New tools for the quantitative assessment of prodrug delivery and neurotoxicity

Abstract

Systemic off-target toxicities, including neurotoxicity, are prevalent side effects in cancer patients treated with a number of otherwise highly efficacious anticancer drugs. In the current study, we have: (1) developed a new analytical metric for the in vivo preclinical assessment of systemic toxicities/neurotoxicity of new drugs and delivery systems; and (2) evaluated, in mice, the in vivo efficacy and toxicity of a versatile and modular NanoDendron (ND) drug delivery and imaging platform that we recently developed. Our paclitaxel-carrying ND prodrug, ND(PXL), is activated following proteolytic cleavage by MMP9, resulting in localized cytotoxic chemotherapy. Using click chemistry, we combined ND(PXL) with a traceable beacon, ND(PB), yielding ND(PXL)-ND(PB) that functions as a theranostic compound. In vivo fluorescence FRET imaging of this theranostic platform was used to confirm localized delivery to tumors and to assess the efficiency of drug delivery to tumors, achieving 25-30% activation in the tumors of an immunocompetent mouse model of breast cancer. In this model, ND-drug exhibited anti-tumor efficacy comparable to nab-paclitaxel, a clinical formulation. In addition, we combined neurobehavioral metrics of nociception and sensorimotor performance of individual mice to develop a novel composite toxicity score that reveals and quantifies peripheral neurotoxicity, a debilitating long-term systemic toxicity of paclitaxel therapy. Importantly, mice treated with nab-paclitaxel developed changes in behavioral metrics with significantly higher toxicity scores indicative of peripheral neuropathy, while mice treated with ND(PXL) showed no significant changes in behavioral responses or toxicity score. Our ND formulation was designed to be readily adaptable to incorporate different drugs, imaging modalities and/or targeting motifs. This formulation has significant potential for preclinical and clinical tools across multiple disease states. The studies presented here report a novel toxicity score for assessing peripheral neuropathy and demonstrate that our targeted, theranostic NDs are safe and effective, providing localized tumor delivery of a chemotherapeutic and with reduced common neurotoxic side-effects.

Keywords: Neuropathy; Protease-mediated drug delivery; Protease-mediated imaging; Theranostic paclitaxel; Toxicity assessment.

Figure 1

Schematic representations of the NanoDendrons (NDs) structures and cleavage (left side) and mechanism of action (right side), illustrating NDs within the vasculature and extravasated into tumor tissue, an MMP9-rich microenvironment. The proteolytic beacon, ND^PB, includes both reference (R) and sensor (S) fluorophors; the prodrug,, ND^PXL, carries PXL (D) attached to the MMP9-cleavable peptide (shown in heavy black).

Figure 2

Imaging of ND activation in vivo for calibration (A) and as a drug delivery surrogate (B-D). In each image (A-C), green, red and yellow represent reference, sensor and co-localization respectively. (A) In vivo calibration of ND^PXL via subcutaneous matrigel phantoms. In this mouse, the phantoms include: PAMAM-AF750 (23 nM, 1), PAMAM-AF700 (10nM, 2), ND^PB (40 nM, 3), MMP9-treated ND^PB (40 nM, 4) and PAMAM-AF750 (7 nM, 5). (B&C) Representative images for measuring ND^PB activation in FVB mice with R221A, (B) and nu/nu mice with MDA-MB-231, (C) orthotopic mammary tumors. (D) Quantification of ND^PB activation used to estimate PXL delivery to the tumor.

Figure 3

Efficacy of ND^PXL-ND and nab-PXL (Abx) versus vehicle in two pre-clinical models of orthotopic mammary cancer; an immunocompetent model of FVB mice with R221A tumors (panels A to D) and a human mouse model of nu/nu mice with MDA-MB-231 tumors (panels E & F). Panels A & E: White light images of typical mice overlayed with either bioluminescence from luciferin in FVB mice (panel A) or GFP fluorescence from tumors in nu/nu mice (panel E), each before treatment and after a 12 day cycle with either ND^PXL-ND, nab-PXL (Abx), each 12.5 mg/kg, or vehicle. Panels B & C: Tumor volume (median with max and min denoted by the vertical bar, panel B) measured with calipers for mice (n =3 or 4 per cohort) treated with either ND^PXL-ND or Abx (either 12.5 or 4.1 mg/kg as indicated) or vehicle; individual tumor volumes at day 12 of treatment (Panel C). Panels D & F: Tumor volume (mean+/−SEM) measured with calipers for mice treated with either ND^PXL-ND (Red), Abx (Blue), each 12.5 mg/kg, or vehicle (black); panel D, R221A tumor model (n = 20 or 22 mice per cohort, Tables S6 &S7); panel F, MDA-MD-231 tumor growth (8 to 15 mice/cohort, Tables S4 & S5) was normalized to the individual mouse tumor volume on the first day of treatment.

Figure 4

Behavioral tests of mice hosting the MDA-MB-231-GFP (nu/nu) model of human breast cancer revealed altered behavior in mice treated with nab-PXL (Abx) but not those treated with ND^PXL-ND or vehicle, indicating neurotoxic effects of Abx but not ND^PXL-ND. Panel A: Hanging wire test results during drug-treatment indicate a trend towards poorer performance in Abx-treated mice. The dotted line and gray bar are the average and SEM, resp., for vehicle-treated mice. Panel B: The toepinch test scored for each test in all mice shows a larger number of low scores in Abx-treated versus vehicle-treated mice, i.e., a poor response not observed in the ND^PXL-treated mice, indicating reduced sensitivity in Abx-treated mice during the treatment cycle (left side) and throughout the whole study (right side). The gray bar identifies the range of scores defined as normal behavior for statistical analysis.

Figure 5

Representative behavior tools to assess and compare the PXL-induced peripheral neuropathy of ND^PXL-ND and nab-PXL (Abx) in FVB mice. Sensory reflexes were scored for a combined total reflex score (panel A), and sensitivity to a toe pinch (panel B); motor coordination was assessed by the pole climb test (panel C). Each graph represents the mean +/− SEM of either all measurements acquired in a given study period (panels A & B) or of individual mouse performance for each day (panel C).

Figure 6

Calculated toxicity scores for non-tumor FVB (left) and nu/nu tumorbearing (right) mice treated with either ND^PXL-ND or nab-PXL (Abx) versus vehicle. The toxicity score for each mouse was calculated from several tests (scaled 0-100) in which higher scores reflect poorer performance indicative of toxicity; the mean (+/− SEM) of all mice in each cohort are denoted by horizontal bars. **, p<0.05; ***, p<0.005.

Figure 7

Efficacy and toxicity comparison in nu/nu tumor-bearing mice treated with either ND^PXL-ND (open circles) or nab-PXL (solid squares). Each dot represents a mouse from day 12 of the study. The rectangle denotes the region with acceptable efficacy (> 60%) but low toxicity score (< 25).