Morphine (DBL^TM morphine sulfate injection BP – 30 mg in 1 mL) was procured from Hospira Pty Ltd. (Melbourne, VIC, Australia). Gabapentin was kindly provided by Dr. Ben Ross, School of Pharmacy, The University of Queensland (Brisbane, QLD, Australia). Amitriptyline (amitriptyline hydrochloride), meloxicam (meloxicam sodium salt hydrate), naloxone (naloxone hydrochloride dihydrate), Triton^TM X-100, Tween 20, paraformaldehyde (PFA) and bovine serum albumin (BSA) were procured from Sigma–Aldrich^® (Castle Hill, NSW, Australia). Isoflurane (IsoFlo^TM) was procured from Abbott Australasia Pty. Ltd. (Macquarie Park, NSW, Australia). Medical oxygen and medical carbon dioxide were procured from Coregas Pty Ltd. (Yennora, NSW, Australia). Triple antibiotic powder (Tricin^®) was procured from Jurox Pty Ltd. (Rutherford, NSW, Australia). Benzylpenicillin (BenPen^TM, benzylpenicillin sodium for injection) was procured from CSL Ltd. (Melbourne, VIC, Australia). Pentobarbitone (Lethabarb^®, pentobarbitone sodium) was procured from Virbac (Australia) Pty Ltd. (Penrith, NSW, Australia). Eye ointment (Refresh Night Time^®) was purchased from Allergan Australia Pty Ltd. (Gordon, NSW, Australia). Ethylenediaminetetraacetic acid, disodium salt, dihydrate (UltraPure^TM EDTA), 4′,6-diamidino-2-phenylindole, dihydrochloride (DAPI), Prolong^® Gold antifade reagent, phosphate-buffered saline (PBS), medium 199 (1×), horse serum, Dulbecco’s phosphate-buffered saline (DPBS, 1×) and 0.25% trypsin-EDTA (1×) were purchased from Thermo Fisher Scientific Australia Pty Ltd. (Scoresby, VIC, Australia). Normal goat serum (NGS) was purchased from Cell Signaling Technology^® (Danvers, MA, United States). Tissue-Tek^® O.C.T. Compound was procured from ProSciTech Pty Ltd. (Kirwan, QLD, Australia). Water for injection BP was purchased from Pfizer Australia Pty Ltd. (West Ryde, NSW, Australia). Ten percent neutral-buffered formalin (NBF) was purchased from Australian Chemical Reagents (Moorooka, QLD, Australia). Methanol and acetone were purchased from EMD Millipore Corporation (Billerica, MA, United States).

W256 breast cancer cells [LLC-WRC 256 (ATCC^® CCL-38^TM)] at passage number 290 were procured from the American Type Culture Collection (ATCC; Manassas, VA, United States). The cells were cultured and passaged following the ATCC guidelines. Cells within passage number 292–319 were used in the present study. To summarize, the cells were thawed from the frozen stocks and cultured in 75 cm² Cellstar^® flasks (Greiner bio-one) at 37°C (5% CO₂: 95% atmospheric air) in 20 mL of Medium 199 (1×) supplemented with 5% horse serum. For detachment of cells, they were initially rinsed gently with 3 mL of DPBS (1×), followed by trypsinization using 2 mL of 0.25% trypsin-EDTA (1×). The cells thus detached were collected by centrifugation with 8 mL of medium for 4 min at 200 × g. The resultant supernatant was discarded, the pellet was re-suspended in 3 mL of DPBS and cell counting was performed using a hemocytometer. After re-centrifuging the pellet for 4 min at 200 × g, the cells obtained were suspended in DPBS in the required concentration (4 × 10³, 1.5 × 10⁴, 4 × 10⁴, 1.5 × 10⁵ and 4 × 10⁵ cells/10 μL DPBS). HK W256 cells were prepared in the same way as the live cells but with an additional step involving heating for 15 min at 90°C.

Female Wistar Han (HsdBrlHan) rats used in these experiments were procured from the Herston Medical Research Centre (Brisbane, QLD, Australia) of The University of Queensland. On arrival at our facility, rats were approximately 3–4 weeks of age with a body weight in the range of ~50–70 g. Rats were caged in groups of two to three in a room with controlled temperature (23 ± 2°C) and a 12 h/12 h light–dark cycle. Standard rodent chow (Specialty Feeds, Glen Forrest, WA, Australia) and tap water were available to the rats ad libitum. Kimwipes (Kimberly-Clark Professional, Kirribilli, NSW, Australia) and Rat Chewsticks (Able Scientific, Canning Vale, WA, Australia) were provided as environmental enrichment. Rats were subject to acclimatization for at least 3 days prior to initiating experimentation. All the experiments herein were performed in the light phase. Approval of experimental procedures was obtained from the Animal Ethics Committee of The University of Queensland (Brisbane, QLD, Australia). The experiments were undertaken in accordance with the requirements of the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (8th edition, 2013).

Unilateral ITIs were performed in a manner similar to that described by others (Mao-Ying et al., 2006; Muralidharan et al., 2013) but with some modifications. Briefly, rats (~80–120 g) were anesthetized deeply with 3% isoflurane delivered in oxygen. Eye ointment was used to avoid drying of eyes during the surgical procedure. Benzylpenicillin injection was subcutaneously administered at a dose of 60 mg per rat. A unilateral rostro-caudal incision of approximately 1 cm in length was created on the upper medial half of the lower left hind limb. After the exposure of the tibia, using a 23-gauge needle, the bone was pierced medial to the tibial tuberosity below the knee joint. A 10 μL injection containing W256 cells in the required number, or HK cells (sham-rats) or DPBS (control rats) was administered into the bone cavity with a Hamilton^® syringe [80508:705SN 50 μL SYR SPECIAL (22/2″/4), Reno, NV, United States]. In all of the experiments (except experiment 4), at least one rat received an ITI of HK cells as a control for lack of pain hypersensitivity, including the experiments designed for pharmacological testing (Supplementary Table 1). The bone was immediately sealed using Ethicon^TM W810 bone wax (Johnson-Johnson International, Diegem, Belgium). The muscles and the skin were stitched in place with non-absorbable USP 5/0 sutures Dysilk^® suture (Dynek Pty Ltd, Hendon, SA, Australia). Topical antibiotic powder was then dusted on the wound. The hindpaw of the injected limb is termed as the ‘ipsilateral’ hindpaw, while that of the non-injected limb is termed as the ‘contralateral’ hindpaw. After the surgery, rats were closely monitored. The general animal health and their body weights were regularly assessed throughout the experimental period.

The number of animals involved in individual groups along with the details of each of the experiments performed together with their corresponding durations are shown in Figure 1. For the ease of editing and typesetting the figure, clinical observations have not been denoted for experiments 2–16 in Figure 1, although they were performed. Baseline pain behavioral tests prior to surgeries/ITI (referred as day 0) were performed on rats from all the experiments (Supplementary Table 2). To comprehensively characterize the model, one group of rats received an ITI of DPBS and another group (naïve) of animals did not undergo surgery or ITI (experiment 4).

The primary aim of Experiment 1 was to perform clinical observations and document temporal changes in general health along with body weights for the study duration. Animal welfare checks for mortality, morbidity and food and water levels were performed twice daily until the study completion. Body weights and clinical observations (general health parameters) were generally monitored at least once per week until the study completion in all the experiments. The observation method and the clinical parameters/symptoms were obtained and modified from the previously established standard methods (Moser and MacPhail, 1990; Haggerty, 1991) and Guidance Document on the Recognition, Assessment, and Use of Clinical Signs as Humane Endpoints for Experimental Animals Used in Safety Evaluation (OECD Environmental Health and Safety Publications Series on Testing and Assessment No. 19, 2000). Briefly, each rat in the study was carefully examined for their appearance, clonic movements, tonic movements, gait, stereotypy and bizarre behavior, if any. Assessment of appearance included parameters such as diarrhea, piloerection, salivation, palpebral closure, lacrimation, lack of grooming, bleeding from orifices and moribundity. Assessment of clonic movements included parameters such as wet dog shakes, clonic convulsions, myoclonic jerks, severe or whole body tremors, mild tremors, quivers of limbs, skin, ears or head, repetitive movements of jaws and mouth. Assessment of tonic movements included severe clonic and/or tonic convulsions causing dyspnoea, postictal depression or death, jumps with all feet leaving the surface, rigid forward extension of head and body, head and body rigidly arched backward and contraction of extensors. Assessment of gait included lameness, body drags, tiptoe walk, dragging or extension of forelimbs or inability to support weight, feet markedly point outward from the body, exaggerated or over-compensated movements of hind limbs, excessive sway, rocks or lurches and ataxia. Assessment of stereotypy involved repetitive sniffing, pacing, stereotypic grooming and circling. Assessment of bizarre behavior involved writhing or flopping, retropulsion, straub tail, self-mutilation and head weaving amongst others. The form used for recording clinical observations in experimental rats is shown in Supplementary Figure 1.

Rats (ITI of 4 × 10⁵ W256 cells, n = 3; ITI of 4 × 10⁵ HK W256 cells, n = 3) from Experiments 8 and 9 were euthanized on day 10 and day 48 with an overdose of pentobarbitone sodium (1 mL/kg of 325.73 g/L; Lethabarb^®). The tibiae were collected and fixed in 10% NBF for at least 2 days. Micro-CT scanning was conducted using a preclinical Inveon Multimodality PET/CT Scanner (Siemens Medical Soln., Knoxville, TN, United States) at the Centre for Advanced Imaging (CAI) at The University of Queensland. The μ-CT images were obtained using Inveon Acquisition Workstation software (IAW version 2.0, Siemens). The X-ray source voltage was set to 80 kV and the current to 250 μA. The scans were conducted using 360° rotation with 360 rotation steps using a medium–high magnification and with a binning factor of 2. The exposure time was 2300 ms and the CT scanning process totally took approximately 60 min. The μ-CT images were reconstructed using a Feldkamp reconstruction software (Siemens) resulting in an isotropic voxel dimension of 27.9 μm. The CT data were calibrated in Hounsfield units (HU) defined such that the water and air have 0 and 1000 HU values, respectively. The images were analyzed using Inveon Research Workstation software (IRW version 4.1, Siemens) to measure the bone volume/total volume (BV/TV ratio), trabecular thickness (T_b.T_h), trabecular spacing (T_b.S_p) and trabecular number (T_b.N) in the proximal diaphyseal regions of the ipsilateral tibial bones, as described previously (Muralidharan et al., 2013).

Rats (ITI of 4 × 10⁵ W256 cells, n = 3; ITI of 4 × 10⁵ HK W256 cells, n = 3) from Experiments 8 and 9 were euthanized on day 10 and day 48 with pentobarbitone sodium (1 mL/kg of 325.73 g/L; Lethabarb^®). Tibiae were harvested and fixed by immersing in 10% NBF (Jin et al., 2016; Liu et al., 2016) for at least 2 days. These tibiae were then immersed in 15% w/v solution of UltraPure^TM EDTA in phosphate buffer for at least 4 weeks, with the EDTA solution being changed twice per week (Hald et al., 2009). The soft decalcified bones were then rinsed, and after dehydration they were embedded in paraffin and cut into 4 μm cross-sections with a RM2235 rotary microtome (Leica Microsystems, Wetzlar, Germany) at the QIMR Berghofer Medical Research Institute, Brisbane, QLD, Australia. Sections of proximal diaphyseal regions of ipsilateral tibiae were mounted on Uber slides (InstrumeC Pty Ltd., Beaumaris, VIC, Australia) and stained using hematoxylin and eosin (H&E) (Mao-Ying et al., 2006) to assess histological changes in the tibial structure.

W256 cells were seeded onto sterile coverslips in 24 well-plates. Once the cells were 80–90% confluent, the culture medium was aspirated and the cells were briefly rinsed using PBS. The cells were fixed with ice-cold methanol (4 min at -20°C). The fixing agent was aspirated and the cells were washed with 0.2% Tween 20 and 0.1% Triton^TM X-100 in PBS for 10 min. The fixed cells were subsequently blocked for 30 min with 1% BSA in PBS at 23 ± 2°C. Anti-Cytokeratin 18 antibody [C04] (Alexa Fluor^® 488) ab187573 (1:20 dilution, Abcam, Melbourne, VIC, Australia), prepared in 1% BSA in PBS, was added to the wells containing the coverslips and the plate was incubated at 37°C for 3 h in the dark (~0.002 lux). The antibody solution in the wells was then aspirated and the cells were rinsed with PBS thrice for 5 min each. A 0.5 μg/mL solution of DAPI was added and the cells were incubated for 5–10 min. The DAPI solution was then aspirated and the cells were washed with PBS. Next, the processed cells on coverslips were mounted on Superfrost^® Plus slides (Lomb Scientific Pty Ltd., Taren Point, NSW, Australia) using Prolong^® Gold antifade reagent. The mounted slides were allowed to dry and stabilize in the dark at 4–8°C overnight, prior to image capture using a fluorescence microscope as mentioned in Section Image Acquisition.

To validate the above results of immunocytochemical staining, anti-Cytokeratin 18 antibody [C-04] ab668 (1:100 dilution, Abcam, Melbourne, VIC, Australia) was used as a second confirmatory antibody, because previous works published in Nature (Leong et al., 2008; Wang et al., 2009) have validated the suitability of this antibody in both immunocytochemistry and immunohistochemistry applications. The procedure for staining of cells with the second antibody was similar to that of the previous antibody as described above. However, after the primary antibody treatment, the cells were incubated with Goat anti-Mouse IgG (H+L) Secondary Antibody, Alexa Fluor 546 A-11030 (1:500 dilution, Thermo Fisher Scientific, Rockford, IL, United States) in PBS with 0.1% Tween 20 (PBST) for around 2 h in the dark (~0.002 lux) at 23 ± 2°C, followed by 2 × 5 min washes with PBST and 1 × 5 min wash with PBS before treatment with DAPI.

Rats (ITI of 4 × 10⁵ W256 cells, n = 3; ITI of 4 × 10⁵ HK W256 cells, n = 2) from experiment 15 and 16 were euthanized on days 7 and 38 with an overdose of pentobarbitone and then perfusion-fixed using 4% PFA. The tibiae were collected and further post-fixed with 4% PFA. The tibiae were stored in 10% NBF for 2 days at 4–8°C. These tibiae were then decalcified in a manner similar to that described in Section Tibial Bone Histology. The decalcified bones were allowed to post-fix for 2 h in 4% PFA solution (4–8°C), cryoprotected successively in 15% sucrose/PBS and 30% sucrose/PBS at 4–8°C and subsequently placed in a 1:1 mixture of OCT:30% sucrose/PBS at 4–8°C, after which they were freeze-mounted in Tissue-Tek^® O.C.T. Compound. Frozen tibial longitudinal sections (7 μm thick) were cut using a Cryostar NX70, (Thermo Fisher Scientific, Waltham, MA, United States) and mounted on Uber Plus charged slides (InstrumeC, Beaumaris, VIC, Australia). In order to identify the tumor infiltration by using immunohistochemistry, tibial sections were washed with a 1× PBS (pH 7.4) solution (3 × 5 min), followed by blocking with 10% NGS containing 0.3% Triton^TM X-100 in PBS for 1–2 h at room temperature. The sections were allowed to incubate with anti-Cytokeratin 18 antibody [C04] (Alexa Fluor^® 488) ab187573 (1:20 dilution) diluted in 2% NGS in PBST overnight at 4–8°C. This was followed by 2 × 5 min washes using PBST and 1 × 5 min wash with PBS. The sections were allowed to incubate with DAPI for 5–10 min and subsequently washed with PBS (2 × 5 min) and the cover-slips were placed along with Prolong^® Gold antifade reagent.

Similar to the procedure outlined in Section Immunocytochemistry of W256 Cells: Cytokeratin 18, anti-Cytokeratin 18 antibody [C-04] ab668 (1:100 dilution) was also used to validate the results of immunohistochemical staining. The procedure for staining of tibial sections with the second antibody was similar to that of the previous antibody described above. However, the sections were allowed to incubate with primary antibody overnight at 23 ± 2°C, and then incubated with Goat anti-Mouse IgG (H+L) Secondary Antibody, Alexa Fluor 546 A-11030 (1:600 dilution) in PBST for 2 h in the dark (~0.002 lux) at 23 ± 2°C. This was followed by 2 × 5 min washes using PBST and 1 × 5 min wash with PBS and the sections were subsequently treated with DAPI and mounted as described above.

For the histology experiments, images were captured with an Aperio ScanScope XT system (Leica Biosystems, Nussloch, Germany) located at the School of Biomedical Sciences, The University of Queensland and processed using Aperio ImageScope v12.3.0.5056 software (Leica Biosystems). Images from immunocytochemistry and immunohistochemistry experiments were captured with an Axioskop 40 microscope (Carl Zeiss, Göttingen, Germany) attached to an Axiocam MRm camera (Carl Zeiss) and processed using AxioVision Rel. v4.8 software (Carl Zeiss). For the immunocytochemistry and immunohistochemistry experiments, images were acquired at a fixed exposure time, optimized using auto-exposure settings of AxioVision Rel. v4.8 software and by using filters suitable for the fluorophore of the secondary antibody.

All the values have been expressed as mean ± standard error of the mean (SEM). For treatments having n ≤ 2, the values are expressed as mean or absolute value, without calculating the error. Generation of graphs and data processing were done using the GraphPad Prism^TM (v7.00) software package. The PWT values of rats administered single bolus doses of test compound or vehicle were normalized by subtracting the pre-dosing baseline values so as to obtain ΔPWT values as follows:

Area under the curve (AUC) for ΔPWT versus time curves (ΔPWT AUC values) was calculated using trapezoidal integration to determine the extent and duration of action of test compounds for relief of allodynia. ΔPWT AUC values were then converted into a percentage of the maximum possible ΔPWT AUC (% MAX ΔPWT AUC) with the following formula:

Dose–response curves were generated by plotting mean (±SEM) % MAX ΔPWT AUC values versus log dose of each of the test compounds. The PPT data and the naloxone data were processed in a similar way as described above. Non-linear regression (GraphPad Prism^TM v7.00) was used to determine the ED₅₀ values for each drug treatment against mechanical allodynia in the ipsilateral as well as the contralateral hindpaws.

Statistical analyses were performed using the GraphPad Prism^TM v7.00 software package. The criterion of statistical significance was p ≤ 0.05. Two-way analysis of variance (ANOVA) followed by the Bonferroni test was used to analyze between-group differences in body weights, behavioral data, pharmacological data and tibial bone morphometric changes. The Mann–Whitney test was used to assess between group body weight differences with missing intermediate values and to compare differences in ΔPPT AUC values for each test compound- and vehicle-treated group. One-way ANOVA followed by the Dunnett’s test was used in order to compare differences in ΔPWT AUC values for each test compound- and vehicle-treated groups. The unpaired t-test was used in order to compare the extent and duration of naloxone induced rescue of the pain phenotype. For statistical comparisons using ANOVA, F-values were reported along with their associated degrees of freedom (treatment, time, interaction and residual). For two-way ANOVA, F-values were expressed as F(df of treatment, time, interaction/residual). Whereas, for one-way ANOVA, F-values were expressed as F(df of treatment, residual).

In terms of body weight, there were no significant differences between animals given an ITI of W256 and HK W256 cells in experiment 1 [F_{(7,17,119/238)} = 1.2, 1125, 1.5; p > 0.05], experiment 2 [F_{(1,43,43/387)} = 2.2, 619.7, 0.9; p > 0.05] (Figure 2A), experiment 3 (p > 0.05) (Figure 2B), experiment 5 (p > 0.05) (Figure 2C), experiment 7 [F_(1,14,14/98) = 0.7, 766.6, 1.3; p > 0.05], experiment 8 [F_(1,8,18/64) = 0.4, 400.8, 2.7; p > 0.05], experiment 9 [F_(1,11,11/88) = 1.1, 364, 0.05; p > 0.05], experiment 10 [F_(1,8,8/64) = 0.3, 180.6, 2.8; p > 0.05], experiment 11 [F_(1,8,8/64) = 1.0, 314.8, 1.6; p > 0.05], experiment 12 [F_(1,7,7/63) = 0.5, 66.0, 0.2; p > 0.05], experiment 13 [F_(1,7,7/49) = 1.1, 195.4, 0.8; p > 0.05] and experiment 14 [F_(1,9,9/63) = 0.7, 141.5, 1.6; p > 0.05]. Additionally, there were no significant differences in between body weights of animals given an ITI of DPBS and age-matched control female Wistar Han rats in experiment 4 (p > 0.05). Although there were slight differences between body weights of rats administered W256 and HK W256 cells in experiment 6 [F_{(1,12,12/108)} = 8.9, 508.1, 3.1; p ≤ 0.05], experiment 15 [F_(1,6,6/18)= 11.9, 84.1, 1.3; p ≤ 0.05] and experiment 16 [F_(1,11,11/33)= 62.6, 314.2, 0.6; p ≤ 0.05] at some time points, these were either due to differences in initial body weights or due to non-cancer factors. Importantly, there were consistent gains in the body weights of all animals throughout the experiments (Supplementary Figure 2).

A total of 173 rats were used in this study, out of which 9 rats (5%) were found to have health-related issues and of these 3 (1.7%) were euthanized for ethical reasons as outlined below. One rat from experiment 1 given an ITI of 4 × 10³ W256 cells had a slight tumor like appearance in the bone near the injection site. Nevertheless, the overall health of the animal was satisfactory and no other complications were observed. Another rat from experiment 1 given an ITI of 1.5 × 10⁵ W256 cells had a mild infection at the sight of surgery with visible pus exudate. Under anesthesia, the infected site was cleaned and topical antibiotic powder was applied, followed by a s.c. injection of benzylpenicillin at a dose of 60 mg. Subsequently, the infection resolved. One rat given an ITI of 4 × 10³ W256 cells, two rats given an ITI of 1.5 × 10⁵ W256 cells and one rat given an ITI of 4 × 10⁵ HK W256 cells from experiment 1 had mild swelling at the injection site, which resolved spontaneously. One rat from experiment 4 given an ITI of DPBS was excluded from the study and euthanized due to lack of normal body weight gain due to a crooked incisor. One rat given an ITI of 4 × 10⁵ W256 cells from experiment 5 was excluded and euthanized due to the presence of a large external tumor growth on the injected tibia. One rat from experiment 6 given an ITI of 4 × 10⁵ W256 cells was excluded and euthanized due to lethargic behavior and low body weight; upon post-mortem investigation, a large metastatic lesion was observed on the intestine.

Using the anti-Cytokeratin 18 antibody [C04] (Alexa Fluor^® 488) ab187573 (Abcam), cultured and fixed W256 cells showed immunofluorescent staining (Figure 11) which was in accord with Cytokeratin 18 expression in ATCC W256 cells reported by others (Lewis et al., 2013). Similar results were obtained with anti-Cytokeratin 18 antibody [C-04] ab668 (Abcam) which was used as a second confirmatory antibody (Supplementary Figure 9).