1.

Introduction

The enzyme, cyclooxygenase-2 (COX-2), is overexpressed in a wide variety of malignant and premalignant neoplasms, contributing to both progression and metastasis.^1–3 COX-2 expression has been particularly well-documented in colorectal cancer (CRC) with an average of 86% of malignancies exhibiting detectable levels of COX-2 in multiple studies.^4–8 Among benign and premalignant colonic lesions, COX-2 expression levels range from 12.5% in hyperplastic polyps to 90% in adenomas.^4–11 In addition, COX-2 expression increases as the lesion progresses, suggesting that it can serve as a biomarker of early CRC. Thus COX-2 is a promising target for the development of imaging agents to enable the detection of premalignant and malignant lesions of the colon. To this end, we developed fluorocoxib A (FA) [Fig. 1(a)], a potent and selective fluorescent COX-2 inhibitor comprising the nonsteroidal anti-inflammatory drug indomethacin conjugated to the fluorophore, carboxy-X-rhodamine (ROX), through a four-carbon linker.^12–18 We hypothesize that FA-mediated fluorescence colonoscopy will enable COX-2-targeted visualization of adenomas of the colon. In earlier preclinical studies, FA enabled visualization of a wide range of COX-2-expressing neoplasms, including intestinal polyps in ${\mathrm{Apc}}^{\text{Min}}$ mice, 1483 human head and neck squamous cell carcinoma tumor xenografts in nude mice,¹⁵ nonmelanoma skin cancer in mice,¹⁴ and head-and-neck cancer and colorectal and bladder cancer in dogs.^12,13 These results suggest that the clinical development of FA has the potential to substantially improve the early detection of neoplasms including CRC. However, development has been hampered by FA’s lack of aqueous solubility. To address this issue, we now report the Pluronic^® formulation of FA and describe its utility in COX-2-targeted colonoscopic optical imaging of azoxymethane/dextran sodium sulfate (AOM/DSS)-induced adenomas of the colon in B6;129 mice.

Fig. 1

(a) Chemical structure of FA and FNC. (b) Pluronic F127 polymeric micellar nanoparticle formulation of FA or FNC in ${\mathrm{CHCl}}_3$ and DPBS following gentle stirring at 25°C for 16 h in the dark.

2.

Materials and Methods

Previously, FA has been solubilized using solvent systems that are not approved by the FDA for clinical applications.^15,19 Thus we selected Pluronic, an FDA-approved organic nonionic triblock copolymer, consisting of a central hydrophobic block of polypropylene glycol (PPG) edged by two hydrophilic blocks of polyethylene glycol (PEG), for the formulation of FA. Pluronic polymer self-assembles into micellar nanoparticles in aqueous systems, enabling solubilization of hydrophobic organic compounds, such as FA, by encapsulating them inside the hydrophobic core surrounded by the hydrated corona.

3.

Results and Discussion

We formulated and determined the photophysical properties of FA-NPs and FNC-NPs in aqueous solutions that resulted in complete retention of excitation and emission maxima of FA and FNC. FNC is an analog of FA with a tether between the indomethacin and ROX moieties that is 2-carbons long rather than the 4-carbon tether in FA. This shorter tether renders FNC unable to inhibit COX-2. In addition, we determined the physicochemical properties including hydrodynamic diameter and zeta potentials. The data showed that the hydrodynamic diameter ($D_h$) of FA-NPs and FNC-NPs were $45.7\pm 2.5\mathrm{nm}$ and $44.9\pm 3.8\mathrm{nm}$, respectively, and the zeta potentials ($\zeta$) were $-1.47\pm 0.3\mathrm{mV}$ and $-1.64\pm 0.5\mathrm{mV}$, respectively. These results suggest that both FA-NPs and FNC-NPs exhibit properties reported to be favorable for tissue penetration and cellular internalization in living animals.¹⁸

We used a colorectal adenoma mouse model²¹ in B6;129 mice by treating them with AOM followed by DSS. In this model, adenomas are typically observed at 8 weeks after AOM administration. The tumors appeared to be predominantly oval with an average length of 1 to 2 mm and width of 0.5 to 1.0 mm.

FA-NP or FNC-NP was injected intravenously and 1 h later mice were lightly anesthetized with 2% isoflurane, and fluorescence colonoscopy was performed using a Karl Storz Veterinary Colonoscope. The delivery of FA-NPs followed by the release of FA resulted in a high adenoma fluorescence compared with that of the surrounding normal colon tissues [Fig. 2(a), Figs. S1 and S2 in the Supplementary Material]. The SBRs were $>29$ ($n=4$, $p<0.0001$). In contrast, minimal fluorescence of adenomas was detected at 1 h postdosing of FNC-NPs in tumor-bearing mice. Similarly, minimal fluorescence was observed in animals treated with the COX-2 inhibitor, celecoxib, before administration of FA-NP [Fig. 2(a)]. We used ImageJ software for postacquisition image analysis [Fig. 2(b)]. The expression of COX-2 in the tumor tissues was confirmed by immunohistological analysis [Fig. 2(c)]. Histologically, the tumors were well-differentiated adenomas. Of interest, the COX-2 immunoreactivity was confined to the stroma of the adenoma, and staining was not observed in the normal colon. These results confirm that the FA delivery and tumor retention occurred by a COX-2-dependent mechanism.

Fig. 2

(a) Representative white light and corresponding fluorescence colonoscopy images (static) of B6;129 mice bearing either colorectal adenomas or normal colon injected with Pluronic micellar nanoparticles of fluorocoxib A (FA-NPs, 0.5 mg/kg) or Pluronic micellar nanoparticles of fluorocoxib negative control (FNC-NPs, 0.5 mg/kg) via tail vein injections. In the COX-2 blocking control experiment, tumor bearing B6;129 mice received FA-NPs (0.5 mg/kg) 1 h after an intraperitoneal injection of celecoxib (10 mg/kg) to block binding of FA to the COX-2 active site. (b) ImageJ software was used for quantification of fluorescence intensity at the ROI of colonic adenomas versus adjacent normal colon epithelium (N.S., not significant). (c) COX-2 immunohistochemical staining of well-differentiated adenoma and adjacent normal colon (Video 1, MOV, 31.9 MB [URL: https://doi.org/10.1117/1.JBO.28.4.040501.s1]).

White light colonoscopy has revolutionized our understanding and ability to characterize colorectal neoplasms in appreciable detail. It also provides a highly effective means to screen, diagnose, and treat preneoplastic and neoplastic colonic lesions. It has been hypothesized that fluorescence colonoscopy using a specific probe targeted to a CRC biomarker can improve the detection of preneoplastic colonic lesions. Past efforts to test this hypothesis have included the development of optical imaging agents targeted to a range of CRC biomarkers.²⁵ Currently, peptide-based compounds²⁶ and antibody-based agents²⁷ are still at the preclinical stages. The present studies with targeted nanoparticles set the framework for endoscopic molecular imaging of COX-2 as a biomarker of CRC in animal models and will facilitate the development of this nanotechnology for early detection colorectal neoplasm in patients.

4.

Conclusions

In this letter, we described FA-NPs, Pluronic polymer-based micellar nanoparticles of FA, and demonstrated their COX-2-dependent delivery of FA to colonic neoplasms, thereby enabling FA-mediated colonoscopic optical imaging of adenomas in AOM/DSS-treated B6;129 mice. The COX-2-dependent delivery of FA within the colorectal adenomas was confirmed by the presence of very low fluorescence signal in adenomas from animals injected with nontargeted FNC-NPs or animals treated with celecoxib prior to injection with FA-NPs. This study establishes a framework for clinical development/deployment of FA-NPs for colonoscopic fluorescence imaging of COX-2 as a biomarker for the early detection of colorectal cancer.