The nociceptin/orphanin FQ receptor system as a target to alleviate cancer‐induced bone pain in rats: Model validation and pharmacological evaluation

Background and Purpose Cancer‐induced bone pain remains inadequately controlled, and current standard of care analgesics is accompanied by several side effects. Nociceptin/orphanin FQ peptide (NOP) receptor agonists have demonstrated broad analgesic properties in rodent neuropathic and inflammatory pain models. Here, we investigate the analgesic potential of NOP receptor activation in a rodent cancer‐induced bone pain model. Experimental Approach Model validation by intratibial inoculation in male Sprague Dawley rats was performed with varying MRMT‐1/Luc2 cell quantities (0.5–1.5 × 106·ml−1) and a behavioural battery (>14 days post‐surgery) including evoked and non‐evoked readouts: paw pressure test, cold plate, von Frey, open field, and weight distribution. Anti‐allodynic potential of the endogenous NOP receptor ligand nociceptin (i.t.) and NOP receptor agonist Ro65‐6570 ( i.p.) was tested using von Frey filaments, followed by a combination experiment with Ro65‐6570 and the NOP receptor antagonist J‐113397 (i.p.). Plasma cytokine levels and NOP receptor gene expression in dorsal root ganglion (DRG, L4‐L6) and bone marrow were examined. Key Results Inoculation with 1.5 × 106·ml−1 of MRMT‐1/Luc2 cells resulted in a robust and progressive pain‐related phenotype. Nociceptin and Ro65‐6570 treatment inhibited cancer‐induced mechanical allodynia. J‐113397 selectively antagonized the effect of Ro65‐6570. MRMT‐1/Luc2‐bearing animals demonstrated elevated plasma cytokine levels of IL‐4, IL‐5, IL‐6 and IL‐10 plus unaltered NOP‐r gene expression in DRG and reduced expression in bone marrow. Conclusion and Implications Nociceptin and Ro65‐6570 selectively and dose‐dependently reversed cancer‐induced bone pain‐like behaviour. The NOP receptor system may be a potential target for cancer‐induced bone pain treatment. LINKED ARTICLES This article is part of a themed issue on The molecular pharmacology of bone and cancer‐elated bone diseases. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v178.9/issuetoc

LINKED ARTICLES: This article is part of a themed issue on The molecular pharmacology of bone and cancer-elated bone diseases. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v178.9/issuetoc

| INTRODUCTION
The most common type of pain in cancer patients is bone cancer pain (O'Toole & Boland, 2006), which has a nociceptive and neuropathic component (Colvin & Fallon, 2008;O'Toole & Boland, 2006). In 1986, the World Health Organization established a three-step guideline for adequate malignant pain treatment, starting with non-opioids (Step 1) for mild pain, weak opioids ± non-opioids and adjuvants for mild to moderate pain (Step 2), and strong opioids ± non-opioids and adjuvants for moderate to severe pain (Step 3;WHO, 2018;Zhu et al., 2015). In particular, for bone pain, it is recommended to combine treatment for pain with radiotherapy and bisphosphonates (WHO, 2018). However, in 43.4% of patients, the cancer was undertreated according to the pain management index (Greco et al., 2014) and 20% of patients rotate through ≥3 opioids before an efficient balance between analgesia and side effects was reached (Sloan, 2008). Nonetheless, implementation of the guidelines has been shown to reduce the most frequently occurring symptoms, for example, impaired activity, mood changes, constipation, nausea and dry mouth (Meuser et al., 2001). Nevertheless, opioid use unwanted effects are still a risk accompanying long-term treatment of chronic pain patients with opioids (Kaye et al., 2017). This has contributed to the current opioid crisis in the United States and constitutes a major health concern as increasing numbers of opioid-related deaths are reported (Koenig, 2018). The opioid crisis in combination with the unmet medical need for a more efficacious and better tolerable treatment specifically for malignant pain is an important current driver for the development of novel analgesic compounds.
The aim of the present study was to investigate the putative analgesic effect of NOP receptor activation in the MRMT-1 rat model of cancer-induced bone pain. Intratibial inoculation with rat mammary gland carcinoma (MRMT-1) cells was previously described as a robust translational model for metastatic cancer-induced bone pain in rats (Medhurst et al., 2002). To validate the robustness of our model, the pain-like phenotype induced by different MRMT-1 cell concentrations was examined. Additionally, the inflammatory component of the model was assessed by plasma cytokine levels and NOP receptor gene expression in dorsal root ganglions (DRGs) and, for the first time, in bone marrow. Thereafter a battery of behavioural tests was conducted in order to determine which readout is best suited to assess pain-related behaviour in this model. What is already known • Current treatment for bone cancer pain is inadequate and represents an unmet medical need.

What does this study adds
• NOP receptor agonists selectively and effectively attenuate mechanical allodynia in a bone cancer pain model.

What is the clinical significance
• The NOP receptor system represents a promising novel target for treatment of bone cancer pain was provided as environmental enrichment. Rats were left to acclimatize to the facility for 1 week prior to initiation of experiments and animal welfare, for example, body weight, grooming, posture, and gait, was assessed on a daily basis. All experiments were performed according to the German Animal Welfare Act and were approved by the local government authority . Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny, Browne, Cuthill, Emerson, & Altman, 2010) and with the recommendations made by the British Journal of Pharmacology.

| Study design
All studies were conducted between January 2017 and September 2018. Behavioural testing occurred between 07:00 am and 01:00 pm and X-ray and bioluminescence imaging was performed afterwards. In case of multiple measurements per day, the order was limb use first, then experimental test and, finally, imaging. All experiments are reported in accordance with the ARRIVE guidelines (Kilkenny et al., 2010). Group sizes were designed to be equal but did eventually vary due to the applied exclusion criteria (Figure 1).
The first study (Study 1, total n = 60) was aimed to determine which quantity of intratibially inoculated MRMT-1/Luc2 cells resulted in a robust cancer-induced bone pain-like phenotype. Therefore, three concentrations of MRMT-1/Luc2 cells (i.e., 0.5 × 10 6 Áml −1 , 1.0 × 10 6 Áml −1 , and 1.25 × 10 6 Áml −1 ) were inoculated and mechanical allodynia (von Frey hair) was tested (16-17 days post-surgery). Additionally, X-ray densitometry and bioluminescence signal images were F I G U R E 1 A brief schematic representation of the experimental design. Generally, prior to surgery (Day 0) training and baseline measurements were performed. Thereafter, animals were allowed to recover for 1 week (red cross) in which no measurements occurred, except for the limb use test, which was performed on a daily basis until day 20 post-surgery (not shown in figure). X-ray densitometry images ("X") and monitoring of bioluminescence signal ("B") to examine bone degradation and tumour growth progression, respectively, were obtained on Days 7, 10, 13, 16, and 19 post-surgery. Animals were subjected to behavioural testing after the second week on Days 14 to 20 post-surgery, exact days depending on the tests, for example, evoked or non-evoked readouts. Blood was obtained or tissue was harvested at the end of the experiment, 21 days post-surgery. The table shows which animal treatment group was included (X) per study. Sham animals were included in each study for behavioural testing, except Study 3C; vehicle-treated animals were always included. Results of the model validation studies "Validation": Study 1 is shown in appendices; Study 2A is shown in Figure 2; Study 2B is shown in Table 2; Results of the pharmacological interventions with NP agonists "Pharmacology": Study 3A-C is shown in Figure 3 T A B L E 1 Levels of inflammatory markers measured in plasma of sham and 1.5 × 10 6 Áml −1 of MRMT-1/Luc2-bearing animals obtained 7, 10, 13, 16 and 19 days post-surgery and post-mortem tumour presence was analysed (see Figure S1 and Data S1). The first study also includes characterization of the model by analysis of inflammatory mediators (Table 1) and ipsilateral DRGs L4-L6 and ipsilateral bone marrow were harvested 21 days post-surgery for NOP receptor (oprl1) gene expression analysis (  Table S1 and corresponding methods). All behavioural tests included a baseline measurement.
F I G U R E 2 Effect of 1.25 × 10 6 Áml −1 of MRMT-1/Luc2 inoculated cells on (a) weight-bearing ratio, (b) cold allodynia, and (c) mechanical allodynia, including the effect of i.p. morphine (3.16 mgÁkg −1 ) administration. All data are presented as mean ± SEM; n = 9, all MRMT-1/ Luc2-bearing animals; *P < .05 versus pretest 15 min prior to the agonist. Here, no morphine group was included, and sham animals were not subjected to behavioural testing. Blood samples were obtained to measure exposure levels of Ro65-6570 and J-113397 2 hr post-administration. Additionally, bioluminescence signal images were obtained on Days 7, 13 and 19 post-surgery, and Xray densitometry was performed at Day 19 post-surgery (see Figure S2 and Data S2).
All behavioural tests and analyses were performed by an operator blinded to the experimental groups. Animals were randomly assigned to treatment groups by an operator unaware of the study hypothesis.
Vehicle, positive control (morphine) and treatment doses were prepared in non-transparent holders by an operator not involved in behavioural testing and administration of the treatment occurred in a randomized fashion. Finally, animals were killed using CO 2 exposure at the end of the experiments, when the humane endpoint was reached, or killed by decapitation under isoflurane for blood sample collection.
Exclusion criteria for MRMT-1/Luc2-bearing animals were as follows: (a) if von Frey response at pretest was >9 g (n = 37) and (b) if no bioluminescence signal was present at day of behavioural testing (n = 23). Exclusion of animals occurred at the end of the experiments after behavioural and bioluminescence data analysis.

| Manual von Frey-Mechanical allodynia
Mechanical allodynia was measured using von Frey monofilaments (Ugo Basile, range 1-15 g), applied to the plantar surface of the ipsilateral hind paw. The paw withdrawal threshold was determined by means of the up-and-down principle, as described previously (Chaplan, Bach, Pogrel, Chung, & Yaksh, 1994). Briefly, animals were placed in a plastic box (L15 × W10.7 × H13.8 cm) located on a wire mesh platform, 30 min prior to first stimulation. Starting at 2 g, stimulation occurred in ascending intensity until a first paw withdrawal response. Consequently, stimulation was performed in descending intensity until no response was observed. Final positive versus final negative response defines the 50% g threshold followed by five consecutive stimulations followed in ascending or descending order, depending on a response. A cut-off value of 15 g was used for cases where no withdrawal response was observed until that value. The 50% response withdrawal threshold was calculated using the formula: 50% g threshold = (10Xf + kδ)/(10.000), where Xf = value (log unit) of the final von Frey monofilament used; k = tabular value for the pattern of up and down responses; and δ = mean difference (in log units) between filaments. MRMT-1/Luc2 animals with a value of >9 g and sham animals with a value <9 g were excluded from the analysis. The von Frey test was used to assess early alterations in cancer-induced bone pain-related behaviour as described previously by Falk, Schwab, et al. (2015) and excludes animals with an advanced cancer stage.

| Weight bearing-Weight distribution
Weight distribution was determined using a rat incapacitance tester (Somedic Sales AB) as previously described by Schiene et al. with modifications as described by Falk et al. (Falk, Al-Dihaissy, et al., 2015;Schiene, De Vry, & Tzschentke, 2011;Schott et al., 1994). Rats were placed in a Plexiglas chamber of the incapacitance tester with their front paws on an angled plate. Their hind paws were located on separate sensors to measure the weight distribution over 4 s in five trials. An average weightbearing ratio was calculated as follows: amount of weight of MRMT-1/ Luc2-bearing limb/the total amount of weight placed on both limbs.

| Cold plate-Cold allodynia
The test for cold allodynia was performed as previously described by Tzschentke, Linz, Frosch, and Christoph (2017)

| X-ray densitometry-Bone density
Relative bone density was measured by X-ray densitometry. Animals were anaesthetized with isoflurane (4% for induction; 2.5% for main-

| Tissue collection
Animals were killed at the end of the experiment via CO 2 exposure and tissue, that is, ipsilateral DRGs (L4, L5, and L6) and ipsilateral bone marrow were harvested from sham and MRMT-1/Luc2 animals. Tibias were dissected, and the bone was cut on the proximal end, close to the knee. Tibias were then centrifuged at 12.000 RPM for 3 s to let bone marrow flow out. Bone marrow and DRGs were collected in 0.5-ml Eppendorf tubes. Tissue was snap-frozen in liquid nitrogen immediately after harvesting and stored in −80 C for further use.

| NOP receptor gene expression (RT-PCR) analysis
Harvested rat DRGs and bone marrow were homogenized using a

| Drugs
The following drugs were used: morphine hydrochloride trihydrate  (Rutten et al., 2018). Nociceptin doses were based on previous publications using inflammatory and neuropathic pain models (Katsuyama et al., 2011;Ma et al., 2003), and morphine dose was based on historical in-house data.

| Study 1
Mechanical hypersensitivity was analysed by one-way ANOVA followed by Bonferroni method for correction of multiple comparisons. X-ray images were analysed by Student's unpaired t-tests and regression analysis was used to analyse bioluminescence signal images (data from Figure S1). Cytokine levels were analysed using Student's unpaired t-test (data from Table 1). Gene expression analysis was performed using fold changes after normalization (data from Table 2).

| Study 2
Mechanical allodynia between ipsilateral and contralateral hind paws of sham and MRMT-1/luc2-bearing animals was analysed by Student's paired t-test. The effect of morphine treatment in MRMT-1/luc2-bearing animals in weight bearing, cold plate, and von Frey was analysed using one-way ANOVA followed by Dunnett's multiple comparisons test (data from Figure 2). Paw pressure test and open field data were analysed with two-way ANOVA followed by Bonferroni multiple comparisons test.

| Study 3
The effect of compounds (morphine, nociceptin, Ro65-6570 and J-113397) were analysed by one-way ANOVA followed by Dunnett's multiple comparisons for multiple comparisons (data from Figure 3).
Corresponding limb use data were analysed with Friedmann's twoway test followed by Wilcoxon's two sample test for post hoc analysis of individual time points. Regression analysis was used to analyse bioluminescence signals, and X-ray images were analysed by Student's unpaired t-tests (data from Figure S2).

| Nomenclature statement of targets and ligands
Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding, Sharman et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2019/20 (Alexander et al., 2019).

| Study 1-Cell quantity validation, inflammatory mediators, and gene expression
The different inoculated cell quantities each resulted in significantly reduced paw withdrawal thresholds; however, the most robust and consistent phenotype with stable bioluminescent signals, significant bone degradation, and minimal extinction of tumours (as assessed post-mortem) was achieved using 1.25 × 10 6 Áml −1 and 1.5 × 10 6 Áml −1 of MRMT-1-/Luc2 cells (see Figure S1 and Data S1). The analysis of inflammatory cytokines in plasma, 21 days post-surgery, showed no difference between sham-and vehicle-treated MRMT-1/ Luc2-bearing animals in the levels of IFN-γ, IL-13, IL-1β, KC-GRO and TNF-α. A significant increase was found in the MRMT-1/ Luc2-bearing animals for IL-10, IL-4, IL-5 and IL-6, compared to sham (see Table 1) NOP receptor gene expression was examined in DRGs and bone marrow of MRMT-1/Luc2-bearing and sham animals (

| Study 2-Behavioural battery validation
Sham animals did not show a difference in von Frey withdrawal thresholds between ipsilateral and contralateral side, t(8) = 0.36.
MRMT-1/Luc2-bearing animals had a significantly decreased ipsilateral paw withdrawal threshold compared to the contralateral side, t  (Table S1).
In The difference in limb use score, bioluminescent signals and relative bone density were only assessed between the sham and MRMT-1/Luc2-bearing animals from Study 3A-C to monitor tumour presence, analysis of bone degradation and as human endpoint respectively. The data provide an illustration for the consistency and robustness of our model and is shown in Figure S2A-C, S2D-F, and S2G-I.

| DISCUSSION
Every year, close to 10 million people are diagnosed with cancer and this is accompanied by pain in 30% to 50% of patients with advanced cancers (Wiffen, Wee, Derry, Bell, & Moore, 2017). However, it is suggested that half of these patients receive inadequate therapy for pain control (Schug & Chandrasena, 2015). Several animal models of cancer-induced bone pain have been developed to further understand its mechanisms. An example is the intratibial inoculation of MRMT-1 cells in Sprague Dawley rats (Medhurst et al., 2002). Accordingly, our study shows that inoculation of 1.25 × 10 6 Áml −1 and 1.5 × 10 6 Áml −1 of MRMT-1/Luc2 cells results in a reliable and robust cancer-induced bone pain-related phenotype with a progressive development of the disease. Similar to the literature, the bioluminescent signal plateaus after the initial 10 days and it was suggested that lack of oxygen renders the core of the tumour necrotic and limits the chemical reaction of luciferase to produce the bioluminescent signal (Appel et al., 2017;Diaz-delCastillo et al., 2018). Although the required cell quantity in our lab was higher in comparison to previous studies (Falk, Al-Dihaissy, et al., 2015), the range of inoculated cell number and the behavioural outcome in pain phenotype is robust and comparable with other literature (Medhurst et al., 2002;Schwei et al., 1999). The up-regulation of IL-4, IL-5 and IL-10 as seen in our MRMT-1/ Luc2-bearing animals is in agreement with previous data and is indicative for cancer physiology modulation (Goldstein et al., 2011;Rosen et al., 1998). In addition, IL-6 is suggested to play a substantial role in the development of pathological pain conditions, for example, cancer pain (Zhou et al., 2016)  receptors (Fang et al., 2015;Remeniuk et al., 2018). Next, little is known about the presence of the NOP receptor in the musculoskeletal tissue. Data from synovial tissue are contradictory (Kumar et al., 1999;Zhang & McDougall, 2006) but suggest that NOP receptor is present in tendon and joint tissue (Ackermann et al., 2001;McDougall, 2003). This study is the first to describe NOP receptor gene expression in rat bone marrow. Interestingly, NOP receptor expression was significantly decreased in bone marrow of MRMT-1/ Luc2-bearing animals, possibly due to invading tumour cells diminishing NOP receptor expressing cells. Further studies investigating the presence of the NOP receptor in bone, cartilage or synovium are warranted, as this yields valuable information for peripheral pain management, for example, intra-articular treatment for jointassociated pain and bone cancer (Bergstrom et al., 2006).
Our behavioural battery identified ipsilateral mechanical allodynia (measured by the von Frey test) as leading phenotype which could be fully reversed by morphine, while cold plate and weight-bearing readouts showed a more moderate phenotype. Cold allodynia has been described in different rodent models of chronic pain (Linz et al., 2014) but appears to be much less pronounced than mechanical allodynia in our cancer-induced bone pain model. Weight bearing has been previously described as a possible readout in cancer-induced bone pain models (Medhurst et al., 2002). The test window for both cold allodynia and weight bearing in our MRMT-1/Luc2-bearing animals was too small to allow detailed pharmacological investigations, although the cancer-induced change in both readouts was partially reversed by morphine. Furthermore, we did not observe a cancer-  Tian et al., 1997;Wang, Zhu, Cao, & Wu, 1999) and anti-hypersensitive actions in rat models of neuropathic and inflammatory pain (Courteix et al., 2004;Fu et al., 2006;Hao, Xu, Wiesenfeld-Hallin, & Xu, 1998;Ju, Shin, Na, & Yoon, 2013;Katsuyama et al., 2011;Ma et al., 2003). Spinal NOP receptor activation has been shown to be efficacious in both neuropathic and inflammatory pain models (Schroder et al., 2014). Overall, NOP receptor activation is suggested to have more potent effects in chronic pain as compared to acute pain conditions in rodents (Schroder et al., 2014).
Accordingly, the potency of nociceptin in our cancer-induced bone pain model is in the same range as the potency observed in other chronic pain models (Courteix et al., 2004;Fu et al., 2006;Hao et al., 1998;Ju et al., 2013;Lee, 2013;Ma et al., 2003).
The NOP receptor agonist Ro65-6570 significantly reversed mechanical allodynia at a dose of 1.0 mgÁkg −1 but showed sedative side effects at a higher dose. This is consistent with what has been described in a rat mononeuropathic spinal nerve ligation model (Rutten et al., 2018). The analgesic effect of 1.0 mgÁkg −1 of Ro65-6570 was blocked by 4.64 mgÁkg −1 of J-113397, a NOP receptor antagonist. As we have previously shown that these doses of both compounds are selective for the NOP receptor (Rutten et al., 2018), this finding implies that the analgesic effect of Ro65-6570 was NOP receptor mediated. Several studies support analgesic efficacy of NOP receptor activation in preclinical models of pain, for example, chronic constriction injury (Wu & Liu, 2018), spinal nerve ligation (Rutten et al., 2018) and complete Freund's adjuvant (Chen & Sommer, 2007).
Moreover, a possible involvement of the NOP receptor in cancerinduced bone pain has been suggested by the efficacy of cebranopadol, a mixed opioid/NOP receptor agonist (Linz et al., 2014), and buprenorphine, a mixed μ-opioid peptide/NOP partial agonist and κ-opioid peptide/δ-opioid peptide antagonist (Gastmeier & Freye, 2009) in rats. However, the present study is the first to reveal highly efficacious and potent relief of cancer-induced bone pain by selectively targeting the NOP receptor in rats.
It has previously been described that spinal and peripheral activation of the NOP receptor contributes to analgesic effects, whereas supraspinal activation may result in hyperalgesia (Schroder et al., 2014). The exact mechanism behind this dichotomy remains unclear, but it is speculated that i.t. NOP receptor activation attenuates mechanical allodynia via hyperpolarization of (glycinergic) interneurons which are disinhibited upon injury, thereby enabling touch sensation to engage with nociceptive projection neurons (Ozawa et al., 2018).
On the other hand, the hyperalgesic effects of supraspinal NOP receptors are suggested to be a result of the inhibition of OFF cells within the rostral ventromedial medulla (Schroder et al., 2014). Descending pathways from the rostral ventromedial medulla to the spinal cord dorsal horn inhibit ascending nociceptive signals (Lau & Vaughan, 2014), and supraspinal NOP receptor activation may thus result in disinhibition at the spinal level (Schroder et al., 2014). Determination of the binding affinity of nociceptin at the NOP receptor has been notoriously difficult and reported K i values vary widely (0.03-2 nM) due to differences in binding assays and unusual binding properties of nociceptin (see Dooley & Houghten, 2000). Furthermore, differences may exist in the binding affinity of nociceptin in the spinal cord and brain (Kusaka, Yamada & Kimura, 2001). Here, full efficacy was