Melatonin alleviates morphine analgesic tolerance in mice by decreasing NLRP3 inflammasome activation.

Morphine is frequently used for pain relief, but long-term morphine therapy in patients with chronic pain results in analgesic tolerance and hyperalgesia. There are no effective therapeutic treatments that limit these detrimental side effects. We found pretreatment with melatonin could decrease morphine-induced analgesic tolerance. There was a significant activation of the NLRP3 inflammasome in the prefrontal cortex and the peripheral blood of morphine-treated mice compared to control animals, which could be blocked by melatonin. The inflammasome activation induced by morphine was mediated by the microglia. SiRNA knockdown or pharmacological inhibition of the NLRP3 abolished the morphine-induced inflammasome activation. Co-administration of melatonin and low-dose morphine had better analgesia effects in the murine models of pain and led to a lower NLRP3 inflammasome activity in brain tissues. Mice deficient for Nlrp3 had a higher nociceptive threshold and were less sensitive to develop morphine-induced analgesic tolerance and acetic acid-induced pain relative to wild-type animals. Concordantly, we observed a significantly elevated level of serum IL-1β, which indicates an increase of NLRP3 inflammasome activity associated with the reduced level of serum melatonin, in heroin-addicted patients relative to healthy individuals. Our results provide a solid basis for conducting a clinical trial with the co-administration of melatonin and morphine for the relief of severe pain.

Tolerance and hyperalgesia are the two main detrimental side effects of morphine treatment and severely limits the clinical usage of the drug [5].
Accumulating evidence suggests that neuroinflammatory are critical for the morphine induced analgesic tolerance and hyperalgesia [6,7]. The NOD-like receptor protein 3 (NLRP3) inflammasome is the beststudied inflammasome which composed of the NLRP3 sensor, caspase-1 and the adaptor molecule apoptosis associated speck-like protein containing a caspase recruitment domain (ASC) [8,9]. Upon NLRP3 inflammasome activation, matured caspase-1 mediates Pro-IL-1β cleavage into mature IL-1β and secretion [8,10]. Dysregulation of the NLRP3 can lead to autoimmune diseases, neurodegenerative diseases, multiple sclerosis and metabolic disorders [11]. Previous study showed that morphine paradoxically prolongs neuropathic pain in rats by amplifying spinal NLRP3 inflammasome activation [12]. However, whether the NLRP3 inflammasome participates in morphine analgesic tolerance is still unclear. As a well-known mitochondrial targeted antioxidant, melatonin can cross the blood-brain barrier and is involved in neuronal protection [13,14], regulation of circadian rhythms [15][16][17], and antinociception [18]. Melatonin also provides an antiinflammatory effect [19,20]. Emerging evidence suggested that melatonin can decrease the antinociceptive tolerance induced by morphine in mouse models [21][22][23], but the exact mechanism of the rescue action induced by this drug has not been fully understood. The lack of a scientific basis has therefore prevented any attempt to use melatonin as an efficient treatment for the alleviation of morphine-induced analgesic tolerance.
In this study, we defined the role of melatonin in morphine-induced analgesic tolerance and hypothesized that aberrant activation of NLRP3 inflammasome may contribute to morphine analgesic tolerance. We first examined the possible ameliorating effect of co-administration of melatonin with morphine on morphine-induced tolerance, then we investigated the role of NLRP3 inflammasome activation in microglia during melatonin-induced reduction of morphine-induced tolerance. We found that the NLRP3 inflammasome plays an important role in antinociceptive tolerance and melatonin pretreatment could decrease morphine-induced NLRP3 inflammasome activity. Deficiency of Nlrp3 in mice blunted morphine-induced analgesic tolerance and acetic acidinduced pain. Our results uncovered the molecular mechanism how melatonin decreases morphine induced-analgesic tolerance.

Reagents, cells and drug treatment
The primary antibodies and chemicals used in this study are listed in Table S1. The BV2 cells were obtained from the Kunming Cell Bank, Kunming Institute of Zoology (KIZ). Melatonin and nigericin were dissolved in ethanol. For melatonin treatment, cells were pretreated with 200 μM melatonin for 30 min before morphine or Lipopolysaccharide (LPS) treatment. LPS was dissolved in endotoxin-free water and 1 μg/ mL LPS was added into medium to treat cells for 6 h. Nigericin (15 μM) was added into medium for 30 min after other drug treatment in order to further stimulate the second signal for activating NLRP3 inflammasomes [24].

Isolation of mouse primary microglia
Mouse primary microglia were prepared and cultured as previously described [25]. Briefly, brain cortices from 1-day old neonatal mice were dissociated with 1-mL pipettes. Debris was removed by filtration with a 70-μm cell strainer (Falcon). Cells were cultured in DMEM plus 10% fetal bovine serum (FBS, v/v) supplemented with 100 IU/mL penicillin and 100 μg/mL streptomycin. After 10 days, confluent mixed glial cultures were shaken at 37 o C for 2 h to promote microglia detachment. Culture medium containing released microglia cells was aspirated, centrifuged at 1000 g for 5 min, and collected microglia were subsequently plated onto poly-D, L-ornithine-coated P100 plates for growth. Primary microglia were stained with antibody to IBA-1 (antimouse IBA-1, Merck Millipore, MABN92, 1:500) and had a purity > 95%.

Heroin-addicted patients
Six female heroin addicts (mean age 31.7 ± 2.4 years) at the withdrawal and rehabilitation stages were recruited from the Kunming Rehabilitation Center on Drug Dependence in Yunnan Province, China. Six drug-free healthy volunteers (mean age 30 ± 1.7 years), without any history of medical, neurological or psychiatric diseases, were used as the control samples. All the subjects provided informed consent prior to the study. These heroin-addicted patients and controls had been described in our previous studies [21,26]. The institutional review board of Kunming Institute of Zoology approved this study (SWYX-20090302-03).

Animal models
The Nlrp3 −/− mouse was created using C57BL/6 as the genetic background and this mouse was previously described [27]. For all behavioral tests, C57BL/6 (as the wild-type animals: WT) and Nlrp3 −/− mice (body weight: 25-30 g) at an age of 8 weeks were used. Mice were housed in clear plastic cages with free access to water and food in an established animal house of 22 ± 2°C and 50% humidity, with a 12-h light/dark cycle. Data acquisition and analyses were performed with a double-blind, controlled design. The related information for grouping of animals, number of animals and the treatments in this study were listed in Table S2 and Table S3. All experiments involving animals and the animal care and experimental protocol were approved by the Institutional Review Board of Kunming Institute of Zoology, Chinese Academy of Sciences (SYDW-2009017, RTSW-2009018 and SMKX-2018011).

Morphine analgesia and tolerance
Morphine hydrochloride was diluted in saline (0.9% NaC1). Melatonin was dissolved in saline with 0.5% ethanol (v/v), and saline with 0.5% ethanol was used as a vehicle. We chose the concentrations of melatonin (0.5 mg/kg) and morphine (10 mg/kg) that had the best effect according to our previous study [21]. Melatonin was injected intraperitoneally 30 min before subcutaneous injection of morphine [21]. Analgesia was assessed using the radiant heat tail-flick latency (Tail-Flick Unit 37360, UGO Baseline, Comerio, Italy) and a hot platform (YLS-21A, Beijing Dameida Technology Co., Ltd, China) as previously described [28]. For the tail-flick test, mice were placed in Plexiglas cages (9 × 6 × 3 cm) on a modified Hargreaves Device. Mice were habituated to the device for 2 min before each test session. A halogen lamp was focused on the tail and the withdrawal reflex time was determined by a photocell. We measured tail flick latency at IR30 and IR50 (decimal selector of heat intensity) every two days from day 0 (baseline) to day 7, 14 and 21. We used 20 s as a cutoff time to avoid damaging the tail. Baseline responses were determined for each mouse before drug injection. For morphine hot plate tolerance, repeated morphine injections (10 mg/kg subcutaneously) were given daily for 7, 14 and 21 days. The hot plate test was performed on a platform heated to 47.5°C, 50°C and 52.5°C with a cutoff of 30 s, and the latency to paw lick or jump was recorded. The baseline response was determined for each animal before treatment. The analgesic response to morphine was assessed by the hot plate test or tail-flick test at 30 min or 60 min after morphine injection (10 mg/kg) [21,29]. For cross-tolerance test, the tolerance was established by repeated injections of either melatonin (or morphine) alone or the saline (or vehicle) twice a day for 2 days. On day 3, melatonin or morphine was injected to mice, followed by the tail-flick test and hot plate test as previously described [30].

Acetic acid-induced writhing test
Pain sensitivity was evaluated by measuring the acetic acid-induced writhing responses as previously described [31], as acetic acid could induce abdominal contractions and hind limb stretching. Mice were placed into open polyvinyl cages (20 × 40 × 15 cm) immediately after acid challenge, and abdominal constrictions were counted cumulatively over a period of 30 min.

Tissue preparation
Mice were anesthetized with pentobarbital at the end of the cycle of drug treatment. After having collected blood by the heart punctures, mice were intracardially perfused with ice-cold saline. Brain was extracted and split into two hemispheres. We isolated prefrontal cortex and other brain tissues from one hemisphere for protein and mRNA analyses. Tissues were frozen in liquid nitrogen and stored at −80°C. Another hemisphere was used for immunofluorescence assay. Briefly, tissues were fixed in 4% paraformaldehyde until further processing.

RNA interference and transfection
siRNA oligos against Nlrp3 (GenePharma, Nlrp3-Mus-384 siRNA oligo) and negative control siRNA (siNC, which was designed to have no known mRNA targets in the cells being used) were obtained from GenePharma (Suzhou, China). The BV2 cells were transfected according to the procedure in our previous studies [32,33]. In brief, cells (1 × 10 5 per well) were seeded in 12-well plates to grow to 50% confluence. Before transfection, culture medium was removed and washed once with Opti-MEM medium (Gibco-BRL, 31985-070). The Nlrp3 siRNA or NC siRNA was dissolved in Opti-MEM medium, and was then mixed with 3 μL Lipofectamine TM 3000 (Invitrogen, L3000008) to achieve a final volume of 100 μL. The siRNA-Lipofectamine mixture was incubated at room temperature for 20 min, and added to each well together with an additional 400 μL Opti-MEM medium. The medium was removed at 6 h after transfection and fresh growth medium was added (1 mL/well) for cell growth. We optimized the siRNA concentration for transfection at a concentration of 50 nM and the time for cell harvest at 48 h.

Assays for reactive oxygen species (ROS) and pyroptotic cell death
The BV2 cells and primary microglia were cultured in 1640 or DMEM medium plus with 10% FBS supplemented with 100 IU/mL penicillin and 100 μg/mL streptomycin in a humidified atmosphere incubator with 5% CO 2 at 37°C. After morphine treatment for 6 h, cells were harvested and washed once with phosphate-buffer saline (PBS). DCFH-DA (2 μM) was used to examine intracellular ROS level. Cells were incubated with the dye at 37°C for 20 min, then washed twice with PBS, resuspended in PBS and kept on ice for an immediate detection on the FACScan (Becton Dickinson, USA). The CytoTox 96 Non-Radioactive Cytotoxicity Assay kit (Promega) was used to assess lactate dehydrogenase (LDH) release to measure cell death according to the manufacturer's instructions.

Quantitative real-time PCR
Total RNA was isolated from cell using TRIZOL (Invitrogen, 15596-018). Total RNA (1 μg) was used to synthesize single-strand cDNA using the M-MLV Reverse Transcriptase (Promega, M170A) in a final volume of 25 μL according to the manufacturer's instructions. The relative mRNA levels of Nlrp3, IL-1β, Asc, Caspase-1, Irf7, Tlr2 and Gsdmd were quantified by using quantitative real-time PCR, with normalization to the Actb/β-actin gene. The quantitative real-time PCR was performed in a total volume of 20 μL containing 2 μL of diluted products, 10 μL of SYBR Master Mix (Takara), 0.2 uL 10 μM each primer (Table S4), on an BIO-RAD Real-time PCR detection system. The quantitative real-time PCR thermal cycling conditions were composed of a denaturation cycle at 95°C for 5 min, followed by 40 cycles of 95°C for 10 s and 57°C for 30 s. For microglia single-cell quantitative real-time PCR analysis, we isolated microglial cells as previously described [34] with some modifications. In brief, adult mice were transcardially perfused with icecold PBS and cortex tissues were collected from control or drug-treated mice in sterile PBS containing penicillin and streptomycin (Gibco, 15140-122). The tissues were washed 5 times with PBS in clean bench, scissored into smaller pieces and enzymatically digested with collagenase IV (1 mg/mL, Thermo Fisher Scientific) and DNase I (0.2 mg/ mL, Sigma). The tissues were incubated in 37°C water bath for 15 min with gentle agitation at a 5-min interval. Single cell suspension was washed 3 times followed by filtration with 40-μm nylon cell strainer. Cell suspensions in 10 mL of cold 0.9 M sucrose solution were centrifuged at 850 g for 10 min to remove the myelin. Floating myelin and the supernatant were discarded and cell pellets were re-suspended in 15 mL centrifuge tube with 300 μL PBS contained 1% bovine serum albumin (BSA). Cells were incubated with CD11b (microglia) antibody (dilution at 1:100, Abcam, ab25533) at 4°C for 40 min. The incubations were manually shaken at 10 min intervals. After incubation, cells were washed 2 times with PBS contained 1% BSA to remove the residual antibody, and were resuspended in DMEM contained 10% FBS for single cell picking. Single cell suspensions were checked under fluorescence microscope to ensure the microglia cells were positively stained for fluorescence. Single cell complementary DNAs (cDNAs) were prepared instantly with the method of Smart-seq2 [35,36] with some modifications. Briefly, single cells with positive fluorescence and good state were picked up with the home-made mouth pipette aided by pulled glass capillaries and washed 3 times in PBS supplemented with 0.1% BSA. Cells were lysed in lysis buffer (2 μL 0.2% Triton X-100 (Sigma), 1 μL oligo(dT) primer and 1 μL dNTP (Thermo Fisher Scientific) and full-length poly(A)-tailed RNA was reversely transcribed and amplified with 23 cycles to increase the cDNA amount. The amplification products were diluted at 1:20 with RNase-free water for the quantification of targeted genes.

Western blotting and ELISA
Western blot assays for the target proteins were performed using the common procedure. Cell lysates of different mouse cerebral cortex, cultured mouse primary microglia and BV2 cells were prepared as described in our previous studies [32,33]. For the supernatants, proteins were extracted using methanol-chloroform precipitation as described previously [27]. A total of 25 μg protein was separated by 12% or 15% SDS PAGE, and was transferred to a polyvinylidene difluoride membrane (Bio-Rad, L1620177 Rev D). The membrane was soaked with 5% (w/v) skim milk for 2 h at room temperature. Antibodies included those in Table S1 and peroxidase-conjugated anti-mouse (lot number 474-1806) or anti-rabbit (lot number 474-1516) IgG (1:5000; KPL). The epitope was visualized using an ECL Western blot detection kit (Millipore, WBKLS0500). ImageJ (National Institutes of Health, Bethesda, Maryland, USA) was used to evaluate the densitometry. The IL-1β levels in culture supernatants, serum from mice and heroin addicts were analyzed by using ELISA kit (R&D Systems, USA).

Statistical analysis
Statistical analysis was performed by using GraphPad prism version 7.0. Sample size calculation was based on our previous study [21]. Power analysis showed that group sizes of 5-16 mice for most of behavioral tests analyzed were sufficient to detect a 20-40% improvement (ɑ = 0.05, 80% power, final sample size should be adjusted for expected attrition, see Table S2 and Table S3). The analgesic tolerance scores and acetic acid-induced writhing of WT and Nlrp3 −/− mice were assessed by using two-way repeated-measures analysis of variance (ANOVA). The data involving more than two groups were analyzed by using one-way ANOVA. Unpaired two-tailed t-test was used for comparing serum levels of IL-1β between healthy individuals and heroinaddicted patients. Data were represented as mean ± SEM or mean ± SD. A P value of < 0.05 was considered to be statistically significant.

Melatonin rescues analgesic tolerance in mice with chronic morphine treatment independent of autophagy
We examined the effect of melatonin on morphine-induced analgesic tolerance in mice by using the standard hot-plate [28] and tailflick tests [21]. Administration of morphine (10 mg/kg, subcutaneous (s.c.)) for seven consecutive days (Fig. 1A) established the antinociceptive tolerance ( Fig. 1B and C). Pretreatment with melatonin (0.5 mg/kg, intraperitoneal (i.p.)) 30 min earlier than the morphine injection in mice significantly prevented the development of morphineinduced antinociceptive tolerance ( Fig. 1B and C). These observations were consistent with our previous study [21] and others [37,38], and indicated that pretreatment with melatonin could prevent the development of antinociceptive tolerance induced by morphine. In mice with established analgesic tolerance (morphine injection for 7 days), co-administration of morphine and melatonin from the 8th day for 7 consecutive days (Fig. 1D) could significantly rescue the analgesic tolerance of morphine according to the tail-flick test with a laser density of IR30 (Fig. 1E) and IR50 (Fig. S1A). The salvaging effect of melatonin on morphine-induced analgesic tolerance could be replicated in mice by using the hot-plate tests at 47.5°C, 50°C or 52.5°C (each with a cutoff time of 30 s), albeit the level could not be restored to the peak level (Fig. 1F). Note that melatonin still had a beneficial effect on the analgesic tolerance in mice with administration of morphine for 13 consecutive days (Fig. S2). These results suggested that melatonin could block the analgesic tolerance in mice with established analgesic tolerance.
Melatonin could salvage autophagy induced by morphine that contributed to the reduced mtDNA copy number and drug addiction [21,32]. We tested whether the beneficial effect of melatonin on morphine-induced analgesic tolerance was also mediated by autophagy, which is characterized by an increased level of MAP1LC3B-II (LC3B-II, microtubule-associated protein 1 light chain 3 beta-II) and a decreased level of SQSTM1 (p62) [39]. Consistent with our previous study [21], chronic morphine treatment elevated LC3B-II:LC3B-I ratio and decreased SQSTM1 protein level in hippocampus tissues, suggesting an increased level of autophagy (Fig. S1B). However, co-administration with melatonin from day 8 had no salvaging effect on autophagy relative to those with morphine treatment alone (Fig. S1B). Evidently, the alleviating effect of melatonin on analgesic tolerance in mice with established analgesic tolerance was independent of autophagy.

The effects of melatonin on chronic morphine-induced NLRP3-caspase-1 inflammasome activation
Following the previous study showing the contribution of inflammasome activation to morphine analgesic tolerance [12] and the fact that many studies had reported anti-inflammatory properties of melatonin [20,[40][41][42][43], we investigated whether melatonin inhibits morphine analgesic tolerance via the blockade of inflammatory responses. We first analyzed the mRNA expression levels of NLRP3 inflammasome-related genes (Table S4) in prefrontal cortex (PFC) tissues from mice with chronic morphine treatment, together with or without melatonin pretreatment. Morphine treatment significantly induced the total mRNA levels of Nlrp3, Il-1β and Caspase-1, whereas melatonin pretreatment could abolish the induced expression of these genes by morphine ( Fig. 2A). Concordantly, the NLRP3 protein level was increased in the PFC tissues from mice with morphine treatment, and this upregulation was attenuated by melatonin pretreatment (Fig. 2B). These results indicated that morphine treatment induces NLRP3 inflammasome activity and a pretreatment with melatonin could abolish this activation.
We analyzed the protein levels of the adaptor ASC (as part of NLRP3 inflammasome), Caspase-1 and IL-1β, to confirm the activated inflammasome in PFC tissues (Fig. 2B). The Caspase-1 activation was assessed by the appearance of cleaved Caspase-1 p20 or p10 [10,44,45]. Compared with the control group, the protein levels of ASC, cleaved Caspase-1 and IL-1β were increased in PFC tissues from mice with morphine treatment, whereas melatonin pretreatment before morphine injection markedly alleviated the induced ASC, cleaved Caspase-1 and IL-1β (Fig. 2B). The serum level of IL-1β was significantly lower in mice with co-administration of melatonin and morphine compared to that of mice with morphine treatment (Fig. 2C). Collectively, these results suggested that the NLRP3-CASP1 inflammasome activation was induced in mice with chronic morphine treatment, and a pretreatment of melatonin before the morphine administration had a beneficial effect.
Microglia are resident macrophages and account for 10-15% of all cells in the brain [46]. Microglia can respond to homeostatic elements by initiating an inflammatory response [47]. We tested the response of microglia to morphine with or without melatonin. We used TMEM119 as a specific marker for microglia [48] and CD68 as a marker for activated microglia that can act as a proxy for phagocytic state [49]. ASC was used to show the NLRP3 inflammasome localization and activation [9]. Morphine markedly increased the ASC specks in microglia and activated microglia cells in mouse PFC. Pretreatment with melatonin could significantly block morphine-induced ASC specks in microglia cells (Fig. 2D-G). We performed single-cell quantitative PCR for NLRP3 inflammasome-related genes in microglia to further demonstrate the responses. We found that the NLRP3 inflammasome-related genes were highly expressed in single microglia from morphine-treated mice, while melatonin pretreatment significantly reversed the NLRP3 inflammasome activation induced by morphine (Fig. 2H). Collectively, these results demonstrated that microglia are actively involved in morphineinduced NLRP3 inflammasome activation.
We used cultured primary mouse microglia and a BV2 microglial cell line to further determine the effect of morphine in activating the NLRP3 inflammasome in microglia. The protein levels of NLRP3, the adaptor ASC, cleaved Caspase-1 and mature IL-1β in primary mouse microglia were remarkably induced in response to morphine and/or LPS treatment (Fig. 2I). The levels of IL-1β and cleaved Caspase-1 (p20) in culture supernatant of morphine-treated microglia were also significantly increased compared to the untreated cells (Fig. 2J). Melatonin pretreatment markedly decreased morphine-induced NLRP3 inflammasome activation in cultured primary microglia (Fig. 2I and J). It should be noticed that melatonin pretreatment also inhibited the LPSinduced NLRP3 inflammasome activation, suggesting its general effect on ameliorating the inflammation ( Fig. 2I and J). Consistently, melatonin pretreatment also blocked the morphine-induced NLRP3 inflammasome activation in mouse BV2 cells at the mRNA (Fig. S3A) and protein (Figs. S3B and C) levels.
The Caspase-1 is activated as a result of inflammasome assembly in the cytosol [50,51]. We examined the subcellular localization of Caspase-1 protein in BV2 cells by immunofluorescence. A diffused distribution of Caspase-1 staining was observed in the cytosol of untreated BV2 cells. Morphine treatment caused a dramatic morphological change of the cells and a perinuclear aggregation of Caspase-1, which could be abolished by a pretreatment with melatonin (Fig. S4A). Together, these results indicated that morphine activated the NLRP3 inflammasome in microglia, and melatonin pretreatment could block this activation. To define the role of NLRP3 inflammasome in response to morphine treatment, primary mouse microglia and BV2 cells were exposed to MCC950, an NLRP3 inhibitor [52], followed by morphine treatment for 6 h. Treatment of MCC950 (100 μΜ) inhibited the morphine-induced Caspase-1 processing, resulting in a reduced level of activated Caspase-1 (p20) compared to cells treated by morphine only (Fig. 2K). Similarly, the increased levels of NLRP3 and ASC by morphine were attenuated by MCC950 treatment (Fig. 2K). Primary mouse microglia cells were transfected with Nlrp3 siRNA (siNlrp3) or negative control siRNA (siNC), followed by the treatment with morphine. Knockdown of NLRP3 protein attenuated the morphine-induced Caspase-1 processing and led to a lower level of activated Caspase-1 (p20) (Fig. 2L). Similarly, NLRP3 knockdown abolished the morphine-induced upregulation of ASC and IL-1β (Fig. 2L). All these results for NLRP3 knockdown (Fig.  S4B) and pharmacological inhibition by MCC950 (Fig. S4C) could be replicated in BV2 cells. Evidently, NLRP3 was involved in the morphine-induced inflammasome activation and proinflammatory cytokine (IL-1β) secretion. Knockdown or inhibition of NLRP3 could block the morphine-induced inflammasome, with a similar effect to melatonin pretreatment.

Melatonin alleviates morphine-induced NLRP3 inflammasome activation through blocking cathepsin B (CTSB) release and oxidative stress
Previous studies have demonstrated that the lysosomal permeabilization, release of CTSB, and generation of mitochondrial reactive oxygen species (ROS) play an important role in activating the NLRP3 inflammasome [9,53,54]. As melatonin is a well-known antioxidant that acts by eliminating cellular ROS release [20,55], we tested whether melatonin alleviates morphine-induced NLRP3 inflammasome activation by mitigating release of CTSB and ROS. Melatonin blocked the elevated levels of ROS (Fig. 2M) and CTSB (Fig. 2N) proteins induced by morphine and LPS in primary mouse microglia, and this blocking effect of melatonin could be replicated in BV2 cells (Figs. S4 D and E). Concordantly, an increased CTSB release in response to morphine treatment was observed in mouse PFC tissues, whereas melatonin pretreatment inhibited the CTSB release induced by morphine (Fig. 2O).
To examine whether the release of CTSB from lysosomes was involved in morphine-induced NLRP3 inflammasome activation, we treated BV2 cells with 10 mM NH 4 Cl to inhibit CTSB activity by moving lysosomal pH above its optimal activity [53]. Treatment with NH 4 Cl caused a reduction of the CTSB protein level and attenuated the activation of NLRP3 inflammasome (Fig. S3D). These results indicated that CTSB was critical for morphine-induced activation of the NLRP3-CASP1 inflammasome.
Activation of NLRP3 inflammasome could induce pyroptosis, a specific type of cell death characterized by cell nuclear condensation, loss of plasma membrane integrity, and late release of intracellular proteins including the cytosolic enzyme lactate dehydrogenase (LDH) [56][57][58]. We found that melatonin pretreatment significantly reduced LDH release (Fig. S3E) and inhibited the condensed nuclei (Figs. S4F and G) induced by morphine. These results demonstrated that the suppression of CTSB and ROS release is critical for the alleviating effect of melatonin on morphine-induced NLRP3-CASP1 inflammasome activation.

The effect of low dose of morphine (1 mg/kg) combined with melatonin in a murine pain model
We used the acetic acid writhing model of pain to test whether melatonin combined with a low dose (1 mg/kg) of morphine could be used for relieving pain. A low dose of morphine combined with melatonin had a better analgesic effect than morphine only (Fig. 3A). Consistent with the previous study [59], we found that melatonin alone also had an analgesic effect (Fig. 3A). Analyses of the protein levels of NLRP3 inflammasome components in PFC tissues from this murine model further confirmed the beneficial effect of melatonin on counteracting the morphine-induced activation of inflammasome (Fig. 3B). Treatment of acetic acid dramatically induced IL-1β release into serum, and melatonin could completely block this effect (Fig. 3C).
The analgesia effect of melatonin on acetic-acid writhing via the NLRP3 inflammasome activation could be further demonstrated by direct inhibition of NLRP3. Compared to the acetic acid group, pretreatment of MCC950 significantly reduced the pain in mice injected with acetic acid, and the analgesic effect of MCC950 (Fig. 3D) was even comparable to that of low dose (1 mg/kg) of morphine combined with melatonin (Fig. 3A). The induced protein levels of NLRP3 and ASC (Fig. 3E), and increased IL-1β release (Fig. 3C) in the acetic acid group were attenuated by MCC950 treatment. These results supported that NLRP3 inflammasome activation plays an important role in pain [12,60], and low dose of morphine combined with melatonin, or direct delivery of NLRP3 inhibitor MCC950, might have better anti-analgesia effects for clinical usage.
morphine administration had no effect on morphine-induced analgesic tolerance in Nlrp3 −/− mice ( Fig. 4A and B), suggesting that the expected beneficial effect of melatonin was mediated by NLRP3. There was no NLRP3 inflammasome activation in PFC tissues of Nlrp3 −/− mice, regardless of morphine treatment (Fig. 4C). We failed to detect the IL-1β release in serum of Nlrp3 −/− mice with or without morphine treatment (Fig. 4D). These findings further showed the active involvement of NLRP3 inflammasome activation in the development of morphine-induced analgesic tolerance.
We assessed the pain in Nlrp3 −/− mice injected with or without acetic acid. The WT and Nlrp3 −/− mice with saline injection had no apparent writhing movements (Fig. 4E). Injection of acetic acid in WT mice resulted in a significant increase of writhing movements relative to the saline controls (Figs. 3A and 4E). Nlrp3 deficiency significantly reduced the number of writhing movements compared to WT mice injected with acetic acid (Fig. 4E), although Nlrp3 −/− mice also (caption on next page) Q. Liu, et al.
Redox Biology 34 (2020) 101560 presented writhing movements in response to acetic acid. The induced levels of NLRP3 and Caspase-1 (p20) (Fig. 4F) and IL-1β release (Fig. 4D) by acetic acid were attenuated in Nlrp3 −/− mice. These results based on Nlrp3 −/− mice showed that the NLRP3 inflammasome was actively involved in morphine-induced analgesic and affected pain response.

Opiate-addicted patients have an increased level of serum IL-1β
Opiate-addicted patients often have opioid-induced hyperalgesia [61,62]. We previously showed that heroin-addicted patients had a significantly lower level of serum melatonin compared to healthy individuals [21], and this might be associated with the pain sensitivity in these patients based on the above results. We found that the level of serum IL-1β in heroin-addicted patients was significantly higher than that of healthy individuals (Fig. 4G). The decreased melatonin in opiate-addicted patients would be associated with an increase of NLRP3 inflammasome, which in turn led to pain sensitivity. It could therefore be useful to use melatonin as an adjunct to relieve the pain of opiateaddicted patients, and a clinical trial could be performed to confirm this practice.

Discussion
Morphine has been widely used to relieve pain, but it causes analgesic tolerance and hyperalgesia [3,63]. It is important to find a way to avoid the side effects of morphine and related opioids. In this study, we have provided several lines of evidence to show that melatonin could alleviate the side effects of morphine by targeting the NLRP3 signaling in microglia and relieved morphine-induced analgesic tolerance. We have shown the potential mechanism by which chronic morphine exposure leads to excess production of cellular ROS and activation of NLRP3 inflammasome in microglia. Melatonin could diminish ROS and consequently inhibited the activation of NLRP3 inflammasome, suppressed the over-activated IL-1β signaling, which finally attenuated the development of morphine analgesic tolerance (Fig. 5). This effect was independent of autophagy, which has been shown to be involved in drug addiction [21,64,65] and presented an Atg5-and Atg7-dependent, dopaminergic neuron specific pattern [32]. These findings indicated that melatonin enacted different roles in counteracting morphine-induced antinociceptive tolerance compared to its salvaging effect on autophagy to prevent morphine-induced behavioral sensitization [21]. The diverse effects of melatonin might be caused by its multiple functions in morphine addiction and analgesic tolerance. There was no cross-tolerance between melatonin and morphine, as we found that repeated injections of melatonin twice daily for 2 days did not affect morphine potency on day 3 compared with vehicle-pretreated controls (Fig. S5A). Similarly, pretreatment with morphine had no effect on melatonin potency (Fig. S5B). Our results offered new insights into the cellular events underlying chronic morphine analgesic tolerance and indicated melatonin as a promising drug for the optimization of the analgesic actions of morphine. Since melatonin pretreatment also blocked NLRP3 inflammasome activation following LPS treatment ( Fig. 2I and J), it would be rewarding to test whether melatonin functioned in a more general mechanism than just blocking morphine induction, and whether it had a similar effect on alleviating the detrimental side effects of other opioid analgesics.
The observation that morphine-induced antinociceptive tolerance could be fully abolished by melatonin pretreatment or MCC950 has provided the basis for initiating a clinical trial for the usage of melatonin together with a low dose of morphine, or simply the use of MCC950 to relieve pain. The increased serum IL-1β level in heroinaddicted patients compared to healthy individuals was in agreement with the expected result of decreased melatonin in these patients [21] considering the alleviating effect of melatonin. This suggested that alleviation of NLRP3 inflammasome could be a target for counteracting morphine analgesic tolerance. Whether other pathways might be also involved in this alleviating process by melatonin, such as the canonical pathway of ion channels [74][75][76], the peripheral GABAergic system [77], the protein kinase C and N-methyl-D-aspartate receptors activity [37], microglia activation and HSP27 expression [38], awaits further study. Note that different dosages of melatonin were required to achieve the same antinociception effect in the tail-flick and hot-plate tests (Fig. 1) and in the acetic acid writhing test (Fig. 3), which were used to test acute thermal pain [78] and acute visceral pain [79], respectively. The underlying molecular mechanisms/circuits of these two types of pain are different, and we believe this is the reason why different dosages of melatonin were needed. Therefore, optimization of melatonin concentrations should be considered to counteract morphine-induced analgesic tolerance for different types of pain.
In summary, we found that chronic morphine treatment caused a release of ROS and CTSB, which led to an excessive activation of NLRP3 inflammasome activity and pyroptosis in microglia. The detrimental effect of morphine could be counteracted by melatonin treatment or NLRP3 inhibition, which prevented the development of morphine analgesic tolerance (Fig. 5). Importantly, a low-dose morphine combined with melatonin had better analgesia effects in murine pain models. Our results provided a possible new solution to improve the analgesic efficacy of morphine by the inhibition of microglial NLRP3 inflammasome. As melatonin was found to be safe and effective for inhibiting morphine tolerance, it may be a drug candidate to reduce morphine analgesic tolerance and to enhance the clinical usefulness of morphine and its related opioids.

Disclosure of potential conflicts of interest
There were no potential conflicts of interest to be disclosed.