Studies of pathology and pharmacology of diabetic encephalopathy with KK‐Ay mouse model

Abstract Aims Pathogenesis of diabetic encephalopathy (DE) is not completely understood until now. The purposes of this study were to illustrate the changes in morphology, function, and important transporters in neurons and glia during DE, as well as to reveal the potential therapeutic effects of medicines and the diet control on DE. Methods Spontaneous obese KK‐Ay mice were used to investigate diabetes‐induced cognitive disorder, the morphology, function, and protein expression changes in impact animal and the cell level studies. The new drug candidate PHPB, donepezil, and low‐fat food were used to observe the therapeutic effects. Results KK‐Ay mice at 5 months of age showed typical characteristics of type 2 diabetes mellitus (T2DM） and appeared significant cognitive deficits. Morphological study showed microtubule‐associated protein 2 (MAP2) expression was increased in hippocampal neurons and glial fibrillary acidic protein (GFAP) expression decreased in astrocytes. Meanwhile, the vesicular glutamate transporter 1 (vGLUT1) expression was increased and glucose transporter 1 (GLUT1) decreased, and the expression of brain‐derived neurotrophic factor (BDNF) and glial cell‐derived neurotrophic factor (GDNF) was also reduced in KK‐Ay mice. Microglia were activated, and IL‐1β and TNF‐α were increased obviously in the brains of the KK‐Ay mice. Most of the above changes in the KK‐Ay mice at 5 months of age could be relieved by diet intervention (DR) or by treatment of donepezil or new drug candidate PHPB. Conclusion KK‐Ay mouse is a useful animal model for studying DE. The alterations of morphology, structure, and function of astrocyte and microglia in KK‐Ay mice might be rescued by DR and by treatment of medicine. The proteins we reported in this study could be used as biomarkers and the potential drug targets for DE study and treatment.


| INTRODUC TI ON
It has been demonstrated that type 2 diabetes mellitus (T2DM)is a major risk factor of the dementia. 1,2 Diabetes-related cognitive dysfunction with the altered morphology and abnormal neuronal physiological function has been defined as diabetic encephalopathy (DE). 3,4 Most of the earlier studies focused on the similarity between DE and vascular dementia (VD), but researchers have recently found that DE is also associated with Alzheimer's disease (AD). 5 Epidemiological data showed that patients of T2DM are about two times higher risk for dementia compared to the people without diabetes, and currently, one in 10 cases of dementia can be attributed to T2DM. 5 Because of the increased incidence of T2DM in worldwide, DE has become a serious medical and social issue. Therefore, it is important to reveal the pathological mechanisms in DE and find therapeutic strategies.
Hippocampus, the key area for learning and memory in the brain, is particularly sensitive to changes in glucose and lipid homeostasis. 6,7 Recently, both astrocyte and microglia, which is not only taking charge in supporting survival and function of neuron but also affecting the capability of learning and memory directly, 8,9 have been caused more and more attention in the DE researches.
A great body of evidence has shown that blood-brain barrier (BBB) injury, energy metabolic disorder, electrophysiological deficits, neuroinflammation, neurotransmitter change, and neuronal apoptosis were found both in the brain of T2DM patients and diabetic animal models and could be involved in the pathological mechanism of DE. 1,6,[10][11][12][13] However, different studies had yielded conflicting results. For example, some studies had indicated that severe loss of cerebral neuron was observed in the diabetic rat in the early stage, but other reports had shown that there was just moderate degeneration of cerebral neuron along in the rat with the diabetes development. 1,2,6,7 Moreover, several researches had found astrogliosis, glutamate excitotoxicity occurred in the brain of diabetic animal models, when other reports showed no obvious change in them or even had opposite results. 1,11 It might be attributed to the differences in experimental models and the observation time. The streptozotocin (STZ)-induced diabetic rat was used in most studies; however, it is difficult to mimic T2DM of humans which is a heterogeneous and multifactorial disease. 14 The KK-Ay mouse, transferred the yellow obese gene (Ay allele) into the Kuo Kondo (KK) mouse which spontaneously developed to severe obesity, hyperglycemia, and insulin resistance, was established as a suitable and useful model mimicking human obesity and T2DM. 15 So far, the study of obese diabetes inducing cognitive deficits was scarcely reported in KK-Ay mice. In our previous study, we had found a high incidence of dementia in KK-Ay mice. 16 In the present study, we simultaneously observed the histological and functional change in both neuron and glia in the hippocampus of 3-, 5-, and 7-month-old KK-Ay mice to elucidate the mechanism of DE in the duration of the disease. In addition, we investigated the protective effects of two agents on DE, one is donepezil which is a classical agent for AD and another is dl-PHPB (potassium 2-(1-hydroxypentyl)-benzoate) which is a novel drug candidate for the treatment of cerebral ischemic stroke. PHPB is the prodrug of listed antiischemia agent 3-n-butylphthalide (dl-NBP), 17 and recently, it was reported to be potent for AD 18 and VD in animal models. 19 Dl-PHPB is now in phase II clinical studies for ischemic stroke. Our present study is helpful to understand the difference between DE and AD and can provide some valuable information for early diagnose and treatment of DE.

| Animal and treatments
Male KK/Upj-Ay/J mice (Beijing HFK Bioscience Co. Ltd.) were fed with a high-fat diet until the experiments were carried out at 3, 5, and 7 months of age. Age-matched male C57BL/6J mice (WT mice; Beijing Vital River Laboratory Animal Technology Co, Ltd) were generally employed as nondiabetic control for the KK-Ay mice.
Male KK-Ay mice were randomly divided into four groups: untreated group (KK-Ay, given high-fat diet) and treated groups.
Treated groups were received dl-PHPB (PHPB; Yunnan Haobang Pharmaceutical Co. Ltd, 150 mg/kg/day) or donepezil group (Don, Eisai Co, Ltd., 3 mg/kg/day) by oral gavage, or diet intervention (DR) group by giving a normal low-fat diet. All treatment continued for two months beginning at 3 months of age. Age-matched C57BL/6J mice were used as the control group (WT) by treatment with physiological saline.
Mice were group-housed in an animal room at a constant temperature of 23 ± 1°C and maintained at a 12 hours light/dark cycle per day. All animals were given food and water ad libitum. All experiments were approved and performed in accordance with the institutional guidelines of the Experimental Animal Center of the Chinese Academy of Medical Science, Beijing, China.

| Morris water maze
The spatial learning and memory of rats were tested according to the method of R. Morris. 9 Briefly, the mouse was gently released into a circle metal pool (120 cm in diameter), filled with water made opaque by the addition of milk powder, from one of the four preplanned starting positions. Spatial training of the hidden platform (submerged 1 cm below the water) in the water maze was performed for five consecutive days. On each day, the mice were allowed to consecutive training trials of which allowed the mouse to swim for a maximum duration of 120 seconds in each trial to find the hidden platform. The duration for finding the platform was recorded as latency. The mice both of which found the hidden platform or not allowed to stay on the platform for 30 seconds. The mouse with blindness was excluded.

| Biochemistry index
Blood samples were collected by cutting the tail or from the heart of mice anesthetized with 4% chloral hydrate in the end of the experimental period. Random plasma glucose, cholesterol, and triglyceride were measured by using test kits (Nanjing Jiancheng Bioengineering Institute, China), and random plasma insulin was analyzed using enzyme-linked immunosorbent assay (ELISA) (DRG Diagnostics).

| Histological staining
After perfusion through the heart in the deeply anesthetized mice, the brains were taken out. The right hemisphere was used for hematoxylin-eosin (HE) and immunofluorescence staining. The hippocampus was dissected from the left hemispheres for western blotting (WB) and ELISA. The right hemisphere was immediately immersed into 4% paraformaldehyde for 18 hours; then, it was sequentially transformed into 10%, 20%, 30% sucrose solution, with 24 hours each time. After complete dehydration, the tissue was frozen and then prepared for frozen section. The brain was cut serially on a Leica microtome into a 20-μm-thick coronal section. After immersing in PBS for 30 minuts to get rid of the OCT, all coronal sections were mounted on glass slides and stained with the routine HE technique. 20

| Immunofluorescence staining
The coronal sections were incubated with 10% normal donkey serum to block nonspecific binding, followed by an overnight incu- Fisher Scientific) for 2 hours at room temperature. The sections were rinsed and transferred on slides, and the cover slipped in an antifade agent. The images were acquired using a Nikon camera mounted on a Nikon Eclipse 80i microscope (Nikon) and analyzed using Image J software. The threshold of detection was held constantly during analysis. Tris-buffered saline with 0.1% Tween-20, incubated with horseradish peroxidase-conjugated secondary antibodies at room temperature for 1 hour, and detected using an enhanced chemiluminescence (ECL) kit. Densitometric evaluation was analyzed using the Quantity One image analysis software (Bio-Rad). β-Actin was used as a loading control.

| Enzyme-Linked Immunosorbent Assay (ELISA)
BDNF, GDNF, IL-1β, and TNF-α in the hippocampal homogenate were detected by ELISA according to the manufacture's instruction (Cusabio). The values were referred as the relative amount per total protein (pg/mg).

| Statistical analysis
All data were expressed as mean ± SEM. Repeated measures analysis of variance or Student's t test or two-way analysis of variance (ANOVA) followed by post hoc LSD test was used for multiple comparisons (SPSS version 19.0, SPSS Inc). Statistical significance was set to a value of P < .05.

| The effects of PHPB, donepezil, and diet intervention on the metabolic syndrome and spatial learning and memory deficits in KK-Ay mice
Our previous study had shown that the body weight, random blood glucose, cholesterol, triglyceride, and insulin of KK-Ay mice were significantly increased compared to control age-matched group at 3,5, and 7 months of age, which suggested that KK-Ay is a reliable model of T2DM. 16 Furthermore, KK-Ay mice showed obvious spatial learning and memory deficits in the diabetic duration. 16 In the present study, we found that PHPB (150 mg/kg) and donepezil (3 mg/kg) treatment had no significant effect on the biochemistry index of 5-month-old KK-Ay mice but DR could reduce the blood glucose and body weight ( Figure 1A-D). In Morris water maze, all treatment groups showed improvements in behavioral performance without any obvious differences in swimming velocity ( Figure 1E,F). These results optimistically indicated that the cognitive deficits in KK-Ay mice were reversible and could be prevented at an early stage.

| Changes in neurons in the hippocampus of KK-Ay mice
Hematoxylin-eosin staining revealed that the neuronal nuclei of KK-Ay mice organized closely and the cytoplasm were stained homogeneously ( Figure 2A

| Changes of astrocytes in the hippocampus of KK-Ay mice
Glial fibrillary acidic protein (GFAP) is an astrocyte-specific cytoskeleton protein which indicates the activation of astrocytes. As shown in GFAP immunostaining ( Figure 3A

F I G U R E 1
The changes of biochemical indexes and cognitive behaviors in 5-mo-old KK-Ay mice and the improvements by PHPB, donepezil, and diet intervention (DR). A-D, Random blood glucose, body weight, cholesterol, and triglyceride were significantly enhanced in KK-Ay mice. No obvious changes in glucose, cholesterol, and triglyceride were detected in KK-Ay mice after treatment with PHPB 150 mg/ kg or donepezil (Don) 3 mg/kg for two months, but DR for two months decreased the glucose and body weight in KK-Ay mice. E, The escape latency of 5-mo-old KK-Ay mice to find the hidden platform was longer than WT mice in Morris Water Maze. All KK-Ay mice in PHPB, donepezil, and DR groups showed good performance. F, No obvious difference in swimming velocity was observed in both KK-Ay and WT mice of all groups, n = 12-16 per group. ** P < .01 versus WT mice, # P < .05, ## P < .01 versus KK-Ay mice

| Microglia was activated in KK-Ay mice, and it was inhibited by PHPB, donepezil, and diet intervention
Activated microglia is regarded as mainly neuroinflammatory cell type.
We investigated the changes of microglia in the hippocampus by immunostaining of ionized calcium binding adapter molecule 1 (IBA-1) which is a prominent marker indicating the activated microglia ( Figure 4A

| Abnormal expression of transporters of glucose and glutamate in the hippocampus of KK-Ay mice
Glucose transporter 1 (GLUT1, 45 kDa) is mainly expressed in the astrocyte and is mainly responsible for maintaining the stability of glucose metabolism. Compared with age-matched WT mice, the expression of GLUT1 was significantly reduced by 26.42% and 28.49% in the hippocampus of 3-and 5-month-old KK-Ay mice ( Figure 5A,B).
However, there was no obvious reduction of GLUT1 in KK-Ay mice at 7 months of age. The data indicated that the GLUT1 might dynamically change in KK-Ay mice with the metabolism disorder. In addition, the expression of vesicular glutamate transporter (vGLUT1) that specifically expressed in the presynaptic membrane to regulate the release of glutamate was obviously increased in the hippocampus of 5-and 7-month-old KK-Ay mice up to 16.1% and 42.0%, but no obvious change in KK-Ay mice at 3 months ( Figure 5A,B). Another glutamate transporter referred to excitatory amino acid transporter 2 (EAAT2), which mainly expressed in astrocytes to take charge of eliminating redundant glutamate in the synaptic cleft, had no change in KK-Ay mice compared to WT mice ( Figure 5A,B). These results illustrate that both GLUT1 and vGLUT1 might involve in the pathology of DE in KK-Ay mice. showed an elevation and was closed to the WT group.

| Reductions of BDNF and GDNF and increase of IL-1β and TNF-α in the hippocampus of KK-Ay mice and the recoverable effect of PHPB and donepezil and diet intervention
The level of BDNF in the hippocampus of KK-Ay mice decreased at 3, 5, and 7 months of age. Similarly, the level of GDNF also showed a significant decrease at age of 7 months and the clear decline trend in 3 and 5 months age compared with WT ( Figure 6A,B). On the contrary, the inflammatory factors IL-1β and TNF-α were significantly increased in the hippocampus of 5-and 7-month-old KK-Ay mice ( Figure 6C,D).
Obviously, PHPB and donepezil could increase the concentration of BDNF and decrease the level of IL-1β of KK-Ay mice ( Figure 6E,F).
Diet intervention could also reduce the level of IL-1β to some content.

| D ISCUSS I ON
Our study demonstrated that KK-Ay mouse is a stable and useful model of T2DM and the mice showed obviously cognitive deficiency in Morris water maze test. It was similar to the previous studies, which KK-Ay mice displayed a cognitive dysfunction at quite young with low activity of long-term potentiation (LTP). 16   Moreover, we also demonstrated that the cognitive deficits in KK-Ay mice could be relieved by a low-fat diet. It could be resulted by the recovery of blood glucose and fatty acids and cholesterol levels, subsequently relieved pathological changes in blood vessels. Our data indicate that the cognitive deficits of T2DM could be recovered at the early stage of the disease. It is different from AD that is almost irreversible. Donepezil and dl-PHPB would be the potential therapeutic option for DE.

| CON CLUS ION
Diabetic encephalopathy study with KK-Ay mice showed an early stage change in neurons and glial cells. Especially, the morphology of astrocytes and its structural marker protein GFAP changed significantly. Meanwhile, neuronal glutamate transporter vGLUT1 increased and astrocyte glucose transporter GLUT1 decreased. All the above pathological changes could be rescued by DR and by neuronal protective agent treatment. The alteration in blood vessels in DE remains to be studied in the future.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.