Passive Diffusion of Transdermal Glucose: Noninvasive Glucose Sensing Using a Fluorescent Glucose Binding Protein

Abstract

The motivation for this study was to determine if a statistically significant correlation exists between blood glucose (BG) and transdermal glucose (TG) collected by passive diffusion. A positive outcome will indicate that noninvasive passive TG diffusion is a painless alternative to collecting blood through a break on the skin. Sampling involves placing a small volume of buffer solution on the surface of membrane or skin for 5 minutes. The sample is then assayed with fluorescent GBP. In vitro testing was done on regenerated cellulose and a porcine skin model to determine diffusion of standard glucose solutions. In vivo testing was done on a healthy subject and a subject with type 2 diabetes. Glucose diffused readily through the regenerated cellulose membrane with good correlation between surface and internal glucose concentrations (R ² = .997). But the porcine skin model required a surface prewash to achieve the same good correlation R ² = .943). Based on this, an optimum prewash step was determined for the in vivo studies. The resulting correlation coefficients between TG and BG after a 15-minute prewash in a healthy subject and type 2 subject were .87 and .93, respectively. Removal of the extraneous glucose in the skin by prewashing was an important step in achieving good correlation between TG and BG. The results suggest that passive collection of TG is a noninvasive alternative to current practice of breaking the skin. Further studies are under way to determine the lag time between TG and BG and for the sampling protocol to be more amenable to point-of-care application.

Keywords: fluorescent glucose binding protein; noninvasive glucose sensing.

Figure 1.

Set-up for in vitro testing of glucose passive diffusion through regenerated cellulose membrane or porcine skin.

Figure 2.

(Left) Emission spectra of GBP with increasing concentrations of glucose. A 5 µL aliquot of each standard glucose solution (4, 8, 10, 20, 50, 100 µM) was added to a 250 µL sample of fluorescent GBP to generate a calibration curve (Right). The linear range between 0.08 and 0.41 µM total glucose was used for analysis.

Figure 3.

Correlation between glucose collected on the surface of regenerated cellulose (10-second sampling time) and the glucose inside the chamber.

Figure 4.

Concentrations of glucose collected on the surface of porcine skin (30-second sampling time) immediately after and 30 and 60 minutes after contact with the glucose solution in the chamber.

Figure 5.

Correlation between glucose collected on the surface of porcine skin (60-minute prewash and 30-second sampling time) and glucose inside the chamber.

Figure 6.

Blood glucose (blue) and transdermal glucose (red) of a 21-year-old healthy subject. The initial drop in TG is due to glucose present in the stratum corneum that is not correlated to BG.

Figure 7.

Effect of the prewash time on achieving the steady state TG of a fasting subject. Data are normalized by the first wash cycle.

Figure 8.

Oral glucose challenge of a healthy subject showing both BG and TG (15-minute prewash and 5-minute sampling time).

Figure 9.

Oral glucose challenge of a subject with type 2 diabetes showing BG and TG (15-minute prewash and 5-minute sampling time).

Figure 10.

Correlation between TG and BG in a healthy subject (15-minute prewash and 5-minute sampling time).

Figure 11.

Correlation between TG and BG in a patient with type 2 diabetes (15-minute prewash and 5-minute sampling time).