + Dia (2 mg/kg) These lines of evidence may indicate the segregation of the Locomotor Activity analgesic and antidepressant-like actions by antidepressants in Drug administration (Vehicle/Cur/Dia) the context of chronic neuropathic pain. This assertion is 600 supported by the current study, as the antinociception and anti-depression exerted by curcumin, in neuropathic mice, 300 * may be pharmacologically segregated, i.e., the antidepression ** ** ** of curcumin is independent of its paralleled antinociception in 0 0 1 2 34 the context of mononeuropathy. Supporting evidence is de- Time (h) rived from the fact that the two paralleled actions of curcumin Cur (45 mg/kg) or Vehicle were discretely influenced by repeated i.t. and i.c.v. adminis- 600 tration of methysergide, a nonselective 5-HT receptor antago- Locomotion counts nist, as well as by acute blockade of mu-opioid receptors and 300 GABAA receptors, respectively. The impairment or disability of brain 5-HTor NA may be a 0 0 15 20 25 30 35 40 potential pathophysiological mechanism for the comorbid Time (days after CCI) depression induced by chronic pain (Kim et al. 2012;Alba- Fig. 9 Effects of curcumin (Cur) treatment on motor performance and Delgado et al. 2013). Given that the monoaminergic mecha- locomotor activity in CCI-operated mice. Cur or vehicle treatment began nisms are implicated in the antidepressant-like and analgesic on day 15 (i.e., 15 days after surgery). For acute Cur or vehicle treatment, properties of curcumin (Xu et al. 2005, 2007; Zhao et al. behavioral tests (rotarod and locomotor activity) were performed from 0.5 to 4 h after Cur or vehicle administration. For chronic Cur or vehicle 2012), we tentatively hypothesize the involvement of mono- treatment, behavioral tests were performed 1 h before first curcumin aminergic mechanisms in the antidepressant-like effects by administration in the morning. a Neither acute nor chronic Cur treatment curcumin in the context of chronic neuropathic pain. This influenced the motor performance (rotarod test) in CCI mice, while acute mechanistic supposition for curcumin action was corroborated diazepam (Dia, 2 mg/kg, p.o.) administration exerted significant inhibi- tory effect. b Neither acute nor chronic Cur treatment influenced the by several pieces of evidence: (a) Chemical depletion of brain locomotor activity in CCI mice, while acute Dia (2 mg/kg, p.o.) admin- 5-HT abolished the antidepressant-like effect of curcumin in istration significantly decreased it. Data are expressed as mean±SEM neuropathic mice; (b) Chronic curcumin treatment increased, – (n =8 12 per group), assessed by multifactor ANOVA followed by in neuropathic mice, the levels of brain 5-HT and the ratio of Duncan test 5-HT/5-HIAA, and concurrently decreased the activity of MAO-A, which preferentially metabolizes 5-HTand NA rath- model from Bennett and Xie (1988) for mimicking the co- er than dopamine in brain (Johnston 1968); (c) Repeated morbidity of depression and chronic neuropathic pain. co-administration of non-selective 5-HT antagonist methyser-

More studies should be performed to ascertain other gide or selective 5-HT1A antagonist WAY-100635 abrogated symptoms of comorbid depression caused by chronic pain, the antidepressant-like effect of curcumin. Furthermore, we such as anhedonia and dysfunction of social communication, showed here the anti-hyperalgesic effect to heat stimuli by and the effects of drugs. curcumin is also dependent on serotonergic system, as the Psychopharmacology (2014) 231:2171–2187 2185 modality specific antinociception by curcumin is reversed by accumulating evidence indicated the involvement of either chemical depletion of 5-HT or blocking 5-HT receptor. GABAergic mechanisms in the chronic antidepressant (espe- The shared serotonergic mechanism by both antinociceptive cially SSRI, e.g., fluoxetine) treatment (Ye et al. 2008;Möhler and antidepressant-like effects of curcumin is consistent with 2012), as well as in mediating the activation of 5-HT receptors a previous study on venlafaxine (Berrocoso and Mico 2009) in the context of neuropathic pain (Viguier et al. 2012). and indicates again the vital relationship between chronic neu- Chronic antidepressant treatment may potentiate some path- ropathic pain and the comorbid affective diseases. However, the ways of signal transduction, such as wnt2 and GSK3β,and two actions, i.e., antinociception and anti-depression, exerted eventually enhance GABAA receptor clustering and by curcumin may be pharmacologically segregated (see afore- GABAergic synaptic inhibition (Okamoto et al. 2010; mentioned discussion). Future studies should be undertaken to Tyagarajan et al. 2011). Thus, the GABAergic mechanism ascertain the precise mechanism on the relevance and segrega- and the activation of GABAA receptor may be an alternative tion of the two concomitant actions by curcumin. downstream target for curcumin action. This is confirmed by To further explore the serotonergic mechanisms behind another series of pharmacological experiments of ours, as curcumin action, we tested different 5-HT receptor antagonists acute co-administration of GABAA receptor antagonist to identify which subtype(s) of 5-HT receptors is(are) respon- bicuculline greatly attenuated the antidepressant-like effect sible for the antidepressant-like effect of curcumin in the of curcumin. The involvement of GABAA receptor in context of neuropathic pain. As shown in our pharmacological curcumin action should be antidepression specific, since the experiments, repeated co-administration of the 5-HT1A recep- same bicuculline challenge did not affect the concurrent anal- tor antagonist WAY-100635 totally abolished the gesic effect of curcumin. Given that the complete abolishment antidepressant-like effect of curcumin, but this is not the case of the curcumin antidepression requires sustained serotonin for 5-HT1B receptor antagonist isamoltane, 5-HT2A/C receptor suppression, we here propose that the GABAergic mechanism antagonist ritanserin, and 5-HT3 receptor antagonist and GABAA receptors may be rendered as the downstream ondansetron. These results strongly argue for an essential role targets, after that the serotonergic tone is upregulated. for 5-HT1A receptor in mediating the antidepressant-like effect In conclusion, the present study demonstrates that curcumin, of curcumin. More precisely, we localized the responding 5- when dosed orally and chronically in mice with

HT1A receptor to supraspinal but not spinal level, as i.c.v. mononeuropathy, can correct their exaggerated depressive-like rather than i.t. injection of WAY-100635 abolished the and pain-related (heat hyperalgesia) behaviors, indicating a pos- antidepressant-like effect of curcumin. The exclusive 5- sible utility of this naturally occurring compound in the clinical

HT1A mechanism responsible for the antidepressant-like ef- setting of comorbid depression induced by neuropathic pain. fects of curcumin indicates again the intimate association Although further studies should be undertaken to elucidate the between 5-HT1A receptor and the etiology of depression, as precise mechanisms for curcumin antidepression in pain-related well as the mechanisms of antidepressants actions (Savitz context, we propose that the antidepressant-like effect of et al. 2009). However, Berrocoso et al. (2006, 2009)showed curcumin may be independent of its paralleled antinociception that WAY-100635 had an opposite influence on the and mediated by supraspinal serotonergic system (coupled to 5- antinociceptive and the antidepressant properties of HT1A receptors) and downstream GABAA receptors. and of venlafaxine. This discrepancy may be ascribed to acute vs. repeated drug regimen and preferential effects engendered Acknowledgments This work was supported by the K.C. Wong Magna by these drugs on presynaptic or postsynaptic 5-HT recep- Fund in Ningbo University, National Natural Science Foundation of China 1A (81201050), Natural Science Foundation of Zhejiang Province tor (e.g., somatodendritic autoreceptors in raphe nuclei vs. (LY12H09002), and Innovative Research Team of Ningbo (2009B21002). postsynaptic receptors located in the forebrain and hippocam- pus areas). Notably, the abolishment of antidepressant-like Conflict of interest The authors declare no conflicts of interest. effect exerted by curcumin requires a repeated procedure of WAY-100635 co-administration, which is analogous, as for the modality and time course, to the relapse of depressive-like References behavior observed after curcumin withdrawal. This temporal delay indicates that the escalation of serotonergic tone Alba-Delgado C, Llorca-Torralba M, Horrillo I, Ortega JE, Mico JA, (coupled with 5-HT1A receptor) may be a primary or upstream Sánchez-Blázquez P, Meana JJ, Berrocoso E (2013) Chronic pain target for curcumin action, and the recruitment of downstream leads to concomitant noradrenergic impairment and mood disorders. targets should be required. Biol Psychiatry 73:54–62 The GABA neurotransmitter constitutes the predominant Anand P, Kunnumakkara AB, Newman RA, Aggarwal BB (2007) inhibitory neural transmission in the brain and has been sub- Bioavailability of curcumin: problems and promises. Mol Pharm 4:807–818 stantially implicated in the regulation of mood and emotion Arnér S, Meyerson BA (1988) Lack of analgesic effect of opioids on (Brambilla et al. 2003;Shenetal.2010). Recently, neuropathic and idiopathic forms of pain. Pain 33:11–23 2186 Psychopharmacology (2014) 231:2171–2187

Arnow BA, Hunkeler EM, Blasey CM, Lee J, Constantino MJ, Fireman Hasnie FS, Wallace VC, Hefner K, Holmes A, Rice AS (2007) B, Kraemer HC, Dea R, Robinson R, Hayward C (2006) Comorbid Mechanical and cold hypersensitivity in nerve-injured C57BL/6J depression, chronic pain, and disability in primary care. Psychosom mice is not associated with fear-avoidance- and depression-related Med 68:262–268 behaviour. Br J Anaesth 98:816–822 Atkinson JH, Slater MA, Williams RA, Zisook S, Patterson TL, Grant I, Hu B, Doods H, Treede RD, Ceci A (2009) Depression-like behaviour in Wahlgren DR, Abramson I, Garfin SR (1998) A placebo-controlled rats with mononeuropathy is reduced by the CB2-selective agonist randomized clinical trial of for chronic low back pain. GW405833. Pain 143:206–212 Pain 76:287–296 Hylden JL, Wilcox GL (1980) Intrathecal morphine in mice: a new Attal N, Cruccu G, Baron R, Haanpää M, Hansson P, Jensen TS, technique. Eur J Pharmacol 67:313–316 Nurmikko T (2010) EFNS guidelines on the pharmacological treat- Jackson KC 2nd (2006) Pharmacotherapy for neuropathic pain. Pain ment of neuropathic pain: revision. Eur J Neurol 17:1113–1123 Pract 6:27–33 Bair MJ, Robinson RL, Katon W, Kroenke K (2003) Depression and pain Jesse CR, Wilhelm EA, Nogueira CW (2010) Depression-like behavior and comorbidity: a literature review. Arch Intern Med 163:2433–2445 mechanical allodynia are reduced by bis selenide treatment in mice Benbouzid M, Gavériaux-Ruff C, Yalcin I, Waltisperger E, Tessier LH, with chronic constriction injury: a comparison with fluoxetine, ami- Muller A, Kieffer BL, Freund-Mercier MJ, Barrot M (2008) Delta- triptyline, and bupropion. Psychopharmacology (Berl) 212:513–522 opioid receptors are critical for treatment of Johnston JP (1968) Some observations upon a new inhibitor of mono- neuropathic allodynia. Biol Psychiatry 63:633–636 amine oxidase in brain tissue. Biochem Pharmacol 17:1285–1297 Bennett GJ, Xie YK (1988) A peripheral mononeuropathy in rat that Kim H, Chen L, Lim G, Sung B, Wang S, McCabe MF, Rusanescu G, produces disorders of pain sensation like those seen in man. Pain 33: Yang L, Tian Y, Mao J (2012) Brain indoleamine 2,3-dioxygenase 87–107 contributes to the comorbidity of pain and depression. J Clin Invest Berrocoso E, Mico JA (2009) Role of serotonin 5-HT1A receptors in the 122:2940–2954 antidepressant-like effect and the antinociceptive effect of Kontinen VK, Kauppila T, Paananen S, Pertovaara A, Kalso E (1999) venlafaxine in mice. Int J Neuropsychopharmacol 12:61–71 Behavioural measures of depression and anxiety in rats with spinal Berrocoso E, Rojas-Corrales MO, Mico JA (2006) Differential role of 5- nerve ligation-induced neuropathy. Pain 80:341–346 HT1A and 5-HT1B receptors on the antinociceptive and antidepres- Leo RJ (2005) Chronic pain and comorbid depression. Curr Treat Options sant effect of tramadol in mice. Psychopharmacology (Berl) 188: Neurol 7:403–412 111–118 Luscher B, Shen Q, Sahir N (2011) The GABAergic deficit hypothesis of Boyce-Rustay JM, Palachick B, Hefner K, Chen YC, Karlsson RM, major depressive disorder. Mol Psychiatry 16:383–406 Millstein RA, Harvey-White J, Holmes A (2008) Desipramine Magni G, Moreschi C, Rigatti-Luchini S, Merskey H (1994) Prospective potentiation of the acute depressant effects of ethanol: modulation study on the relationship between depressive symptoms and chronic by alpha2-adrenoreceptors and stress. Neuropharmacology 55:803– musculoskeletal pain. Pain 56:289–297 811 Marchand F, Ardid D, Chapuy E, Alloui A, Jourdan D, Eschalier A Brambilla P, Perez J, Barale F, Schettini G, Soares JC (2003) GABAergic (2003) Evidence for an involvement of supraspinal delta- and spinal dysfunction in mood disorders. Mol Psychiatry 8:721–737 mu-opioid receptors in the antihyperalgesic effect of chronically De Souza MM, Pereira MA, Ardenghi JV,Mora TC, Bresciani LF, Yunes administered in mononeuropathic rats. J Pharmacol RA, Delle Monache F, Cechinel-Filho V (2009) Filicene obtained ExpTher307:268–274 from Adiantum cuneatum interacts with the cholinergic, dopaminer- Max MB, Culnane M, Schafer SC, Gracely RH, Walther DJ, Smoller B, gic, glutamatergic, GABAergic, and tachykinergic systems to exert Dubner R (1987) Amitriptyline relieves diabetic neuropathy pain in antinociceptive effect in mice. Pharmacol Biochem Behav 93:40–46 patients with normal or depressed mood. Neurology 37:589–596 Di Pierro F, Settembre R (2013) Safety and efficacy of an add-on therapy Micó JA, Ardid D, Berrocoso E, Eschalier A (2006) Antidepressants and with curcumin phytosome and piperine and/or lipoic acid in subjects pain. Trends Pharmacol Sci 27:348–354 with a diagnosis of peripheral neuropathy treated with dexibuprofen. Möhler H (2012) The GABA system in anxiety and depression and its J Pain Res 6:497–503 therapeutic potential. Neuropharmacology 62:42–53 Duarte FS, Lach G, Martins PR, Romeiro GA, de Lima TC (2008) Nozaki C, Kamei J (2006) Possible involvement of opioidergic systems Evidence for the involvement of the monoaminergic system in the in the anti-nociceptive effect of the selective serotonin reuptake antidepressant-like action of two 4-amine derivatives of 10,11- inhibitors in sciatic nerve- injured mice. Eur J Pharmacol 552:99– dihydro-5H-dibenzo [a, d] cycloheptane in mice evaluated in the 104 tail suspension test. Prog Neuropsychopharmacol Biol Psychiatry Ohayon MM, Schatzberg AF (2003) Using chronic pain to predict 32:368–374 depressive morbidity in the general population. Arch Gen Gilhotra N, Dhingra D (2010) GABAergic and nitriergic modulation by Psychiatry 60:39–47 curcumin for its antianxiety-like activity in mice. Brain Res 1352: Okamoto H, Voleti B, Banasr M, Sarhan M, Duric V, Girgenti MJ, 167–175 Dileone RJ, Newton SS, Duman RS (2010) Wnt2 expression and Gonçalves L, Silva R, Pinto-Ribeiro F, Pêgo JM, Bessa JM, Pertovaara A, signaling is increased by different classes of antidepressant treat- Sousa N, Almeida A (2008) Neuropathic pain is associated with ments. Biol Psychiatry 68:521–527 depressive behaviour and induces neuroplasticity in the amygdala of Porsolt RD, Le Pichon M, Jalfre M (1977) Depression: a new animal the rat. Exp Neurol 213:48–56 model sensitive to antidepressant treatments. Nature 266:730–732 Gustorff B, Dorner T, Likar R, Grisold W, Lawrence K, Schwarz F, Radhakrishna Pillai G, Srivastava AS, Hassanein TI, Chauhan DP, Rieder A (2008) Prevalence of self-reported neuropathic pain and Carrier E (2004) Induction of apoptosis in human lung cancer cells impact on quality of life: a prospective representative survey. Acta by curcumin. Cancer Lett 208:163–170 Anaesthesiol Scand 52:132–136 Reeta KH, Mehla J, Gupta YK (2010) Curcumin ameliorates cognitive Haley TJ, Mccormick WG (1957) Pharmacological effects produced by dysfunction and oxidative damage in phenobarbitone and carbamaz- intracerebral injection of drugs in the conscious mouse. Br J epine administered rats. Eur J Pharmacol 644:106–112 Pharmacol Chemother 12:12–15 Sanacora G, Mason GF, Rothman DL, Krystal JH (2002) Increased Hargreaves K, Dubner R, Brown F, Flores C, Joris J (1988) A new and occipital cortex GABA concentrations in depressed patients after sensitive method for measuring thermal nociception in cutaneous therapy with selective serotonin reuptake inhibitors. Am J hyperalgesia. Pain 32:77–88 Psychiatry 159:663–665 Psychopharmacology (2014) 231:2171–2187 2187

Savitz J, Lucki I, Drevets WC (2009) 5-HT1A receptor function in major Vo T, Rice AS, Dworkin RH (2009) Non-steroidal anti-inflammatory depressive disorder. Prog Neurobiol 88:17–31 drugs for neuropathic pain: how do we explain continued wide- Shen Q, Lal R, Luellen BA, Earnheart JC, Andrews AM, Luscher B spread use? Pain 143:169–171 (2010) Gamma-aminobutyric acid-type A receptor deficits cause Vollenweider I, Smith KS, Keist R, Rudolph U (2011) Antidepressant- hypothalamic–pituitary–adrenal axis hyperactivity and antidepres- like properties of α2-containing GABA(A) receptors. Behav Brain sant drug sensitivity reminiscent of melancholic forms of depres- Res 217:77–80 sion. Biol Psychiatry 68:512–520 Wang S, Tian Y, Song L, Lim G, Tan Y, You Z, Chen L, Mao J (2012) Shishodia S, Sethi G, Aggarwal BB (2005) Curcumin: getting back to the Exacerbated mechanical hyperalgesia in rats with genetically roots. Ann N Y Acad Sci 1056:206–217 predisposed depressive behavior: role of melatonin and NMDA Steru L, Chermat R, Thierry B, Simon P (1985) The tail suspen- receptors. Pain 153:2448–2457 sion test: a new method for screening antidepressants in Xu Y, Ku BS, Yao HY, Lin YH, Ma X, Zhang YH, Li XJ (2005) The mice. Psychopharmacology (Berl) 85:367–370 effects of curcumin on depressive-like behaviors in mice. Eur J Suzuki T, Li YH, Mashimo T (2005) The antiallodynic and Pharmacol 518:40–46 antihyperalgesic effects of neurotropin in mice with spinal nerve Xu Y, Ku B, Cui L, Li X, Barish PA, Foster TC, Ogle WO (2007) ligation. Anesth Analg 101:793–799 Curcumin reverses impaired hippocampal neurogenesis and increases Suzuki T, Amata M, Sakaue G, Nishimura S, Inoue T, Shibata M, serotonin receptor 1A mRNA and brain-derived neurotrophic factor Mashimo T (2007) Experimental neuropathy in mice is associated expression in chronically stressed rats. Brain Res 1162:9–18 with delayed behavioral changes related to anxiety and depression. Yadav VS, Mishra KP, Singh DP, Mehrotra S, Singh VK (2005) Anesth Analg 104:1570–1577 Immunomodulatory effects of curcumin. Immunopharmacol Tanabe M, Tokuda Y, Takasu K, Ono K, Honda M, Ono H (2007) The Immunotoxicol 27:485–497 synthetic TRH analogue taltirelin exerts modality-specific Yalcin I, Tessier LH, Petit-Demoulière N, Doridot S, Hein L, Freund- antinociceptive effects via distinct descending monoaminergic sys- Mercier MJ, Barrot M (2009) Beta2-adrenoceptors are essential for tems. Br J Pharmacol 150:403–414 desipramine, venlafaxine or reboxetine action in neuropathic pain. Tyagarajan SK, Ghosh H, Yévenes GE, Nikonenko I, Ebeling C, Neurobiol Dis 33:386–394 Schwerdel C, Sidler C, Zeilhofer HU, Gerrits B, Muller D, Yalcin I, Bohren Y, Waltisperger E, Sage-Ciocca D, Yin JC, Freund- Fritschy JM (2011) Regulation of GABAergic synapse formation Mercier MJ, Barrot M (2011) A time-dependent history of mood and plasticity by GSK3beta-dependent phosphorylation of gephyrin. disorders in a murine model of neuropathic pain. Biol Psychiatry 70: Proc Natl Acad Sci U S A 108:379–384 946–953 Viguier F, Michot B, Kayser V, Bernard JF, Vela JM, Hamon M, Ye ZY, Zhou KQ, Xu TL, Zhou JN (2008) Fluoxetine potentiates Bourgoin S (2012) GABA, but not opioids, mediates the GABAergic IPSCs in rat hippocampal neurons. Neurosci Lett 442: anti-hyperalgesic effects of 5-HT7 receptor activation in rats 24–29 suffering from neuropathic pain. Neuropharmacology 63: Zhao X, Xu Y, Zhao Q, Chen CR, Liu AM, Huang ZL (2012) Curcumin 1093–1106 exerts antinociceptive effects in a mouse model of neuropathic pain: Villarinho JG, Pinheiro KD, Pinheiro FD, Oliveira SM, Machado P, descending monoamine system and opioid receptors are differen- Martins MA, Bonacorso HG, Zanatta N, Fachinetto R, Ferreira J tially involved. Neuropharmacology 62:843–854 (2013) The antinociceptive effect of reversible monoamine oxidase- ZhenL,ZhuJ,ZhaoX,HuangW,AnY,LiS,DuX,LinM,WangQ,XuY, A inhibitors in a mouse neuropathic pain model. Prog Pan J (2012) The antidepressant-like effect of fisetin involves the Neuropsychopharmacol Biol Psychiatry 44:136–142 serotonergic and noradrenergic system. Behav Brain Res 228:359–366