Purpose: Tumor hypoxia, a hallmark of tumor progression, causes upregulation of genes related to tumor angiogenesis, survival, and metabolism, and promotes resistance to various cancer modalities. Upregulated VEGF expression during hypoxia produces excessive and dysfunctional tumor vasculature which further reduces oxygen delivery that generates an acidic pH. Acidic pH neutralizes weak acid chemotherapies used as standard care for breast cancer while lack of oxygen reduces effectiveness of radiation contributing to treatment resistance. Furthermore, tumor-associated macrophages (TAMs) become immobilized in hypoxic areas. TAM presence correlates with poor prognosis in breast cancer as they produce factors which promote dysfunctional angiogenesis creating an immunosuppressive environment that counteracts the use of immunotherapies. In our previous work, we demonstrated that intratumoral delivery of high dose (100 ng in 50 µl) of GM-CSF abolished angiogenesis via soluble VEGFR-1 expression by VEGF sequestration leading to a near anoxic tumor microenvironment (TME). Further, we showed that GM-CSF induced TAMs to re-polarize into a pro-inflammatory M1 phenotype leading to reduced anti-inflammatory IL-10. Because anti-angiogenic therapies have proven unsuccessful in breast cancer, and because inducing hypoxia only supports an aggressive tumor program, we hypothesized that small doses of GM-CSF delivered systemically and released into the TME may 1) induce small amounts of sVEGFR-1 to normalize vessels and rescue oxygen delivery and pH, and 2) drive tumor supportive anti-inflammatory TAMs to pro-inflammatory tumor-fighting (M1-like) phenotype to augment cytotoxic T cell killing. In our current study, we are optimizing methods to deliver GM-CSF systemically using FDA-approved poly(ethylene glycol)--poly(lactic-co-glycolic acid) (PEG-PLGA) size-specific nanoparticles. As proof of concept to show oxygen and pH recovery with GM-CSF as our cargo, we performed low dose intratumoral injections of rmGM-CSF or saline in an orthotopic PyMT mouse model and subjected these mice to EPR spectroscopy for in vivo determination of tumor oxygen and pH. We also began optimization of PEG-PLGA/GM-CSF nanoparticle synthesis and characterized these nanoparticles by size and zeta potential. Finally, we performed in vitro experiments on murine primary macrophages using empty PEG-PLGA nanoparticles for their ability to induce an inflammatory response from nanoparticles without GM-CSF. In summary, we hypothesize that targeting TAMs with PEG-PLGA delivered GM-CSF can overcome tumor hypoxia and pH in breast cancer.
PyMT breast cancer cells (ATCC Py8119) were cultured and 5x105 cells were orthotopically-implanted into an inguinal mammary fat pad of C57BL/6J mice. Once tumors were palpable, mice were treated with recombinant murine (rm)GM-CSF (5 or 15 ng) or isotonic saline (vehicle) three times a week for 2.5 weeks by intratumoral injection. After, intratumoral oxygen and pH were measured by Electron Paramagnetic Resonance (EPR) spectroscopy using a soluble trityl compound (p-TAM) sensitive to oxygen (pO2) and extracellular acidosis (pHe) and an L-band EPR spectrometer. PEG-PLGA diblock copolymer (50:50) was purchased from Polysciences. Nanoparticles were formed by nanoprecipitation followed by freeze drying in the presence of 5% mannitol and sterile filtered before being placed in culture. Particles were characterized by size on the NanoSight and by zeta potential on a ZetaSizer. Bone marrow was collected by flushing the femurs of C57Bl/6J mice. Macrophages were differentiated over 5 days using rmCSF1 (5 ng/mL), endotoxin-free FBS, endotoxin inhibitor polymyxin B (pB; 10 µg/mL) in low endotoxin RPMI-1640 media. Macrophages were then cultured in pB-free media with empty PEG-PLGA nanoparticles for 48 hrs to determine any inflammatory response elicited on the macrophages by the nanoparticles. Macrophages were dissolved in TRIzol reagent and total RNA was extracted and cDNA synthesized from 900 ng total RNA. Primers for classic murine macrophage markers (M1: Nos2 and Tnfα; M2: Arg1 and Il10) were used in qRT-PCR analysis and results were normalized to Rpl4 and Cap1 housekeeping RNAs.
Our recent pilot data suggests that low dose GM-CSF-treated tumors had increased oxygen and elevated pH compared to vehicle-treated tumors. Empty PEG-PLGA nanoparticles had a size distribution ranging from 31-282 nm with a mean particle size of 119 nm and zeta potential of -14.8 mV. GM-CSF-loaded PEG-PLGA nanoparticles had a size distribution ranging from 36-354 nm with a mean particle size of 116.5 nm and zeta potential of -15.0 mV. Finally, our empty PEG-PLGA nanoparticles do not elicit an inflammatory response on murine macrophages, in vitro.
Our data suggests that low dose GM-CSF-induced macrophage sVEGFR-1 expression is acting to stabilize vessels thereby improving tumoral vessel function and perfusion as concluded from observed recovery of oxygen and pH. We can also conclude that consistency and reproducibility is present in early stages of GM-CSF-loaded nanoparticle development as size and zeta potential are uniform between GM-CSF-loaded and unloaded batches. Also, we observe that PEG-PLGA alone will not elicit an unwanted immunogenic effect. References: Eubank et al, Immunity, 2004 and Eubank et al, Cancer Research, 2009.
Pushkar Saralkar– Graduate Research Assistant, West Virginia University, Morgantown, West Virginia
Andrey Bobko– Research Assistant Professor, West Virginia University
Valery Khramtsov– Director, In Vivo Multifunctional Magnetic Resonance center; Professor, Department of Biochemistry, West Virginia University
Werner Geldenhuys– Associate Professor of Pharmaceutical Sciences, West Virginia University, Morgantown, West Virginia
Tim Eubank– Associate Professor, School of Medicine, West Virginia University