Nanoparticle-mediated Synergistic Drug Combination for Treating Bone Metastasis

Abstract

Bone metastasis at an advanced disease stage is common in most solid tumors and is untreatable. Overexpression of receptor activator of nuclear factor κB ligand (RANKL) in tumor-bone marrow microenvironment drives a vicious cycle of tumor progression and bone resorption. Biodegradable nanoparticles (NPs), designed to localize in the tumor tissue in bone marrow, were evaluated in a prostate cancer model of bone metastasis. The combination treatment, encapsulating docetaxel, an anticancer drug (TXT-NPs), and Denosumab, a monoclonal antibody that binds to RANKL (DNmb-NPs), administered intravenously regressed the tumor completely, preventing bone resorption, without causing any mortality. With TXT-NPs alone treatment, after an initial regression, the tumor relapsed and acquired resistance, whereas DNmb-NPs alone treatment was ineffective. Only in the combination treatment RANKL was not detected in the tumor tibia, thus negating its role in tumor progression and bone resorption. The combination treatment was determined to be safe as the vital organ tissue showed no increase in inflammatory cytokine or the liver tissue ALT/AST levels, and animals gained weight. Overall, dual drug treatment acted synergistically to modulate the tumor-bone microenvironment with encapsulation enhancing their therapeutic potency to achieve tumor regression.

Graphical Abstract

1. Introduction

Bone is the most common site for cancer metastasis as the bone marrow provides favorable conditions [1] such as sluggish blood flow, cell adhesion receptors on sinusoidal walls, growth factors, and cytokine-rich microenvironment that help in establishing metastatic foci [2–4]. Although bone metastasis occurs in most cancers at an advanced disease stage, it is particularly prevalent (68–73%) in prostate and breast cancer patients [5]. Quality of life of patients with bone metastasis declines rapidly as they experience skeletal-related events (SREs) [6], and their life expectancy (<3 years) is significantly shortened [7]. Crosstalk between the cancer cells and the bone cells in the bone marrow, mediated by the transcription factor, receptor activator of nuclear factor-κB ligand, RANKL, drives tumor growth progression and bone resorption [8]. Further, the RANK-expressing cancer cells are specifically attracted to the bone [9–11], where a high local concentration of RANKL exists, thus further aggravating metastasis [12]. Hence RANKL is considered a master player with multifunctional roles—from cancer cell seeding into the bone marrow to modulating the bone microenvironment — to foster the vicious cycle of tumor progression and bone resorption [13]. Although bone modifying therapies such as bisphosphonates [14] and Denosumab (DNmb), humanized monoclonal antibody that binds to RANKL [15] improve the quality of life by minimizing SREs, these treatments are associated with serious toxicities (e.g., osteonecrosis, hypocalcemia) and patient survival is not prolonged [16]. Localized radiation treatment is mainly used as palliative therapy to relieve bone pain [17]. Since bone is less perfused (e.g., 7% of the cardiac output goes to the bone), intravenously administered anticancer drugs do not achieve therapeutic levels at the bone metastasized sites to cause tumor regression [18].

In addition to improving drug delivery to the tumor tissue in bone marrow, a mechanistic approach is needed to treat bone metastasis. In this study, we tested the efficacy of sustained-release, biodegradable nanoparticles (NPs), designed to localize to the tumor tissue in the bone marrow [19], encapsulating docetaxel, an anticancer drug (TXT-NPs), and separately Denosumab (DNmb-NPs). With TXT acting on cancer cells and DNmb binding to RANKL, we hypothesized that the combination treatment (TXT-NPs + DNmb-NPs) would modulate the tumor-bone marrow microenvironment, particularly inhibiting RANKL formation, ultimately affecting tumor progression. We also anticipated that NPs with sustained release characteristics would potentiate the therapeutic efficacy as well as enhance biocompatibility of the treatment. To mimic the advanced disease stage of prostate cancer, we tested the efficacy of the treatments in an intraosseous model of bone metastasis induced via intratibial injection of prostate cancer cells (PC3-Luc), which forms an osteolytic bone lesion and separately, in a hematogenous metastasis model induced via intracardiac injection of PC3-Luc that metastasizes to different soft tissue organs.

2. Material and Methods

2.1. Materials:

Poly (d,l-lactide co-glycolide, 50:50; 0.76–0.94, PLGA) was purchased from Evonik Corporation (Birmingham, AL). Poly Vinyl Alcohol 8–88 (PVA) was purchased from EMD Millipore (Burlington, MA). L-(+)-Tartaric acid dimethyl ester (DMT), Sodium azide, Glucose, and Cremophor EL (CrEL) were purchased from Millipore Sigma (Burlington, MA). Docetaxel (TXT) was purchased from Alpha Aesar (Tewksbury, MA). Denosumab (DNmb) (XGEVA, Amgen Inc, Thousand Oaks, CA) was obtained from Cleveland Clinic Pharmacy. Human serum albumin (HSA) was purchased from Lee Biosolutions (Maryland Heights, MS). Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (DMEM-F12), Dulbecco’s phosphate-buffered saline (DPBS), 0.05% Trypsin-EDTA, Penicillin, and Streptomycin were obtained from the Cell Services and Media Core Facility of the institute. Heat-inactivated Fetal Bovine Serum (FBS) was purchased from ThermoFisher Scientific (Waltham, MA). Luciferin was purchased from Promega Corporation (Madison, WI). Organic solvents used were of HPLC grade.

2.2. TXT-NPs and DNmb-NPs:

TXT-NPs were formulated using a single, oil-in-water (o/w) emulsification solvent evaporation method, whereas DNmb-NPs were developed using a multiple, water-oil-in-water (w/o/w) emulsification solvent evaporation method. In a typical protocol for the formulation of TXT-NPs, 810 mg PLGA, 90 mg DMT, and 100 mg TXT were dissolved in 10 mL chloroform. DMT, a plasticizer and pore-forming agent, was used to facilitate the release of the encapsulated drug [20]. The polymer-drug solution was emulsified into 60 mL of 6% w/v PVA solution using sonication (UP200St, Hielscher Ultrasonic, Teltow, Germany), followed by homogenization at 10,000 psi for 15 min (EmulsiFlex-C5, Avestin, Inc., Ottawa, ON, Canada). The emulsion formed was stirred overnight to evaporate chloroform; the formed TXT-NPs were recovered using Tangential Flow Filtration (TFF, KrosFlo^® KR2i, Spectrum, now Repligen, Waltham, MA) using the column, D04-S05U-05-N (0.05 μm, Repligen). The conditions for TFF were optimized so that >95% of PVA was removed from the formulation. Previously, we have shown that a fraction of PVA remains associated with PLGA-NPs at the interface, termed the “residual PVA,” that cannot be removed from the NP interface despite repeating washings and modulates the surface properties (zeta potential) of NPs [21].

To formulate DNmb-NPs, 81 mg PLGA and 9 mg DMT were dissolved in 2 mL chloroform. Separately, to 300 μL DNmb solution, 21 mg HSA was dissolved by mixing; the formed solution was emulsified into the PLGA solution, first by vortexing for 2 min, followed by sonication (40% output, Model Q500, QSonica Sonicators, Newtown, CT) using a microtip probe for 2 min over an ice bath to form w/o emulsion. The above w/o emulsion was further emulsified into 3% w/v, 18 ml PVA solution, first by vortexing followed by sonication as above for 4 min to form w/o/w emulsion. Following overnight (~18 hours) evaporation of the chloroform with stirring, the formed DNmb-NPs were recovered by ultracentrifugation at 4 °C for 30 min (30,000 rpm Rotor 50.2Ti, Beckman L80, Beckman Coulter, Inc., Brea, CA). The supernatant was collected, and the pellet was re-suspended in 10 mL of MilliQ water by vortexing and then sonicated as above for 2 min. The formed dispersion of NPs was centrifuged again, the supernatant was collected, and the pallet was suspended and centrifuged again. The supernatant/washings were analyzed for the DNmb levels using ELISA. Following the last washing step, the sediment was resuspended into 750 μL water, to which 750 μL 2% w/v glucose solution (glucose used as a cryoprotectant) was added. The formulations were lyophilized for 2 days at −48 °C, 3.5 Pa (FreeZone 4.5, Labconco Corp., Kansas City, MO).

2.3. Physical Characterization of NPs:

Mean hydrodynamic diameter and zeta potential of TXT- and DNmb-NPs were determined in water using a dynamic light scattering (DLS) technique (Nicomp 380 ZLS, Particle Sizing Systems, Santa Barbara, CA). The size of NPs was measured at a scattering angle of 90° at 25 °C and zeta potential in the phase-analysis mode and the current mode at a scattering angle of −14°. For transmission electron microscopy (TEM), 5 μL of the NP-dispersion in water (~100 μg/mL) was placed on a 200 mesh Formvar-coated TEM grid with a size of 97 μm (TED PELLA, Redding, CA). The samples were negatively stained with 2% w/v uranyl acetate solution, the excess stain was removed with filter paper, and the grid with coated NPs was washed with sterile MiliQ water and air-dried for 3 hours. The coated samples were observed with an electron microscope operating at 80 kV (Tecnai G2 SpiritBT, FEI Company, Hillsboro, OR). The mean TEM NP diameter (n=125–175) was measured using Image J software from ~ 20 images.

2.4. TXT and DNmb Loading in NPs:

To the lyophilized TXT-NP formulation, methanol (1 mg/mL) was added, and the samples were kept on a shaker for 48 hours (Environ Orbital Shaker, Labine, Melrose Park, IL). The samples were centrifuged at 4,000 rpm for 10 min at 4 °C (Sorvall Legend RT Centrifuge), and the supernatants were analyzed using HPLC. An indirect method was used to determine DNmb loading in NPs, i.e., by analyzing the amount of DNmb that is not encapsulated in NPs. For this, the supernatant and washings from the DNmb-NP formulation were collected and analyzed for the DNmb content using ELISA (SHIKARI^® Q-DEN kit, Iwai North America Inc., San Carlos, CA). From the total amount of DNmb added into the formulation and that detected in the washings (non-encapsulated), the encapsulated DNmb amount was calculated. This method was adopted since extracting macromolecules from PLGA-based NPs using organic solvents is incomplete. Also, there is a possibility of denaturation/inactivation of macromolecules when these are extracted from NPs using organic solvents. Previously, we have used the above indirect method of analyzing protein or enzyme loading in the PLGA-based formulation of NPs [22].

2.5. TXT and DNmb Release from NPs:

Release of TXT and DNmb from the respective formulations of NPs was carried out in vitro in double diffusion chambers, separated by a Millipore membrane (0.05 μm porosity, VMWP01300 EMD Millipore, Burlington, MA). For the TXT release study, each donor chamber was filled with 2.5 mL of TXT-NP suspension (1 mg/mL) prepared in a pH 7.4 mannitol citrate buffer containing 0.1% (v/v) Tween-20 (Sigma) and 10% ethanol to maintain sink condition and 0.1% sodium azide as a preservative. The receiver chamber was filled with buffer only. Diffusion cells were placed in a closed box to avoid exposure to light, and the box was placed on a shaker rotating at 100 rpm at 37 °C (Environ Orbital Shaker). At regular time intervals, the entire content from each receiver chamber was removed and replaced with a fresh buffer. The samples collected were lyophilized and extracted in 1 mL of methanol and analyzed for TXT levels using the HPLC method as described below. For the release study from DNmb-NPs, the formulation was dispersed at 1 mg/mL concertation in pH 5.2 Sorbitol Acetate buffer (4.6% w/v sorbitol, 18 mM Sodium Acetate, 0.1% sodium azide as preservative). We selected this buffer because DNmb (XGEVA, Amgen Inc.) is provided in this buffer. The released samples collected from the receiver chambers were measured for DNmb levels using ELISA. Similar studies were conducted with drugs (TXT and DNmb) in solution to ensure that the membrane partitioning the chambers is not a barrier to diffusion of the released drugs from the donor to the receiver chamber.

2.6. Analytical Methods:

HPLC (Shimadzu Scientific Instruments, Inc., Columbia, MD) conditions were: mobile phase – methanol: acetonitrile: water (50:30:20 v/v); flow rate – 1mL/min; detector UV; wavelength – 230 nm. Injection volume – 25 μL; retention time – 4.1 min. Reverse phased C18 column (Supelcosil; dimension – 25 cm × 4.6 mm; porosity 5 μm) purchased from Sigma-Aldrich. To determine TXT levels in the tissue sample, LC/MS/MS was used. In brief, the system included a Vanquish UHPLC binary pump with an online degasser, autosampler, column heater, and a TSQ Quantiva triple quadrupole mass spectrometer with a heated electrospray (H-ESI) ion source (ThermoFisher Scientific, Waltham, MA). The column used was a Gemini C18, 2 × 150 mm, 3 μm analytical column (Phenomenex, Rancho Palos Verdes, CA). Two eluents were used: eluent A contained water with 0.2% formic acid, and eluent B contained acetonitrile with 0.2% formic acid. A gradient was formulated as follows: 40%B to 100% B over 8 min, hold at 100% B for 4 min and return to 40% B for 8 min. The flow rate was 0.3 ml/min, and the injection volume was 5 μL. Mass spectrometer source parameters included a positive spray voltage of 3500 V, sheath gas of 60 units, aux gas of 15 units, sweep gas of 2 units, ion transfer tube temperature of 380 °C, and vaporizer temperature of 350 °C. Positive ionization with selective reaction monitoring (SRM) precursor/product ion pairs was collected: TXT at m/z 808.4/105 and the sodium adduct of TXT at 830.4/304. The SRM sodium adduct gave higher sensitivity and was used for tissue drug amount calculation. Collision energies were 40 V for non-sodium adducts and 30 V for sodium adducts. The integrated switching valve was configured to divert UHPLC flow to waste rather than the ion source for the first 2 min to prevent unrelated compounds from fouling the ion source. Software XCaliber was used to process the data and get the peak areas of TXT.

2.7. Cell Line:

PC3-Luc cell line (luciferase gene expressing human prostate cancer cells) was purchased from PerkinElmer^® (Waltham, MA) and cultured in T-75 culture flasks (USA Scientific, Ocala, FL) with DMEM-F12 medium supplemented with 5% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C with 5% CO₂ atmosphere. Human aortic smooth muscle cells (SMCs) were obtained from Cell Applications Inc (San Diego, CA). The cells were cultured in the smooth muscle cell growth medium (Cell Applications Inc., Cat# 311–500) as above in T-75 culture flasks.

2.8. Cytotoxicity Study:

PC3-Luc cells (2,500 cells/well) and SMCs (3,000 cells/well) were seeded into 96-well plates (Corning Inc., Corning, NY) for 24 hrs prior to the treatment either with TXT in solution (8 mg/mL dissolved in ethanol stock followed by serial dilutions in cell culture media) or the equivalent doses of TXT-NPs dispersed in cell culture media. For treatments, the medium in 96-well plates was replaced with a 200 μL medium containing drug (TXT-solution/TXT-NPs) and incubated for 5 days. Cell viability was determined using the CyQUANT^™ NF Cell Proliferation Assay kit (Life Technologies, Carlsbad, CA). The IC₅₀ for each treatment was calculated using the non-linear curve fit (logistic) using OriginPro 8 (OriginLab Corp., Northampton, MA). Similarly, the cytotoxicity of DNmb solution, DNmb-NPs, and control NPs was determined.

2.9. Tumor Inoculation:

The Cleveland Clinic’s Institutional Animal Care and Use Committee approved all animal procedures, which were carried out according to Federal and internal guidelines. Athymic^nu\nu mice (Male, aged 6–8 weeks, Envigo, Indianapolis, IA) were used. Before inoculation, PC3-Luc cells cultured as above were harvested using 0.05% Trypsin-EDTA solution, resuspended in serum-free DMEM-F12 medium, and counted using Countess Automated Cell Counter (Invitrogen, Waltham, MA). The cells were kept in an ice bath and were used within 30 min after harvesting for tumor inoculation. The studies were carried out in intraosseous and hematogenous models of metastases. The intraosseous model of bone metastasis was established by injection of 5×10⁴ PC3-Luc cells in 10 μL of serum-free DMEM-F12 medium into the medullary cavity of the right tibia using a 26G-1/2” needle mounted on a Hamilton syringe (Hamilton Company, Reno, NV). To establish the hematogenous metastasis model, 1×10⁵ PC3-Luc cells in 100 μL of serum-free DMEM-F12 medium were injected into the left ventricle (a little left from the midpoint between the top of the sternum and xiphoid process, left 3^rd intercostal space) using 27G-1/2” needle attached to ½ mL syringe (BD bioscience, Mississauga, Ontario). Tumor development was confirmed from the bioluminescence signal of PC3-Luc cells, as described below.

2.10. Treatment Protocol:

The formulation of TXT-NPs (50 mg/mL) was dispersed in normal saline by sonication for one minute over an ice bath as above, whereas the formulation of DNmb-NPs (10 mg/mL) was dispersed into normal saline by gentle mixing as it contained glucose as cryoprotectant. TXT solution (1 mg/ml) was prepared into Cremophor EL (Cr-EL) and diluted 1:1 v/v with ethanol (200 proof Ethyl Alcohol, Pharmco Inc., Brookfield, CT). DNmb (XGEVA) solution was prepared separately by diluting (1 mg/mL) in normal saline. Mice with established tumors in the tibia (bioluminescent signal, 1.3 × 10⁴ to 3.98 × 10⁵, Average 1.11 × 10⁵) were divided into five treatment groups: saline control, DNmb-NPs, TXT-NPs, TXT-NPs + DNmb-NPs in combination, and TXT-CrEL + DNmb Sol in combination. The treatment dose of TXT (NPs or drug in Cr-EL) was 12 mg/kg, and that of DNmb (NPs or solution form) was 3 mg/kg. Following the first dose, subsequent doses were administered at 6 weeks intervals. In the hematogenous model study, mice were divided into two groups (bioluminescent signals ranging from 2.05×10⁴ to 1.10×10⁷, Average 2.92×10⁶), saline control and combination treatment (TXT-NPs + DNmb-NPs), and the treatments were given every 4 weeks.

2.11. Monitoring Tumor Progression:

Tumor growth was monitored by measuring the bioluminescence signal of cancer cells. Mice were injected intraperitoneally with 250 μL of Luciferin solution (20 mg/mL in sterile normal saline); after 3 minutes of wait, consecutive images were taken, each with 3 minutes of exposure time using IVIS Lumina II^® in-vivo live animal imaging system (PerkinElmer, Waltham, MA) until the tumor bioluminescence signal reached the peak. For the intraosseous bone tumor model, average radiance (p/s/cm²/sr) was measured after manually selecting the tumor perimeter as the region of interest (ROI) with Living Image Software (IVIS Imaging Systems). All the tumor foci were individually selected as the region of interest (ROI) for the hematogenous metastasis model, and the average radiances were measured for all the metastatic foci and combined to represent the total tumor burden. Background signal was obtained from the animals which were not inoculated with tumor cells but received Luciferin solution as above. Animals were euthanized if any of the following occurred: loss of 20% or more of the initial body weight of the animal, tumor mass exceeding 10% of body weight, tumor ulcer, necrosis or bleeding, or severe interruption in locomotion due to tumor mass, were considered as the endpoint for those animals. The tissues collected at the end stage were used to analyze DNmb, RANKL, and inflammatory cytokine levels, as described below.

2.12. Micro-CT Scans to Determine Osteolytic Bone Lesions:

Representative animals from each treatment group were imaged using micro-computed tomography (micro-CT) in situ to determine bone loss. The images were acquired on the GE eXplore Locus Micro-CT (General Electric Company; Boston, MA). Mice were scanned in the supine position using an X-ray source voltage of 80kVp, tube current of 490 μA, and a CCD exposure time of 1800 ms. The GE Locus Micro-CT is equipped with a rotating gantry that rotates by a 1° angle of increment, acquiring 360 views and 2 frames per view. The detector data binning was set at 1:1 to achieve an isotropic voxel resolution of 0.02 mm. The resultant image was utilized for post-acquisition reconstruction on GE Reconstruction Utility Software provided by the manufacturer to process the acquisition data.

2.13. Calculating Bone Parameters from Micro-CT Images:

Following post-acquisition reconstruction, micro-CT volumes were further calibrated using air, water, and cortical bone phantoms. The micro-CT volume grayscale values (corresponding to Hounsfield Units) were normalized to the phantom. A standard grayscale threshold level could be applied to all volumes for bone metrics. Bone structure metrics, bone mineral density (BMD), bone mineral content (BMC), and bone volume fraction (BV/TV) were determined using the GE Microview Software (Version MicroView ABA 2.2). All the volumes were registered to the same orientation in 3D space and cropped to a standard ROI for each animal. The right tibia with tumor and the left tibia (control) were evaluated for bone losses. The ROIs were manually marked using the cortical/trabecular bone segmentation tool. The bone analysis tool determined subsequent analysis and generation of global threshold-based metrics. Bone mineral content and bone density from each treatment group were compared with a healthy mouse bone of similar age (having no tumor and no treatment).

2.14. Biodistribution of TXT:

Mice with intratibial tumors with an average radiance of 10×10⁶ p/s/cm²/sr were administered, either TXT-NPs or TXT-CrEL. One week after receiving the treatments, animals were euthanized, blood was drained via cardiac puncture, and different organs, including tumor tissue from the bone marrow, were harvested, pressed on blotting paper to remove residual blood, and frozen at −80 °C before taking for the analysis. After thawing the tissue to room temperature, ~100 mg from each organ tissue was weighed, chopped into small pieces, and homogenized in 1 mL of tissue lysis buffer (RIPA Buffer, Millipore Sigma) using a tissue homogenizer (Minilys, Bertin Instruments, France). Before homogenization, paclitaxel (TXL, Alfa Aesar) 1 ng/mL was added to each tissue sample as an internal standard. The lysed tissue samples were lyophilization (FreeZone 4.5) for 48 hours, and the drug was extracted in 1 mL of methanol for 48 hours at 37 °C on an orbital shaker. The samples were analyzed using LC-MS/MS as described above.

2.15. Biocompatibility:

To athymic mice (without tumor), either TXT-NPs or TXT-CrEL (12 mg/kg) was administered intravenously. One week post-administration, animals were euthanized, and the blood was collected via cardiac puncture. Subsequently, the blood was drained prior to harvesting vital organs (lung, heart, liver, kidney, and spleen). To obtain serum, blood was kept at room temperature for 30 minutes for clot formation and then it was centrifugation at 1400 g for 20 minutes, and the supernatant was collected. A piece of each collected tissue sample (~100 mg) was homogenized as described above. The serum samples and tissue homogenates were analyzed for alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels using ELISA kits (Abcam, Cambridge, MA). The remaining harvested organ tissues were fixed with 4% paraformaldehyde, paraffin-embedded, sectioned (20 μm), and stained with hematoxylin and eosin for histological analysis. Serum and tissue from animals that did not receive any treatment were used as control.

2.16. End-stage Tissue Analysis:

The tissues collected at the end of each study were analyzed for DNmb, RANKL, ALT, AST, and inflammatory cytokines (TNF-α, IL-6). The tissues were homogenized in lysis buffer and analyzed using ELISA kits (RANKL, IL-6, Abcam, Waltham, MA; and TNFα, Thermofisher Scientific). The data were normalized to protein levels (BCA Protein Assay kit, ThermoFisher Scientific).

2.17. Statistical Analysis:

Data are represented as the mean ± s.e.m. Statistical significance between the groups was calculated using GraphPad Prism. Animal survival data were plotted using the Kaplan-Meier method, and statistical significance was determined using Log-rank (Mantel-Cox) test. Differences were considered significant at p < 0.05.

3. Results and Discussion

3.1. Characteristics of NPs and cytotoxic effect in vitro:

In our previous study, we determined that the NPs with close to neutral zeta potential (+/− 5 mV) have better uptake in the tumor-bearing tibia than those with highly cationic or anionic zeta potential. For this, we modulated the surface-associated PVA. Greater localization of neutral NPs in the tumor tissue in the bone marrow is attributed to their reduced opsonization [19]; hence they remain in the circulation longer than highly anionic or cationic NPs for extravasation through the fenestrations or clefts between the endothelial cells of the sinusoidal capillaries into the bone marrow [23]. Some of these clefts are as wide as ~170 nm [24, 25], and these openings are significantly larger (380 to 780 nm) in the tumor tibia due to overexpression of multiple proangiogenic growth factors, creating vascular niches [26]. Thus, the circulating NPs smaller than the openings in the clefts can pass through these fenestrations into sinusoidal capillaries in the bone marrow [27]. Accordingly, we formulated TXT-NPs and DNmb-NPs with almost neutral zeta potential (Figures 1A, B, and Table 1).

Figure 1: Physical characterization of TXT-NPs and DNmb-NPs and their cytotoxic effects in PC3 cells:

A) Characterization of TXT-NPs (a) Histogram showing Gaussian Size Distribution, (b) Transmission Electron Micrograph (TEM), (c) Release of TXT from NPs in vitro under sink condition, Data as mean ± s.e.m., n=3, and (d) Cytotoxicity of TXT-NPs as compared to TXT-Sol in PC3 Luc cells. The cells were incubated with TXT-NPs or TXT-Sol for 5 days; data as mean ± s.e.m. (n=6). B) Characterization of DNmb-NPs (a) Histogram showing Gaussian Size Distribution, (b) Transmission Electron Micrograph (TEM), c) Release of DNmb in vitro under sink condition, Data as mean ± s.e.m., n=3, and (d) Cytotoxicity of DNmb-NPs and control NPs (without any encapsulated drug) with an incubation time of 5 days; data as mean ± s.e.m. (n=6).

Table 1:

Physical characterization of TXT- and DNmb-NPs

Parameter	TXT-NPs	DNmb-NPs
Hydrodynamic diameter	252 ± 21nm	330.5 ± 17 nm
Polydispersity Index	0.07 ± 0.02	0.05 ± 0.02
TEM diameter	104 ± 3.5 nm	112 ± 2.9 nm
Zeta Potential	−0.6 ± 0.2 mV	−0.05 ± 2.2
TXT/DNmb loading	6.6 ± 0.05% w/w	14.7 ± 0.12%
Encapsulation Efficiency	66.3 ± 0.05%	69.8 ± 0.6%

Data as mean ± s.e.m., from n=5 to 7 batches of NP formulations. TEM diameter, an average of n=125–174 nanoparticle size measurements

The hydrodynamic diameter determined using DLS measures the size of NPs in a hydrated state, with the hydration of the NP surface-associated emulsifier, polyvinyl alcohol (PVA), contributing to the diameter, whereas the TEM diameter measures it in a dry state, hence a discrepancy in the two types of size measurements [28]. Polydispersity indices <0.1 indicate uniform particle size distribution (Table 1). Interestingly, despite very close TEM diameters of TXT- and DNmb-NPs (104 ± 3.5 nm vs. 112 ± 2.9 nm), there was a noticeable difference in their hydrodynamic diameters (252 ± 21 nm vs. 330.5 ± 17 nm) (Table 1). Since DLS measures size in a hydrated state of NPs, it is possible that DNmb, which is an antibody and hydrophilic in nature, creates DNmb-NPs with a greater degree of hydration than TXT-NPs, as TXT is hydrophobic. This is because a fraction of the encapsulated therapeutics would migrate to the NP interface, influencing the hydration of NPs and hence their hydrodynamic diameter. Such changes in size are not reflected in the TEM images as they are measured in a dry state.

In addition, in the TEM images, TXT-NPs show a white and clear surface, whereas DNmb-NPs appear spotty with a more contrasting border than TXT-NPs (Figures 1A–a and 1B–b). For TEM imaging, 2% w/v uranyl acetate solution is used to counterstain NPs; and it is most likely binding more to the DNmb present at the NP interface because of its macromolecular nature than to the TXT at the NP interface. Another possibility could be that DNmb-NPs are porous because of the macromolecular nature of the encapsulated antibody and HSA (used as an inert protein) that are heterogeneously dispersed in the PLGA-NP polymer matrix; hence uranyl acetate could potentially diffuse through the surface of DNmb-NPs but not through the TXT-NPs as TXT is hydrophobic and homogeneously dispersed in the PLGA-NP polymer matrix.

In the formulation of DNmb-NP, HSA was used in the aqueous phase of w/o emulsion to stabilize the emulsion due to its emulsifying property, without which the emulsion with DNmb alone is unstable and could not form NPs [20]. In addition, HSA protected DNmb from interfacial inactivation, which commonly occurs when macromolecular therapeutics (e.g., proteins and peptides) are exposed to the water-organic solvent interface, such as in this case, while making the w/o emulsion. The exposed hydrophobic domains of macromolecules at the water-oil interface can interact to denature the protein. To support the above role of HSA, DNmb solution with and without HSA was vortexed with chloroform and then kept on an orbital shaker at 37 °C for 1 hour. The samples were centrifuged at 4,000 rpm for 10 min at 4 °C (Sorvall Legend RT Centrifuge), and the top aqueous phase was used to measure antibody levels using ELISA. There was no effect of HSA on the standard plot of DNmb. The data demonstrated significantly higher retention of DNmb activity in the presence of HSA as compared to without it (59% vs. 86%) (Supplemental Data 1). This protective effect of albumin is attributed to its preferential accumulation at the water-oil interface and is resistant to interfacial inactivation [29]. We have used this strategy to encapsulate antioxidant enzymes into PLGA-based NPs [22].

The formulations of TXT-NPs and DNmb-NPs demonstrated sustained release of the encapsulated therapeutics, with the release of DNmb relatively slower than that of TXT due to the differences in their molecular weights (TXT, small molecular vs. DNmb, macromolecule) (Figures 1A–c and 1B–c). Since TXT or DNmb in solution equilibrates rapidly across the membrane placed between the donor and receiver chambers; sustained release seen is due to their encapsulation into NPs (Figures 1A–c and 1B–c). With TXT-NPs, IC₅₀ in PC-3 cells was higher than with TXT-solution (IC₅₀ = 1.040 ng/mL vs. 0.575 ng/mL) (Figure 1A–d), which is expected considering sustained drug-releasing characteristics of NPs. DNmb-NPs did not show any cytotoxic effect on cancer cells, and the results were similar to that of control NPs (Figure 1B–d). Similar experiments with TXT-NPs and TXT-solution in SMCs used as a representative normal cell line showed IC₅₀ of 27.5 ng/mL for TXT-NPs and 15.8 ng/mL for TXT-solution. These values are significantly higher (~25-fold) than for PC-3 cells, indicating the selective cytotoxicity effect of TXT towards PC-3 cancer cells than normal cells. Others have also reported the differential effect of anticancer drugs, less toxic to normal cells than to cancer cells, and this phenomenon has been attributed to drug effects on the cell cycle and mitosis [30, 31]. Although not determine in this study, it is also possible that cancer cells have higher uptake of free or encapsulated drugs as they divide more rapidly than SMCs (~25 hrs vs. ~75 hrs), as the uptake could occur during cell division [32].

3.2. Effect of treatments in bone metastasis model of prostate cancer:

Saline control and DNmb-NP-treated animals showed rapid tumor progression, declining body weight, and increasing mortality with time (median survival saline vs. DNmb-NPs 9.5 vs. 11.5 wks, p = n.s.). The treatment with TXT-NPs showed an initial tumor regression, and the animals gained weight, but after ~ 24 weeks, the tumor relapsed, and further dosing of TXT-NPs was not effective in suppressing the tumor growth, resulting in a rapid decline in body weight and increased mortality. Overall, the median survival was significantly better for the animals treated with TXT-NPs than with DNmb-NPs or compared to saline control (Median survival =39.5 wks vs. 9.5 wks for saline or 11.5 wks for DNmb-NPs, p= 0.01) (Figure 2). We did not include the TXT-CrEL treatment group in this study because, in our previous studies, we found no significant effect of it either on tumor growth inhibition or survival, and the results were very similar to the saline control [33].

Figure 2: Effect of different treatments on tumor regression, body weight change, and survival in an intraosseous model of prostate cancer bone metastasis:

Following tumor induction, animals were treated intravenously every 6 weeks as indicated by the arrows on X-axis either with TXT-NPs, DNmb-NPs or combination of TXT-NPs+ DNmb-NPs (TXT dose =12 mg/kg; DNmb dose =3 mg/kg). The arrow on Y-axis in Fig 1A indicates the background signal from normal animals (not inoculated with tumor cells) but received Luciferin injection. Change in A) Bioluminescence signal, B) Body weight, and C) Survival with different treatments. D) Representative bioluminescence images at different time points post-treatment, showing tumor growth. Combination treatment (TXT-NPs + DNmb-NPs) demonstrated complete tumor regression, treated animals gained weight, and there was no mortality until the study endpoint at 40 weeks. Data represented as mean ± s.e.m. (n=4).

The combination treatment (TXT-NPs + DNmb-NPs) was the most effective, regressing the tumor completely. Based on the bioluminescence signal, there was no tumor signal after 33.3 ± 6.4 weeks. The treated animals also gained weight and showed 100% survival until the study endpoint at ~40 weeks (Figure 2). The combination treatment with equivalent doses of drugs in solutions (TXT-CrEL + DNmb sol) showed a steady increase in tumor growth (Supplemental Data 2), indicating a better efficacy of the combination treatment with the encapsulated therapeutics.

The micro-CT analysis showed the loss of cortical bone of tumor-induced tibiae in saline control and DNmb-NPs and TXT-NPs treated animals but not in combination-treated animals, and they looked similar to the contralateral tibia (without tumor) or normal animal tibia (without tumor or drug treatment) (Figure 3). These findings correspond with the bioluminescence signals in the respective treatment groups (Figures 2A and D).

Figure 3: Micro-CT analysis for bone loss.

Micro-CT images of the tumor tibia from representative treatment groups. The results show that the combination treatment (TXT-NPs+ DNmb-NPs) inhibited bone resorption, whereas the other groups showed bone resorption.

Quantitative analysis for the bone mineral content and bone volume fraction showed loss of bone in the tumor tibia from saline control and DNmb-NPs and TXT-NPs treated animals (Table 2). Interestingly, the contralateral tibiae from these groups also showed lower bone mineral content than in normal animal tibia. In combination-treated animals, however, the bone mineral contents and bone volume fraction in the tumor tibia were very close to that in normal animal tibia (Table 2). Overall, the data demonstrate that the combination treatment not only prevented bone resorption in the tumor tibia but also the loss of mineral content in the contralateral bones.

Table 2:

Effect of different treatments on bone characteristics of the tibia.

Treatment Group	Sample	Bone Mineral Content (mg)	Bone Mineral Density (mg/cc)	Bone Volume Fraction (Tumor Leg v/s Non- tumor Leg)
Normal	Right Leg	18.8	696	1.00
	Left Leg	18.6	681
Saline control	Tumor Leg	02.7	610	0.27
	Non-tumor Leg	12.1	679
DNmb-NPs	Tumor Leg	02.8	659	0.21
	Non-tumor Leg	14.4	731
TXT-NPs	Tumor Leg	04.3	647	0.36
	Non-tumor Leg	13.1	711
TXT-NPs + DNmb-NPs	Tumor Leg	19.8	736	1.15
	Non-tumor Leg	18.7	774

Our previous study with paclitaxel-loaded NPs (PTX-NPs) in the same prostate cancer intraosseous model demonstrated slowing down of tumor progression with respect to untreated control, but even with a weekly dosing, the tumor did not regress, and animals showed weight loss, indicating toxicity [19]. In the subsequent study with TXT-NPs, we accelerated the drug release profile (~1 %/day vs. ~10% /day, based on the one-week release profile) using DMT as a plasticizer/pore-forming agent. The treatment with TXT-NPs alone in our previous study with once-in-4-week dosing demonstrated better tumor regression and survival than with PTX-NPs and significantly better than with TXT-CrEL, indicating that encapsulation and accelerated release improved the therapeutic outcome. Thus, in addition to size and zeta potential, modulating the drug release profile from NPs is an important parameter so that the drug dose released is effective in causing tumor regression [33]. However, as shown in this and previous studies [33], despite improved outcomes with TXT-NPs, after the initial regression, the tumor relapsed and developed resistance (Figure 2A).

For the combination treatment, previously, we tested TXT-NPs and DNmb-solution administered every 4 weeks. The combination treatment demonstrated tumor regression and no bone loss, and histochemical analysis of the tumor tibia indicated normal bone morphology and osteoblast and osteoclast cell activities [33]. The treatment with DNmb-solution alone was not effective, suggesting that the combination treatment produced a synergistic effect. Since DNmb has a long circulation half-life (15 days at 60 mg dose, 26 days at 120 mg dose in humans) [34], it may be bioavailable at the tumor site in the bone marrow along with the TXT released from the tumor-localized TXT-NPs to cause the synergistic effect. However, with the combination treatment, there was animal mortality at a later stage (20% at 32 weeks post-treatment). In addition, DNmb administered as a solution in humans is associated with serious toxicities, including osteonecrosis of the jaw, hypocalcemia, and atypical femoral fracture events, with multiple vertebral fractures, which required treatment discontinuation [15].

In this study, we encapsulated both TXT- and DNmb into separate NPs with similar properties, with the idea that both the drugs will have similar pharmacokinetics of drug distribution and localization in the tumor tissue, whereas administering one drug (TXT) in NPs and other as a solution (DNmb) might have different pharmacokinetics and bioavailability. In this study, the combination treatment was given every 6 weeks instead of every 4 weeks used in the previous study to minimize the risk of toxicity. Despite a lower frequency of dosing, the combination treatment in NPs resulted in complete tumor regression, and there was no bone loss or mortality until the study endpoint at 40 weeks (Figure 2). Although encapsulating both DNmb and TXT in NPs provides better therapeutic efficacy, the alternative of administering TXT-NPs and DNmb-solution in combination has the advantage, considering that DNmb-solution is an approved product and is given to patients with bone metastasis to minimize SREs [15].

3.3. TXT biodistribution and biocompatibility:

The mean TXT level in the tumor tibiae with TXT-NPs was 11.5-fold higher than with TXT-CrEL and 4-fold higher than in the contralateral (non-tumor) tibiae (Figure 4). The difference in the TXT levels in the tumor tibia with TXT-NPs than in the contralateral non-tumor tibia could be due to overexpression of multiple proangiogenic growth factors forming vascular niches that could open the cleft in the endothelial lining of the sinusoidal capillaries, resulting in greater localization of the circulating NPs in the tumor tibia than in non-tumor tibia [35]. In addition, the NPs are large enough to minimize their passage through fenestrations (~75 nm) in the liver’s sinusoidal endothelial cell lining to reduce hepatic uptake and clearance [36]. The data show that the ratio of TXT levels in the tumor tibia tissue to that in the liver tissue on a per-gram weight basis is ~ 11.6-fold (Figure 4). PEGlylation is commonly used to prolong the systemic circulation time of NPs, but PEG, due to its steric hindrance, can interfere with the extravasation of these NPs to the bone marrow [37]. With TXT-NPs, the TXT level in the prostate was ~3.7-fold higher than with TXT-CrEL, which could be advantageous in treating localized prostate tumor that has metastasized to the bone. Overall, TXT-NPs demonstrated favorable drug biodistribution critical for enhanced therapeutic efficacy. The effect seen could also be due to the sustained release properties of NPs, thus retaining the drug longer than with TXT-CrEL. Following a single-dose administration as above, there were no changes in the ALT or AST levels in the liver and heart tissue (Figure 4B) as well as in the serum (Figure 4C) in either TXT-CrEL or TXT-NPs treatment group. The histological analysis of the vital organ tissue also did not show any sign of toxicity (Figure 4D). The biocompatibility data are important considering that in the biodistribution study, higher TXT levels were found in the lungs, heart, and kidney with TXT-NPs than with TXT-CrEL (Figure 4A). Although TXT-CrEL did not show any toxicity following a single dose, it is not effective in treating bone metastasis, as has been demonstrated in our previous study [33].

Figure 4: Biodistribution and biocompatibility of TXT with and without encapsulation in NPs.

A) Animals with intraosseous bone metastasis were administered intravenously either as TXT-NPs or TXT- CrEL (12 mg/kg), and the tissues were harvested one-week post-treatment for TXT analysis using LC-MS. The TXT level in the tumor tibia was 11.5-fold higher with TXT-NP than with TXT- CrEL and 4-fold higher than in the contralateral tibia (no tumor). ALT and AST ELISA assay of B) Tissues from the liver and heart and C) Serum show no significant change with TXT-NPs or TXT- CrEL treatment compared to those in the saline-treated animal. D) Histological analysis of lung, heart, liver, kidney, and spleen using hematoxylin and eosin staining. Entire tissue sections were checked at 20X magnification for tissue necrosis. (Scale bar = 50μm). Data as the mean ± s.e.m., n= 3, *p<0.05, ** p<0.005.

3.4. End-stage tissue analysis:

DNmb was detected in the tumor tibia of the animals treated with combination and DNmb-NPs alone, but not in saline control or TXT-NPs treated animals (Figure 5A). The tumor tibia of saline control, DNmb-NPs, and TXT-NPs showed RANKL but not in the combination treatment (Figure 5B). The inefficacy of DNmb-NPs alone treatment to impact tumor growth (Figure 2A) or prevent bone resorption (Figure 3) can be implicated due to the presence of RANKL in the tumor tibia. DNmb-NPs have no antiproliferative effect on cancer cells (Figure 2A–d); hence the tumor growth is not directly impacted by its treatment [38]. On the other hand, cancer cells in the bone marrow tumor environment can produce RANKL, which, like that produced by the bone cells, participates in the bone resorption process [39]. In other words, metastatic prostate tumor cells may be able to act as surrogate osteoblasts and circumvent the necessity for osteoblast-derived RANKL by interacting directly with the immature osteoclast cells in order to mediate localized osteolysis [40]. In addition, cancer cells in the marrow can adapt to the RANKL-independent pathways [41] by developing dependence on other growth factors in the bone marrow to sustain the growth and subsequently produce RANKL [4]. In TXT-NPs alone treated animals, the relapsed tumor was resistant; and this treatment group also showed RANKL in the tumor tibia (Figure 5B), indicating that the RANKL could have modulated the response of anticancer drugs to develop resistance to sustain tumor growth. In fact, it has been reported that RANKL expressed by the bone marrow stromal cells contributes to anti-cancer drug resistance by activating various signaling molecules that are tumorigenic [42]. Drug resistance to TXT is also considered one of the main factors responsible for its failure in treating prostate cancer [43].

Figure 5: End-stage tissue analysis for DNmb, RANKL, inflammatory cytokines, and liver ALT/AST levels.

Tissue samples were homogenized and analyzed using ELISA. Analysis of A) DNmb and B) RANKL in tibiae. C & D) Analysis of inflammatory cytokines (TNF-alpha and IL-6) in vital organ tissue, and E) Liver tissue ALT/AST levels. DNmb was detected in the tibia of the animals that received DNmb-NPs or the combination (DNmb-NPs + TXT-NPs). RANKL was detected in the tumor tibia in the animals treated with DNmb-NPs, TXT-NPs, or saline control but not in the tumor tibia of the animals treated with the combination. There were no changes in the inflammatory cytokines and the liver tissue ALT and AST levels in different treatment groups, and were similar to that found in the normal tibia. Normal = animal without tumor or treatment; T= tumor tibia; N = contralateral (nontumor) tibia. Data as the mean ± s.e.m., *p<0.05, ** p<0.005, *** p<0.001.

Relatively lower bone mineral content seen in the contralateral tibia (non-tumor tibia) in saline control, and TXT-NPs, and DNmb-NPs treated animals than in normal animal tibia could have been due to the effect of RANKL produced in tumor tibia, leaching into the systemic circulation and then acting non-specifically onto the contralateral tibia, perhaps also on other bones in the body [44]. RANKL belongs to the TNF cytokine superfamily and plays a role in osteoporosis of the bone [45]. Since there was no RANKL in the tumor tibia in the combination treatment, bone mineral content in the contralateral tibia in this treatment group was very similar to that in the normal animal tibia (Table 2). Hence, inhibiting RANKL is key to preventing metastatic tumor progression and bone-associated complications [46]. Considering that multi-dose chronic treatment is needed to treat bone metastasis, we analyzed the end-stage tissue for each treatment (Figure 5). The animals in the combination treatment group gained weight (Figure 2B), there was no increase in inflammatory cytokines in the vital organs (Figures 5C and D) or the liver tissue ALT/AST levels (Figure 5E), and it also had no non-specific effect on the bones (Table 2), indicating that the combination treatment was not only effective but well tolerated.

3.5. Efficacy in hematogenous metastatic model (induced via intra-cardiac of PC3 Luc cells):

At an advanced stage of the disease, in addition to the bone, tumor often metastasizes or is already metastasized to soft tissue. Therefore, we tested the efficacy of the treatment in the hematogenous metastatic model. Based on the bioluminescence signal, saline control showed an increase in tumor burden, rapid loss of body weight, and increased mortality, whereas the combination-treated animals showed a significant delay in tumor progression compared to saline control, there was no body weight loss, and the treated demonstrated significantly improved survival than saline control (median survival saline 4.5 vs. 18 weeks, p <0.05) (Figure 6). Considering the aggressive nature of the hematogenous metastasis model, the results are encouraging, and further optimization of the treatment dose and dosing frequency might achieve tumor regression. Nonetheless, the data from the two model studies suggest that combination treatment could be used to treat advanced-stage cancer that has metastasized extensively, including to the bones. Wu et al., with encapsulated TXT in PLGA-NPs that were surface modified with hyaluronic acid, demonstrated 4.4-fold higher drug accumulation in tumor tissue than with free drug in an orthotopic model of human lung cancer. Further, the treatment with NPs demonstrated better tumor growth inhibition and improved survival than with free drug [47].

Figure 6: Effect of combination treatment in the hematogenous metastatic tumor model.

Hematogenous metastatic tumor model induced following cardiac injection of PC3-Luc cells were treated intravenously every 4 weeks with the combination treatment (TXT-NPs + DNmb-NPs; TXT dose =12 mg/kg, DNmb dose = 3 mg/kg). A) Effect of treatment on change in bioluminescence signal, body weight, and survival. B) Representative images from saline control and the combination-treated animals. The arrow in Fig 6A indicates the background signal from the animals not inoculated with tumor cells but injected with Luciferin solution. Combination treatment shows tumor growth inhibition and improved survival compared to saline control. Data as the mean ± s.e.m., n=4, Median survival = 18 wks vs. 4.5 wks, p=0.006.

With monotherapy (DNmb-NPs or TXT-NPs), there appear to be multiple pathways via which the tumor continues to grow, with RANKL modulating the bone marrow tumor environment (Figures 7 c & d). Its inhibition is key to enhancing the sensitivity of TXT to cancer cells to prevent drug resistance and relapse. In all, the NP-mediated favorable drug biodistribution to the tumor tibia and sustained drug effect of the combination treatment with TXT acting on cancer cells, inhibiting the formation of growth factors needed for the progression of osteoblast to osteoclast conversion and, at the same time, DNmb preventing osteoblast to osteoclast conversion to inhibit the production of growth factors needed for tumor progression. Based on the data, the combination treatment interrupted the destructive cycle of cancer cell-bone cell interaction by inhibiting the formation of RANKL (Figure 7e) to cause tumor regression. In all, the sustained releasing NPs played critical roles in achieving therapeutic efficacy since the combination therapy in solution was ineffective in regressing tumor growth (Supplement data 2).

Figure 7: Schematic depicting efficacy of combination treatment.

Following bone metastasis (a) tumor continues to grow with RANKL promoting cross-talk between cancer and the bone cells, supporting osteoblasts (bone cells) to transform into osteoclasts that cause bone lysis. The transformation of osteoblasts to osteoclasts produces growth factors that support cancer cell progression (b). Monotherapy with DNmb-NPs alone treatment is not effective in neutralizing RANKL that is produced by cancer cells and bone cells. DNmb does not have any effect of its own on cancer cell growth; hence tumor progression continues (c). TXT-NPs show tumor relapse as the bone marrow tumor microenvironment adapts and produces RANKL, desensitizing cancer cells to the cytotoxic effect of TXT to sustain tumor growth (d). In combination treatment, by impacting cancer cells and RANKL at the same time, interrupted the RANKL medicated cross-talk between cancer and the bone cells; in the absence of RANKL, enhanced the sensitivity of TXT to cancer cells to prevent tumor relapse and development of drug resistance, thus regressing tumor completely and preventing bone loss (e). NPs provided favorable biodistribution and sustained effects to modulate the tumor bone microenvironment to inhibit the production of RANKL (schematic created using BioRender.com).

Although prostate cancer is often osteoblastic (80%), we used the PC3 cell line that induces an osteolytic tumor. However, it is the most used cell line, isolated from the metastasized bone of a prostate cancer patient, to model an aggressive end-stage prostate cancer disease phenotype [48]. Our previous data in different prostate cancer cell lines in vitro has shown the efficacy of TXT-NPs [33]; hence, we anticipate that the combination treatment would be effective in different types of prostate cancer as well as in other cancers of bone metastasis as RANKL is considered as the key mediator of tumor progression. In this regard, RANKL serum levels could serve as a promising biomarker to further optimize the combination treatment [49].

The prior efforts, such as conjugating anticancer drugs or drug-loaded NPs to bone-seeking agents, such as bisphosphonates, tetracycline, or E-selectin (overexpressed in bone marrow endothelium), remained ineffective in regressing bone metastasized tumors, at the most, they slow down tumor growth [50] but do not prevent bone remodeling [51]. Our approach of delivering two drugs encapsulated in NPs and administered in combination is relatively simple yet effective and safe. It is possible to further improve our treatment efficacy by optimizing the dose and dosing frequency and demonstrating its effectiveness in different tumor models of bone metastasis.

According to the American Cancer Society, about 400,000 new cases of malignant bone metastasis are diagnosed each year, estimated >350,00 patients with bone metastasis dying each year in the United States [52]. The incidence of advanced malignant tumors with bone metastasis is 30–75%, more prevalent in patients with advanced prostate cancer and breast cancer [1]. Based on the autopsy results, about 80% of death in prostate cancer patients is due to bone metastasis [53]. Hence, there is an unmet clinical need to develop an effective therapy for treating advanced staged metastatic cancers. Our approach of delivering two drugs in combination in NPs could be effective yet simple for clinical translation. In this regard, a few nanomedicines for cancer therapy have been approved and a few more are at different stages of clinical trials for metastatic and advanced-stage cancers [54].

4. Conclusions

The combination treatment with encapsulated TXT and DNmb regressed the tumor completely in a prostate cancer model of bone metastasis, preventing bone resorption without causing any mortality. There was no increase in inflammatory cytokine levels in vital organ tissues, the liver ALT/AST levels, and non-specific effects on the bones, indicating the treatment’s safety. The histological analyses of vital organ tissue did not show any necrosis. The mechanism of efficacy is attributed to the synergistic effect of the combination treatment that inhibited RANKL formation in the bone tumor microenvironment. The combination treatment could benefit early on by minimizing SREs and associated morbidity due to inhibition of tumor progression. Since NPs are formulated with the FDA-approved biodegradable polymer, and both the drugs used are FDA-approved oncology drugs, the combination treatment has significant potential for clinical development as a therapy for treating advanced-stage cancers involving bone metastasis.

HIGHLIGHTS.

• Bone metastasis at an advanced disease stage is common and is untreatable.

• Cross-talk between cancer and bone cells drives tumor progression.

• Docetaxel and Denosumab encapsulated in nanoparticles regressed the tumor completely.

• The dual synergistic drug treatment prevented bone resorption.

• Encapsulation enhanced the therapeutic efficacy and safety of the treatment.

Abbreviations

ALT

Alanine Aminotransferase

AST

Aspartate Aminotransferase

FBS

Fetal Bovine Serum

CrEL

Cremophor EL

DNmb

Denosumab monoclonal antibody

ELISA

Enzyme-linked immunosorbent assay

HPLC

High-performance liquid chromatography

HSA

Human serum albumi

IC₅₀

Half concentration required for 50% cell growth inhibition

IL-6

Interleukin-6

LC-MS

Liquid chromatography-mass spectrometry

Luc

Luciferase

micro-CT

Micro-computed tomography

NPs

Nanoparticles

M.W.

Molecular Weigh

PC3-Luc

Prostate Cancer Cell Line expressing luciferase

PLGA

d,l-lactide co-glycolide

PTX

Paclitaxel

PVA

Poly (vinyl alcohol)

RANKL

Receptor activator of nuclear factor κB ligand

ROI

Region of interest

SREs

Skeletal-related events

SRM

Selective Reaction Monitoring

TNF-α

Tumor Necrosis Factor-Alpha

TXT

Docetaxel

UHPLC

Ultra-High-Performance Liquid Chromatography

VSMCs

Vascular Smooth Muscle Cells

Data and materials availability:

All data are available in the main text. The nanoparticle formulation described in the manuscript will be made available to investigators under the material transfer agreement for research purposes.