PLGA-nanoparticle preparation

PLGA-NPs were prepared following the nanoprecipitation technique with 40 mg of Resomer 502H (PLGA Mw = 7–17 kDa, 50:50 L:G ratio; Sigma-Aldrich, St. Louis, MO, USA) dissolved in 1 mL of acetone and added dropwise to 10 mL of 1.5% (w/v) polyvinyl alcohol (PVA; Sigma-Aldrich), stirring at 1000 rpm, using a syringe pump at 10 mL/hr. The nanoparticle (NP) suspension was left to stir overnight at 500 rpm to evaporate acetone. PLGA-NPs were purified by five rounds of centrifugation at 15,000×g and resuspension in 1 mL of DI water. Finally, the NP suspension was filtered through a 1 µm strainer to remove aggregates. Following purification, PLGA-NPs were suspended in DI water at 10 mg/mL and stored at 4 °C. For celecoxib-encapsulated PLGA-NPs (Cx-NPs), 200 mg of celecoxib (Sigma-Aldrich) was dissolved with 40 mg PLGA in 1 mL acetone. Rhodamine-encapsulated PLGA-NPs (Rhod-NPs) were prepared by adding 500 µL of rhodamine 6G (Acros Organics, Waltham, MA, USA), dissolved in ethanol at 20 mg/mL, to 40 mg PLGA in 1 mL acetone. Concentration was determined using NanoSight NS300 Nanoparticle Tracking Analysis (Malvern Panalytical, Malvern, Worcestershire, UK).

Preparation of poly(lactic acid) hydrogel

One gram of poly(d, l-lactide)-b-poly(ethylene glycol)-b-poly(d, l-lactide) (PLA-PEG-PLA; PolySciTech, West Lafayette, IN, USA) was dissolved in 5 mL of DI water by stirring at 4 °C for 72 h to prepare 20% (w/v) solution. To prepare PLA-gel with 5% (w/v) NPs (NP/PLA-gel), 200 µL were prepared at a time by centrifuging 1 mL of NP suspension at 15,000×g for 10 min, followed by removal of the supernatant and resuspension in 20% PLA-gel. Rheological data of the PLA-gel is publicly available, provided by the manufacturer [28].

Nanoparticle characterization

The average nanoparticle diameter, polydispersity index (PDI), and zeta potential were measured by dynamic light scattering (DLS) using a zeta-sizer (Malvern Panalytical). NPs were suspended in purified DI water at 5 mg/mL and filtered with a 1 µm strainer before measurements. All measurements were performed in triplicate.

SEM imaging

SEM images were captured by Quattro ESEM (Thermo Fisher Scientific, Waltham, MA, USA). NPs were imaged using conventional methods with high vacuum following sputter coating of lyophilized NPs with iridium. For the PLA-gel, the Environmental SEM method was employed in which the samples were imaged without drying or sputter coating and remained in the solution/gel state.

HPLC

High-performance liquid chromatography was performed using UltiMate 3000 UHPLC (Thermo Fisher Scientific) with Acquity UPLC HSS T3 column (1.8 µm, 2.1 × 150 mm; Waters, Milford, MA, USA). The samples were prepared by adding 500 µL of dichloromethane (DCM) to 500 µL Cx-NP, creating a two-phase solution, followed by vortexing for 30 s to rupture the NPs. Then, the top aqueous phase was removed and centrifuged to confirm the presence of PLGA. The DCM was evaporated in a fume hood and the remaining celecoxib was dissolved in 1 mL of a 75:25 methanol and water mixture. The extracted celecoxib samples were diluted five times with 0.1% formic acid in a 1:1 mix of water and acetonitrile, and 10 µL was loaded at a time with a flow rate of 0.30 mL/min and detection wavelength of 254 nm.

PLGA-NP uptake

Bone marrow-derived cells (BMDCs) were freshly isolated from female B6 mice and cultured in RPMI media containing 0.2 mg/mL Rhod-NPs. After 24 hr of incubation, the cells were filtered with a 20 µm strainer, fixed with 10% buffered formalin, and adhered onto a poly-lysine-coated glass slide using StatSpin CytoFuge 2 cytocentrifuge (Beckman Coulter, Brea, CA, USA) for 10 min at 1,200 rpm then counterstained with DAPI (Vector Laboratories, Newark, CA, USA). Immunofluorescence staining was performed following incubation with Rhod-NPs for 24 hr. The cells were stained for CD3, CD45, or Ly6G with Alexa 647-labeled antibody for 1 h at RT, followed by filtration, fixation, and transfer to a glass slide. The cells were then counterstained with DAPI and imaged using a Leica DM 2500 (Leica Biosystems, Nussloch, Germany) fluorescence microscope with a 40× objective.

Release profile of PLGA-NP from PLA-gel

Rhod-NPs were prepared and suspended at 5% in 20% PLA-gel (Rhod-NP/PLA-gel). Then, 200 µL of the Rhod-NP/PLA-gel was solidified in 1.5 mL tubes and incubated in 200 µL of 50 mM Tris–HCl buffer (pH 7.5) containing 4 µg/mL proteinase K at 37 °C (n = 5). PLA-gel loaded with NPs with no Rhod was prepared as a baseline control. The supernatant was collected every two days and supplemented with fresh buffer until the gel was completely degraded. The fluorescence intensity was measured using EnVision Multilabel Plate Reader (Perkin Elmer, Shelton, CT, USA) with excitation at 544 nm and emission at 579 nm. A standard curve was generated by measuring the fluorescence intensities of NP suspensions with known concentrations, ranging from 0.5 to 100 µg/mL, to determine the concentration of released NPs. NP concentration was converted to mass and plotted as the cumulative fractional release of the total Rhod-NP collected as the PLA-gel degraded.

Cell culture

The Py230 mouse breast cancer cell line carrying mCherry reporter gene was cultured in Ham’s F-12 K (Kaighn’s) Medium (Thermo Fisher) supplemented with 5% fetal bovine serum (FBS), 1% antibiotic–antimycotic (Thermo Fisher) and 0.1% MITO + Serum Extender (Corning, NY, USA) as described previously [13]. Primary BMDCs were isolated from the bone marrow of female B6 mice and suspended in RPMI medium [29]. Following the removal of erythrocytes, the cell suspension was cultured in RPMI supplemented with 5% FBS and 1% antibiotic–antimycotic in an uncoated 3 cm dish. All cells were cultured in a humidified incubator at 37 °C, 5% CO2.

Mouse model of breast cancer

C57BL/6J (B6; strain #000664) mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA) and maintained in a pathogen-free facility. Py230 cells at a density of 2 × 106 cells per 100 µL in a 1:1 mix of RPMI-1640 and Matrigel GFR (Corning) were injected into the abdominal mammary fat pad of 6-week-old female B6 mice. Tumor size was measured by caliper every 3 days and volume was calculated using the formula L × W2 × 0.5, where L is the long diameter and W is the short perpendicular diameter. A needle biopsy was performed once tumor volume reached approximately 120 mm3 using a vacuum-assisted device connected to a 20-gage needle. The needle was inserted 5 mm (length equivalent to approximately half the major axis) into the tumor horizontally and slowly removed once the pressure reached -70 kPa. Immediately following needle biopsy, 2 µL of 20% PLA-gel containing 5% NPs, with no payload or loaded with celecoxib or rhodamine, was slowly injected into the biopsy cavity using a nano-syringe (Hamilton Company, Reno, NV, USA) connected to a 33-gage blunt needle to ensure injection into the biopsy cavity. Following injection, the needle was held inside the cavity for 10 sec to ensure gelation of the PLA-gel. The mice were euthanized, and the tumors were resected 1 or 15 days later. To quantify mCherry expressing breast cancer cells, the lungs were digested with 62.5 μg/mL Liberase (MilliporeSigma, Burlington, MA, USA) and 500 units/mL DNase I (MilliporeSigma). The cell suspension was incubated with a Zombie Aqua viability kit (BioLegend, San Diego, CA, USA) and analyzed using an S1000 EON benchtop flow cytometer (Stratedigm, San Jose, CA, USA). For in vivo release, the amount of Rhod-NP equal to one intra-biopsy cavity injection (0.1 mg) was suspended in 100 µL of PBS and intravenously injected via the tail vein. Tumors and organs were resected 1 and 15 days after the injection and embedded in OCT to generate frozen sections at a 6 µm thickness. The sections were counterstained with DAPI and mounted with Vectashield Antifade Mounting Medium (Vector Laboratories). Images were visualized by Leica DM 2500 (Leica Biosystems) fluorescence microscope and scanned using ZEISS Axio Scan.Z1 Digital Slide Scanner (Carl Zeiss Microscopy, Oberkochen, Germany).

Immunohistochemistry

Immunohistochemistry was performed on FFPE sections (4 μm) as described previously [13] using antibodies listed in Supplementary Information. The slides were scanned using ZEISS Axio Scan.Z1 Digital Slide Scanner (Carl Zeiss Microscopy) with a 40× objective. Image data were acquired using ZEN lite software (Carl Zeiss Microscopy) and quantified using Image J (NIH, Bethesda, MD) with the Color Deconvolution plugin [30] (n = 10). Multi-color immunofluorescence staining was performed using FFPE sections (4 μm). Slides were incubated with a combination of the primary antibodies (CD45, CD31, E-cadherin, pan-cytokeratin (CK), and vimentin) and corresponding secondary antibodies. Slides were quenched with TrueVIEW Autofluorescence Quenching Kit (Vector Laboratories), counterstained with DAPI, and mounted with Vectashield Plus Antifade Mounting Medium (Vector Laboratories). Fluorescently stained slides were scanned using ZEISS Axio Scan.Z1 Digital Slide Scanner (Carl Zeiss Microscopy) with 40× objective. Image data were acquired using ZEN lite software (Carl Zeiss Microscopy). For quantification of E-cadherin+(green)/CK+(red) cells and vimentin+(red)/CK+(green) cells, the number of yellow pixels (green and red overlapping pixels) was counted and normalized with the number of blue pixels (nucleus) per field of view at a final magnification of 40× (n = 4), using Image J.

ELISA

Freshly isolated mouse Mϕ were treated with Cx-NPs or control NPs without payload (Ctrl-NPs) at a concentration of 0.5 mg/mL in a 96-well plate for 6 h followed by incubation with or without 1% wound fluid, isolated from biopsied Py230 mouse breast tumors, for 4 h. The resulting supernatants were subjected to ELISA assay to quantify the concentration of PGE2 following the vendor’s instructions (Cayman Chemical, Ann Arbor, MI, USA). PGE2 concentration was compared using one-way ANOVA. Data is expressed as pg per mL with mean ± SD (n = 3).

Wound fluid preparation

Gas-sterilized surgical sponge (Medtronic, Minneapolis, MN, USA) at a size of 7 × 5 × 2.5 mm was implanted into Py230 mouse breast tumor. The surgical sponge was collected 7 days later and centrifuged at 10,000 rpm on a spin column. The supernatants were used for assays at 1% final concentration.

Statistical analysis

Data obtained from experiments were statistically analyzed to provide 95% power for a test’s significance level of 0.05, as depicted by *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001, CI = 95%. Two group comparisons were done using the t test if the assumptions of normal distribution were appropriate. Otherwise, the Wilcoxon test was used. One-way ANOVA was used to compare multiple groups. All statistical analyses were performed using Prism 9 software (GraphPad Software, La Jolla, CA, USA) or R 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria).