Fig 1.
A photo image of a rat APS system.
The rat APS system consists of an insulin pump, a CGM, an Intel Edison loaded with OpenAPS, a laptop as a control server, an electronic food scale, and a restrainer to hold a rat.
Fig 2.
BGL and insulin infusion rate of a diabetic rat controlled by OpenAPS.
A representative Nightscout image is shown with color-coded BGL including normal (green), moderate hypoglycemia (yellow), and severe hypoglycemia (red). The blue shaded boxes show basal insulin infusion rate controlled by OpenAPS. The big red circles represent BG readings determined by Contour Next glucose meter, which were used to calibrate the sensor. Other pertinent information such as current BG (86 in green), the trend of BG (horizontal arrow in green), the current time (9:24), sensor reading (7 min ago), and OpenAPS looping (1 min ago), etc. are also shown. The image shows a representative of 7 different experiments using two different diabetic rats.
Fig 3.
Determination of an optimal concentration of Humalog for diabetic rats.
Representative Nightscout images of 1:3, 1:7, and 1:20 dilutions of Humalog are shown. OpenAPS controlled BGL was observed for the same diabetic rat using different dilutions of Humalog. Each individual experiment was run following the same procedure as described in the Methods section. A total of 20 independent experiments using 5 different diabetic rats and various dilutions of Humalog were performed.
Fig 4.
Effects of different diets on BGL and insulin infusion in diabetic rats.
Representative Nightscout images of diabetic rats fed with chow diet (upper panel) or sugar pellets (lower panel) are shown. A total of 207 closed-loop hours (5 different rats with 10 independent experiments) using 1:7 dilution of Humalog was collected for diabetic rats fed with chow diet. Using the data analysis program described in the Methods section, the percent of time in hyper-, hypo- and normal glycemia as well as insulin infusion rate were calculated as shown in Table 1. A total of 175 closed-loop hours (3 different rats with 5 independent experiments) were collected for diabetic rats fed with sugar pellets (lower panel).
Table 1.
Summary of the data showing food type, closed-loop hour, and % BGL in hyperglycemia, normal, and hypoglycemia for healthy and diabetic rats.
Fig 5.
Effects of different diets on BGL and suggested insulin infusion by OpenAPS in healthy rats.
Representative Nightscout images of healthy rats fed with chow diet (upper panel) or sugar pellets (lower panel) are shown. A total of 281 closed-loop hours (4 different rats with 8 independent experiments) was collected for healthy rats fed with chow diet. Using the data analysis program, the percent of time in hyper-, hypo- and normal glycemia as well as OpenAPS-suggested insulin infusion rate were calculated as shown in Table 1. A total of 146 closed-loop hours (3 different rats with 5 independent experiments) were collected for diabetic rats fed with sugar pellets (lower panel).
Fig 6.
BGL after intraperitoneal injection of 4 g/kg 50% dextrose in healthy and diabetic rats.
(A) and (C) Representative Nightscout images of BGL for healthy (n = 3) and diabetic (n = 3) rats, respectively, are shown. The red arrow shows the time of injection. Blue lines show suggested insulin infusion by OpenAPS. BGL above 200 mg/dl and below 180 mg/dl are shown in red and green dots, respectively. (B) and (D) BGL determined by Contour Next glucose meter (closed circles) and Enlite sensor (closed inverted triangles) for healthy and diabetic rats, respectively. (E) Area under curve for BGL determined by the glucose meter for healthy and diabetic rats. Values are presented as means ± SEM (n = 3).
Fig 7.
(A) BGL after intraperitoneal injection of 2 g/kg and 3 g/kg 50% dextrose in healthy and diabetic rats. Panels a and b show representative Nightscout images for BGL for healthy rats (n = 3) injected with 2 g/kg and 3 g/kg 50% dextrose, respectively, and Panels c and d show those for diabetic rats (n = 3). (B) Area under curve for BGL determined by Enlite sensor for each condition as indicated on the figure. Values are presented as means ± SEM (n = 3).
Fig 8.
Representative images of islets from healthy and diabetic rats.
Panels a-e (healthy) and f-j (diabetic) show frozen pancreas sections immunostained for insulin (red) and glucagon (green) to localize β- and α-cells, respectively. Pancreas sections (n = 3 for each condition) were processes for immunostaining using appropriate primary and secondary antibodies as stated in the Methods section. DAPI was used for nuclear staining (blue). Florescent images were obtained using a 40X objective in an Olympus FlowView confocal microscope.
Fig 9.
Assessment of glycogen storage in the liver.
(A) Panels a and b show representative images of Periodic acid-Schiff (PAS) staining of frozen liver sections isolated from healthy and diabetic rats, respectively. Panels c (healthy) and d (diabetic) show the images of PAS staining after amylase treatment as a control. (B) Integrated intensity of the PAS staining with and without amylase. Data are the means ± SEM of ~26–30 liver sections from 3 rats per condition.
Fig 10.
Assessment of lipid accumulation in the liver.
(A) Panels a and b show representative images of Nile red staining of frozen liver sections isolated from healthy and diabetic rats, respectively. Fluorescent images were obtained using a 40X objective in an Olympus FluoView confocal microscope. (B) Integrated fluorescent intensity over number of nuclei for each liver section was determined using ImageJ image processing and analysis program. Data are the means ± SEM of ~26–30 liver sections from 3 rats per condition.
Table 2.
Plasma TAG, total cholesterol, and HDL levels of healthy and diabetic rats (n = 5).
Table 3.
Body weight and HbA1c levels of healthy and diabetic rats.