PD-L1-Targeting Nanoparticles for the Treatment of Triple-Negative Breast Cancer: A Preclinical Model

Abstract

Triple-negative breast cancer (TNBC) is a highly aggressive form of breast cancer. Common treatments following surgical resection include PD-1-targeting checkpoint inhibitors (pembrolizumab), as 20% of tumors are PD-L1 positive with or without systemic chemotherapy. Over the last several years, our laboratory has developed nano-immune conjugates (NIC) in which hydrophobic chemotherapy drugs like paclitaxel (PTX) and SN38, the active metabolite of irinotecan, are made water soluble by formulating them into albumin-based nanoparticles (nab) that are hydrophobically linked to various IgG1 monoclonal antibodies, creating an antigen-targetable nano-immune conjugate. To date, we have successfully tested PTX containing NICs linked to either VEGF- or CD20-targeted antibodies in two phase I clinical trials against multiple relapsed ovarian/uterine cancer or non-Hodgkin's lymphoma, respectively. Herein, we describe a novel NIC created with either PTX or SN38 that is coated with anti-PD-L1-targeting antibodies for the treatment of a preclinical model of TNBC. In vitro testing suggests that the chemotherapy drug and antibody retain their toxicity and ligand binding capability in the context of the NIC. Furthermore, both the PTX and SN-38 NIC demonstrate superior anti-tumor efficacy relative to antibody and chemotherapy drugs alone in a PD-L1 + MDA-MB-231 human TNBC xenograft model, which could translate clinically to patients with TNBC.

Keywords: mouse xenograft model; nano-immune conjugate (NIC); programmed death ligand 1 (PD-L1); triple-negative breast cancer (TNBC).

Figure 1

Nanoparticle manufacturing schematic and physiochemical properties of nanoparticles and NIC formation; (a) is a schematic representation of the manufacturing process for nab-PTX and nab-SN38 made in Biorender (b) Nanoparticles were diluted to 100 µg/mL, and their size was determined with a zeta-sizer, and sizes are represented in nm. (c) HPLC was performed to measure the total amount of drug that was bound to the nanoparticles and 90–100% of the 10 mg/mL bound to albumin. (d) Nanoparticles were diluted in saline or simulated plasma from 20 to 100 µg/mL, and the size of the particles was determined with the zeta-sizer (left panel); nanoparticles were diluted in saline at 50 µg/mL, and their stability was determined over time (middle panel) and pH (right panel). (e) STI-3031 was incubated with the nanoparticles in increasing concentration (0–8 mg/mL), and zeta potential was determined with a zeta-sizer.

Figure 2

Antibody–Ligand Binding. (a) Cell line expression of PD-L1 was determined by flow cytometry using FITC anti-PDL1, alexa fluor 488-conjugated atezoluzumab, and STI-3031. Data are represented in counts of PD-L1-positive cells in triplicate, and the p-value was determined with non-parametric two-tailed Student’s t-test comparing cells stained with antibody relative to isotype control. (b) PDL1-positive cell lines (MDA-MB-231 (left panel) and A375-PDL1 (right panel)) were utilized to measure the ability of antibody only, nab-PTX, STI3031:PTX NIC, nab-SN38, and STI3031:SN38 NIC to bind cell-bound PD-L1 prior to staining with alexa fluor-conjugated ST-3031. Cells were run by flow cytometry, and data are represented as counts of PDL1+ cells. PD-L1 isotype control and STI-3031 only were utilized as negative and positive controls, respectively; p-values were determined using non-parametric Student’s t-test comparing cells stained with antibody relative to isotype control or conjugated STI-3031 only. (c) Competitive ELISA was employed to confirm antibodies, nab-PTX, STI3031:PTX NIC, nab-SN38, and STI3031:SN38 NIC to inhibit PD-1/PDL1 blocking. Buffer only and PD-1-neutralizing antibodies were used as negative and positive controls, respectively. Curves were compared by two-way ANOVA, and curves were found to be significantly different (p = <0.0001). (ns) not significant, (*) p = </= to 0.05, (**) p = </= to 0.01, (***) p = </= 0.001, (****) p = </= to 0.0001.

Figure 3

Toxicity of PTX and SN-38. MDA-MB-231, MCF-7, Sk-Br-3, A375 WT, and A375-PDL1 cells were utilized to determine toxicity of STI-301, nab-PTX, and STI3031:PTX NIC (a) and irinotecan, free SN-38, nab-SN38, and STI3031:SN38 NIC (b). Cells were exposed to drugs overnight, and proliferation was measured using DNA-bound EdU. Two-way ANOVA was utilized to determine the significance of curves in treatment groups for each cell type. Curves were significantly different (p = </= 0.00001). (c) IC50 values were calculated using non-linear regression in GraphPad Prism (version 10.4.1) and expressed in µg/mL.

Figure 4

In vivo testing of nab-PTX and nab-SN38 nanoparticles and PD-L1 antibody-coated NIC. First, nab-PTX NIC made with atezolizumab (atez:PTX NIC) was compared to NIC made with STI-3031 (STI3031:PTX NIC). (a) Tumor growth (left panel), tumor response (middle panel), and survival curves are shown (right panel). (b) The second in vivo experiment compared PD-L1-coated nab-PTX NIC made with 4 concentrations of antibody, 0.5 (STI-3031:PTX NIC 0.05), 1 (STI-3031:PTX NIC.1), 2 (STI-3031:PTX NIC.2), and 4 (STI-3031:PTX NIC.4) mg/mL. Tumor growth (left panel), tumor response (middle panel), and survival curves (right panel) are shown. (c) In the final in vivo experiment, PD-L1-coated nab-SN38 NIC was compared to nab-SN38, irinotecan, and antibody only. Tumor growth (left panel), tumor response (middle panel), and survival curves (right panel) are shown. Tumor volume was determined by (length × width2)/2. The line graphs represent the average tumor volumes from 5 mice/group. Tumor response is shown as % change from baseline = [((Volume Day 0) − (Volume day 14 or 21))/Volume Day 0] × 100. For nab-PTX, tumor response was calculated on day 14, and the nab-SN38 tumor responses were calculated on day 21. Kaplan–Meyer curves were generated to show survival. To determine the significance for tumor growth, a 2-way ANOVA was used (p = <0.0001). For tumor response, NIC made with 4 mg/mL were compared to each other group individually with non-parametric two-tailed student’s t-test. Mantel–Cox log-rank test was utilized to determine median survival significance between NIC made with 4 mg/mL to each other group individually. All statistics were performed in GraphPad Prism software; p-values are less than 0.0001 (****), less than 0.001 (***), less than 0.01 (**), and less than 0.05 (*), not significant (ns). Each in vivo experiment was performed 2 times for confirmation of results.