(264a) In Vivo Characterization of Glucose Responsive Insulin Delivery Systems

In Vivo Characterization of Glucose
Responsive Insulin Delivery Systems
Lisa R Volpatti, Morgan A. Matranga,
Abel C. Cortinas, Robert S. Langer, Daniel G.
Anderson

Koch
Institute of Integrative Cancer Research & Dept. of Chemical Engineering,
MIT, Cambridge, MA 02139

Motivation: Diabetes mellitus is a disease characterized by poor
glycemic control as a result of deficient insulin production (Type 1) or
signaling (Type 2). Chronic hyperglycemia often leads to severe complications including
retinopathy, cardiovascular disease, kidney failure, and cancer. Therefore, Type
1 and advanced Type 2 diabetic patients are dependent on exogenous insulin to
combat hyperglycemia. To improve patient compliance, an attractive diabetes
therapy is the self-regulated delivery of insulin to mimic native insulin
production or signaling. Many chemically glucose-responsive systems which employ
the enzyme glucose oxidase as a glucose sensor have been reported in the
literature. However, only ~25% of publications on glucose-responsive insulin
release even include in vivo studies,
and there have been very few reports of true glucose-responsiveness in vivo. Here we show extensive in vivo results including pharmacodynamic and pharmacokinetic studies of both healthy
and diabetic mouse models using our acetalated-dextran
based glucose-responsive nanoparticle system.

Results: Acetalated-dextran nanoparticles
were formulated containing insulin and glucose oxidase. Glucose oxidase
converts glucose to gluconic acid thus reducing the
pH of the microenvironment when glucose levels are high. Acetalated
dextran rapidly degrades under acidic conditions, releasing insulin on demand.
Several in vivo studies were conducted
to determine the safety and efficacy of these nanoparticles. Safety studies
were performed by administering nanoparticles to healthy mice, measuring their
blood glucose levels (Fig.1a), and calculating the corresponding hypoglycemic
index (Fig 1b). The hypoglycemic index is an indication of the severity of
blood glucose depression under normal physiological conditions. We show that two
different doses of our nanoparticles have significantly lower hypoglycemic
indices than that of naked insulin, indicating that they are safe to inject in
the absence of a high blood sugar state.

We further show the
efficacy of the nanoparticles using a diabetic mouse model with three simulated
meals (Fig 1c) and compare the area under the curve for naked insulin, two
doses of nanoparticles, and a dose of nanoparticles with a skipped simulated
meal (Fig. 1d). These results indicate that if a diabetic patient were to miss
a meal, then he/she would not be at risk of a hypoglycemic episode and would
still maintain tight glycemic control. We next compare our nanoparticles to
control particles omitting either insulin or glucose oxidase (Fig. 1e) and
quantify the area under the curve to show that the enzymes are necessary to
achieve normoglycemia (Fig. 1f). We also perform two
week experiments by encapsulating the nanoparticles in a hydrogel network to
create a long acting insulin depot. Additionally, we perform pharmacokinetic
studies by quantifying the amount of insulin in mouse serum samples for both
healthy and diabetic models to show glucose responsiveness. Finally, we show by
in vivo fluorescent imaging that our
nanoparticles are glucose responsive.

Conclusions: In vivo studies using both healthy
and diabetic animal models are key to advancing glucose-responsive insulin
therapies. Here we show in a broad range of in
vivo studies that our optimized acetalated-dextran
nanoparticles release insulin in response
to high levels of glucose, achieving tight glycemic control, and creating a self-regulated system
that could be used as an injectable therapy for diabetic patients.