Patients with disseminated metastatic disease from breast cancer are likely to have liver involvement in >50% of cases at some point during disease progression. These patients have a poor prognosis; and, when treated with the standard of care systemic therapy they have a median survival of <9-months. Increasing survival in breast cancer patients will likely require the administration of better therapies that are specifically targeted to treat distant metastases. One approach to increasing treatment efficacy for breast cancer liver metastases is through the application locoregional therapies. Locoregional therapies are an appealing interventional approach for breast cancer patients with liver metastases since these tumor lesions are accessible via minimally invasive procedures that can be administered using either ultrasound or CT imaging. Current locoregional therapies to treat breast cancer liver metastases are non-specific and have not produced significant increases in survival. The goal of this study was to design and test a targeted locoregional therapeutic intervention for breast cancer liver metastases. The lead candidate, a fixed-dose small-molecule drug called MBC-005, was tested in vitro and then the efficacy was evaluated in a BALB/c mouse liver metastases model. A novel formulation of N-allyl noroxymorphone hydrochloride incorporated into an alginate-based gel overcomes many of the limitations associated with the administration of small-molecule drugs, which include solubility, off-target toxicity, and enzymatic degradation. In vitro results demonstrated that MBC-005 mediated its anti-tumorigenic effect through a p21-dependent mechanism via a novel molecular pathway, in which N-allyl noroxymorphone component of MBC-005 stimulated the opioid growth factor receptor to increase p21 expression. Intratumoral administration of MBC-005 increased survival 3.9-fold in mice and significantly decreased tumor volume 4-fold. While many cytotoxic therapies increase p21 expression as a response to DNA damage, MBC-005 increased p21 expression independent cytotoxic DNA damage. MBC-005 did not induce off-target toxicity; and, as such, would be amenable to multiple rounds of administration. Nevertheless, it is notable that the positive effects of MBC-005 treatment on increasing survival and decreasing tumor volume in mice was achieved using a single dose.

Material Studies

A modified USP apparatus 2 was fitted with in-situ UV fiber optic probes that was used to perform the dissolution evaluation. Solubility was assessed in 10 dissolution media (e.g., buffers with and without surfactants) to determine the ideal ‘sink’ conditions for dissolution. UV wavelengths were scanned to identify the spectra and the lambda max. The limit of detection and limit of quantification were identified, and a dissolution profile was determined. Each method was run in triplicate over 72-hours using fiber optic dip probes.

Viscosity was assessed using the SV-10 tuning-fork sine wave viscometer (A&D Weighing). Samples were prepared and mixed for 2-minutes, allowed to sit for 5-minutes, then briefly mixed a second time prior to testing. The tuning-fork sensor plates were inserted into MBC-005 and viscosity and temperature data were collected at 5-second intervals.

In vitro Studies

Primary human mammary epithelial cells (HUMEC) were used as a non-transformed control cell line. HUMEC’s were maintained in Leibovitz’s L-15 media (Lonza), without serum but supplemented with the Mammary Epithelial Cell Growth Kit (ATCC). The MCF7 human breast cancer tumor cell-line (HR+/HER2-) is derived from a metastatic adenocarcinoma that possesses both estrogen and progesterone receptors (e.g., hormone receptor positive). The BT474 human breast cancer tumor cell-line is a ductal carcinoma derived from a solid mass (HR+/HER2+). MCF7 and BT474 cells were maintained in Dulbecco's Modified Eagle Medium (DMEM) containing 10% fetal calf serum (FCS, v/v) and 1% penicillin–streptomycin (Cellgro, Corning Life Sciences) with 1% glutamine (Glutagro, Corning Life Sciences). All cells used in this study were obtained from ATCC.

Cell Viability Assays

Cells were treated with serial doses of our formulation of N-allyl noroxymorphone dihydrate (Mallinckrodt Pharmaceuticals) solubilized in acetic acid, citric acid monohydrate, and 0.9% normal saline (Spectrum Chemical). Viable cell number was determined with the methyl tetrazolium (MTT) assay. After 72-hours of treatment, 5-mg/mL of the MTT reagent (w/v, Sigma) was added to each well and incubated for 2-hours, after which the cells were lysed with 500-µL of DMSO (Sigma). 50-µL of solution was added in duplicate to a 96-well plate and MTT absorbance was measured at 570-nm on a Tecan Spark monochromator. The effects on cell proliferation were determined by normalizing treated wells relative to mean values from non-treated wells, as follows: Fold change in cell number = 100*[treated cells optical density/mean control optical density].

The CyQuant lactate dehydrogenase cytotoxicity assay (LDH; Invitrogen) was used to determine cell death following treatment with N-allyl noroxymorphone. Briefly, at the appropriate time points after treatment with N-allyl noroxymorphone, 50-µL of cell culture media from each experimental plate was transferred to a 96-well plate, in duplicate. 50-µL of Reaction Mixture was aliquoted into each sample well and mixed by gentle tapping. The plate was incubated for 30-minutes at 37ºC in the dark, and absorbance was immediately read at 490-nm and 680-nm. The background measured at 680-nm was subtracted from the 490-nm reading to determine LDH activity. Fold change in LDH was determined using the following equation: LDH Fold Change = 100*[treated cells optical density/ control well optical density].

Spheroid Assay

The spheroid assay was used to determine long terms effects of N-allyl noroxymorphone therapy on tumor cell survival. Tumor cells in regular DMEM (5x10⁴ per implant) were mixed with equal parts (v/v) of Matrigel (Corning Life Sciences), allowed to settle in an incubator for 30-minutes, scanned using an Epson 700 scanner to determine initial spheroid area, and then carefully covered with media and placed back in the cell incubator. At the conclusion of the experiment, cells were fixed in 70% EtOH and stained with 1% crystal violet and then scanned again. Area fractions were determined using FIJI (ImageJ), and the initial area fraction for each implant was subtracted from the area fraction at the conclusion of the experiment.

The p21 Assay

Cells were disassociated using a commercially available lysis buffer (Cell Signaling Technology), to which 1-mM of the protease inhibitor phenylmethylsulfonyl fluoride was added. Protein lysates were analyzed using the Pierce BCA assay to determine total protein concentrations (ThermoFisher). Total p21^Waf1/Cip1 protein concentrations were assayed using a commercially available colorimetric sandwich ELISA assay (Cell Signaling Technology). Data were compared to a standard curve supplied with the ELISA kit.

Mouse Xenograft Studies

The efficacy of MBC-005 in treating breast cancer metastasis in the liver was evaluated using a syngeneic mammary carcinoma model utilizing orthotopic implantation of 4T1-Luc2 cells into the intrahepatic space in 8-weeks old female BALB/c mice (N=96 total/ N=12 per group). A Cox Proportional Hazard Model was employed to assess the Hazard Ratio (HR) to determine if the following parameters contributes to animal survival: treatment (Control versus MBC-005), initial tumor volume, final tumor volume, median tumor volume, animal weight at the conclusion of the studies, metastases, bioluminescence analysis, and liver weight.

The 4T1-luc2 murine mammary gland tumor cell-line is an implantable triple-negative breast cancer that was obtained from Perkin-Elmer. The cell line was maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum, 2-mM glutamine, 100-units/ mL penicillin G sodium, 100-µg/mL streptomycin sulfate, and 25-µg/mL gentamicin. Cell-line bioluminescent intensity was verified by limiting dilution before and after cell implant.

For tumor implantation, the 4T1-luc2 tumor cells were resuspended in 50% Matrigel (Corning) at 4x10⁶ cells/ mL. Prior to undergoing intrahepatic implantation surgery, the animals were administered 0.15-mg/ kg buprenorphine. Each mouse was surgically injected intra-hepatically into the left lateral lobe with 2x10⁵ cells in a 50-μL suspension under isoflurane anesthesia. On day 7 after the implantation of the 4T1-luc2 tumor cells, mice were randomized into treatment groups and all animals except those the no treatment control group were administered with doxorubicin (5-mg/ kg) intravenously. Mice in the MBC-005 treatment groups were concurrently treated with 25-mL of MBC-005 (Table 1) via intra-hepatically injection into the left lateral lobe.

The surgeries, animal monitoring, and imaging were completed at Charles River Laboratory Morrisville (Morrisville, NC). Animals were maintained in barrier housing in a clean room with 12-hours light/ dark cycle as part of an IACUC approved study. Animal welfare considerations were observed during this study, including efforts to minimize suffering and distress, which included subcutaneous administration of 0.15-mg/ kg of buprenorphine as an anesthetic immediately following surgery and as needed subsequently. Animals were anesthetized via inhalation using 4% isoflurane to reach a surgical plane of anesthesia and then maintained with 1% to 2% isoflurane. Following tumor implantation and treatment, animals were monitored daily and weighed every 2-days. Animals were euthanized when they were observation have lost 30% of body weight or on three consecutive weight measurements lost 25% of body weight. Animals were also euthanized when they lost 15% of body weight and, in conjunction, had impaired hind limb function, had hindlimb paralysis, or were moribund. Animals were also euthanized when the tumor volume was greater than or equal to 1000-mm³. Animals that completed the study and survived to day 60 following treatment were also euthanized. Animals were euthanized using CO₂ asphyxiation followed by cervical dislocation, after which animals were bled via cardiac puncture for clinical chemistry and hematology assessments. Necropsies was performed for individual animal and livers were removed for histologic processing.

Liver Weight Measurements

Following gross necropsy, livers were collected. Liver weight was obtained before being fixed in 10% neutral buffered formalin. Liver weights were measured using a balance and corrected for the animal weight to produce a % liver weight.

Bioluminescence Imaging

Bioluminescent images were acquired on day 0 and then bi-weekly until the end of the study.  Luciferase activity was measured using the IVIS SpectrumCT (Perkin Elmer) equipped with a CCD camera, mounted on a light-tight specimen chamber.  On each imaging day, animals were injected with 0.22-µm of VivoGlo D-Luciferin substrate (150-mg/ kg; i.p. at 10-mL/kg, split across two injections) and placed in an anesthesia induction chamber (2.5% to 3.5% isoflurane in oxygen). Upon sedation, animals were placed in a ventral position in the imaging chamber, equipped with stage heated at physiological temperature, for image acquisition starting ten minutes after luciferin injection. Flux was quantified and reported as 10⁶ photons/ second (p/s). 

Histology

After weighing the livers, samples were placed in 10% neutral buffered formalin and allowed to fix for at least 72-hours. Livers were dehydrated using serial concentrations of ethanol (e.g., 30%, 50%, 70%, and 100%) and then toluene. Livers were then laid flat, embedded in paraffin, and then sectioned at 5-mm on a rotary microtome. Sections were then stained with hematoxylin and eosin. Livers were sectioned at three levels (L1, L2, and L3). One section was evaluated from each level, which were separated by 645-mm (e.g., L1 was 645-mm from L2 and L2 was 645-mm from L3). Liver sections were imaged at 4x using a Nikon Eclipse Ni microscope and a Nikon DS-Fi2 digital camera. Composite images were generated using a tiling method, in which images were taken periodically in a grid pattern and then assembled into a final image using Photoshop (Adobe).

Histomorphology

Histomorphometric analysis was used to determine tumor volume and the volume of necrotic tissue in the left lobe of the liver, in which the tumor and MBC-005 were implanted. Volumes for the whole liver lobe were determined using Fiji (v1.54f) using image thresholding. Tumor volumes and the volume of necrotic tissue were determined using QuPath (v0.5.1-arm64) pathology analysis software. Volumes were estimated based on the work of Cruz-Orive and the Cavalieri Principle. Briefly, the area fraction (A) for a particular level (e.g., A_L1) was multiplied by the distance (d) between sections located in adjacent levels (i.e., between L1 and L2), in which d=645-mm. The composite volume (V_t) for each parameter (e.g., total volume, tumor volume, of volume of necrosis) was calculated by summing the volumes for each level, such that: V_t = V_L1 + V_L2 + V_L3 = (A_L1*645-mm) + (A_L2*645-mm) + (A_L3*5-mm). The final term A_L3 is multiplied by 5-mm to account for thickness of the final section. The volume fraction (V_f) was calculated by dividing the tumor volume (V_t(tumor)) by the total liver volume (V_t(liver)). Likewise, the volume fraction of necrosis (V_f) was calculated by dividing the volume of necrosis (V_t(necrosis)) by the total liver volume (V_t(liver)).

Statistical Analyses

Means and standard deviations were calculated for the MTT assay data, the LDH assay data, and the spheroid assay data. The normalized dose-response data from the MTT assay were assessed using an ‘inhibitor concentration versus normalized dose response’ non-linear regression analysis. Dose-response regression lines were assessed for statistical differences between groups using an Extra-Sum-of-Squares F-test to determine if the groups were statistically different or, alternatively, could be modeled using a single regression line (a=0.05). LDH and the spheroid assay data were assessed for significant differences using either one-way ANOVA followed by correction for multiple comparisons with Dunnett’s test (a=0.05).

Bioluminescent data for the mouse studies were analyzed using a Malthusian exponential growth model was developed using non-linear least-squares regression to estimate the doubling time for the tumor volume, which served as a surrogate for local tumor growth. Regression lines were assessed for statistical differences between groups using an Extra-Sum-of-Squares F-test to determine if the groups were statistically different or, alternatively, could be modeled using a single regression line (a=0.05).

Histomorphometric analyses were first evaluated normality using the Shapiro-Wilks test. All of the groups were found to be normal or lognormal, and a one-way ANOVA was used to assess differences between treatment groups. Consistent with Glantz’s et al. Welch’s correction with an unpaired ‘t’ was used to identify differences between the NTC group and the MBC-005 treatment groups (a=0.05).

A Cox Proportional Hazard Regression Model was used to assess survival and to calculate the parameter estimates (b_i) and the hazard ratio (HR = exp(b_i)) for the following parameters: treatment (Control versus MBC-005), initial tumor volume, final tumor volume, median tumor volume, fracture incidence, the presence of metastases, and final body weight.  All data analyses were performed using GraphPad Prism version 9.4.0, GraphPad Software, San Diego, California USA.