In vivo performance evaluation of a transdermal near-infrared fluorescence resonance energy transfer affinity sensor for continuous glucose monitoring

The in vivo performance of a transdermal near-infrared fluorescence resonance energy transfer (FRET) affinity sensor was investigated in hairless rats, in order to validate its feasibility for glucose monitoring in humans. The sensor itself consists of a small hollow fiber implanted in dermal skin tissue, containing glucose-sensitive assay chemistry composed of agarose-immobilized Concanavalin A (ConA) and free dextran. The glucose-dependent fluorescence change is based on FRET between near-infrared-compatible donor and quencher dyes that are chemically linked to dextran and ConA, respectively. We conducted an acute in vivo evaluation of transdermal sensors with an optical fiber-coupled setup over 4 h, and a chronic in vivo evaluation of fully implanted sensors for up to 16 days. The fiber-coupled sensors followed trends of blood glucose concentrations very well with a delay of less than 5 min. The acute performance of the implanted sensors at the day of implantation was similar to that of the fiber-coupled sensors. After 2 weeks the implanted sensors remained functional, evidenced by an adequate correlation between sensor signal and changes in blood glucose excursions, but exhibited delays of approximately 10-15 min. Preliminary characterization of host response showed signs of mild inflammations around the implanted sensor, which were characterized by formation of a 10-20-mu m-thick collagen band, typical for capsule formation. An acute study of systemic ConA biotoxicity was also conducted. A histological analysis of various organs and of clinical chemistry data showed no significant differences between rats receiving intradermal injections of ConA at 10 times the concentration in the sensor and rats in a control group (injection of saline solution). The absence of a toxicological or systemic response to ConA at a 10-fold larger amount than in the sensor should dispel concerns over the in vivo safety of ConA-based sensors. This study clearly demonstrates the feasibility of the proposed transdermal FRET-based sensor interrogation concept for glucose monitoring.