1. Introduction

Type 2 diabetes mellitus (T2DM) is becoming an epidemic worldwide, and due to diet or genetic predisposition, an estimated 6.28% of the population experiences this chronic disease [1]. As one of the most serious health problems, T2DM is associated with an increased risk of renal and cardiovascular diseases (CVDs) [2], as well as neurological complications in both the peripheral and central nervous systems [3]. Accumulating evidence demonstrates that T2DM presents significant challenges related to maintaining physical independence and cognitive ability in the context of an ageing population, particularly with regard to cognitive impairment [4]. Individuals with T2DM develop cognitive impairment early and are likely to experience Alzheimer’s disease, characterised by cognitive dysfunction [5], which is expected to continue to increase in prevalence over time [1]. Therefore, it is important to alleviate the cognitive impairment caused by T2DM to improve patient independence and quality of life.

Aerobic exercise is increasingly suggested as an effective nonpharmacological strategy to combat cognitive decline in T2DM patients [6]. For instance, Choi et al. identified one of the main proposed mechanisms whereby exercise affects cognitive function by increasing the brain-derived neurotrophic factor (BDNF) [7]. Furthermore, aerobic exercise can also upregulate the expression of synaptic plasticity-associated proteins [8]. These findings imply that exercise may be an effective treatment for diabetes-related cognitive impairment. However, the molecular mechanism by which aerobic exercise improves cognitive function is poorly understood.

Janus kinase 2 (JAK2) is a vital component of nonreceptor tyrosine kinases [9]. Once activated, JAK2 phosphorylates and signal transducer and activator of transcription (STAT) established the downstream target [10]. Increasing evidence suggests that JAK2/STAT3 proteins have important functions in apoptosis and also participate in food intake under leptin-mediated control [11, 12]. Moreover, the JAK2/STAT3 pathway has been found to be related to memory dysfunction [13]. A recent study revealed that mice injected with the JAK inhibitor, AG490, exhibited spatial memory impairment [14]. Based on these data, we speculated that the JAK2/STAT3 pathway is involved in regulating cognitive impairment induced by T2DM; it is unclear whether exercise plays a role in regulating the JAK2/STAT3 pathway.

In addition, an increase in peripheral blood glucose concentration is a typical characteristic of T2DM, which leads to hippocampal energy metabolism disorder [15]. Adenosine monophosphate-activated protein kinase (AMPK), as well as sirtuin 1 (SIRT1), represent two enzymes involved in the regulation of energy metabolism [16]. AMPK, a regulator of body glucose and hippocampal energy balance [17], is widely administered to increase food intake [18] and forms one of the insulin receptor signalling pathways [19]. SIRT1 is one of the most important protective factors for metabolic syndrome. SIRT1 exhibits the ability to significantly extend cell lifespan and enhance neurogenesis by regulating various pathways [20], making it an exciting therapeutic target for neurodegenerative diseases. Thus, the essential role of AMPK/SIRT1 activity in energy metabolism led us to hypothesise that the AMPK/SIRT1 pathway may be related to the exercise-induced regulation in T2DM mice.

Therefore, in the present study, we explored the effect of aerobic exercise on cognitive impairment in T2DM mice and investigated whether the JAK2/STAT3 and AMPK/SIRT1 pathways are involved in the underlying mechanism.

2. Materials and Methods

2.1. Animals

Each animal experiment was carried out according to the recommendations of the ethics committee guidelines of Wenzhou Medical University (Protocol Number: wxdw 2016-0266). Specific pathogen-free (SPF) male leptin receptor-deficient db/db mice (C57BLKsJ/Nju, 8 weeks) and age-matched male db/m mice (C57BL/KsJNju) were obtained (Strain ID: T002407) from GemPharmatech Co., Ltd. All mice were maintained in conventional laboratory conditions (22°C–25°C, 50%–60% relative humidity), 12 : 12 h light-dark cycle. The time duration of animal study is shown in Figure 1.

3. Results

3.1. Aerobic Exercise Improved Blood Glucose and Insulin Level in T2DM Mice

After 12 weeks of treadmill intervention, T2DM+Exe mice displayed significantly improved in GTT and ITT compared to T2DM mice (P < 0.05, Figures 2(a)–2(d)). Compared to the Con group, the levels of plasma insulin and blood glucose were significantly higher in the T2DM group. Exercise improved the insulin and blood glucose levels (P < 0.05, Figures 2(e) and 2(f)).

3.2. Aerobic Exercise Improved T2DM-Related Cognitive Impairment

In this study, we performed the MWM test to investigate the influence of the treadmill on the spatial learning and memory functions of the T2DM mouse. From the representative swimming trajectories of the four groups of mice on the first day of navigation training and latency period, initial differences among the 4 groups were not significant (Figures 3(a) and 3(c)). After training, a significant delay was observed in the T2DM mice, which showed prolonged escape latency and path length compared with the Con mice on the fifth day. The escape latency and path length of the T2DM+Exe mice remarkably decreased compared with the T2DM mice (P < 0.05, Figures 3(c) and 3(d)). The representative trajectory of navigation training on the fifth day showed similar results (Figure 3(b)). During the probe test, the number of platform crossings and the time spent in the target quadrant of the T2DM group remarkably decreased in comparison with the Con group. Moreover, the T2DM+Exe mice displayed a higher crossing number and more time staying in the target quadrant compared with the T2DM group (P < 0.05, Figures 3(e) and 3(f)). The result of total swimming distances among the four groups was not significant (P > 0.05, Figure 3(g)).

3.3. Aerobic Exercise Improved Nerve Cell Damage and Protected Synaptic Damage in the Hippocampus Caused by T2DM

To observe the damage of hippocampal nerve cells in diabetes, hematoxylin and eosin-stained paraffin sections were used to observe the hippocampal CA1 region. In the Con and Con+Exe groups, most of the neurons were arranged regularly and contained large, round nuclei. In the T2DM group, more nuclei had become pyknotic (contracted and dark, arrowheads in Figure 4(a)) and were surrounded by swollen and broken protrusions. In the T2DM+Exe group, the abnormal structure of the neurons was improved (Figure 4(a)). The number of damaged neurons significantly increased in the T2DM mice and markedly decreased after exercise intervention (P < 0.05, Figure 4(e)). In addition, immunohistochemical and Western blotting analyses showed that the positive expression of ADPN, PSD95, and NMDAR1 significantly declined within the T2DM group in comparison with the Con group. This improved in the exercise intervention group (Figures 4(b)–4(d) and 4(g)). Moreover, the protein expression and the mRNA levels of BDNF and SYN1 markedly decreased in the T2DM mice compared with the Con mice and evidently increased after training (P < 0.05, Figures 4(c), 4(f), and 4(g)).

3.4. Aerobic Exercise Attenuated the JAK2/STAT3 Pathway in T2DM Mice

We also observed that the hippocampal phospho-JAK2 and phospho-STAT3 levels in the T2DM group were significantly higher than those in the Con mice. With regard to the Con+Exe and T2DM+Exe mice, long-term treadmill significantly decreased phospho-JAK2 and phospho-STAT3 expression (P < 0.05, Figures 5(c), 5(f), and 5(g)). As depicted in the immunofluorescence of Figure 5, the same tendency of phospho-JAK2 and phospho-STAT3 levels was observed (P < 0.05, Figures 5(a), 5(b), 5(d), and 5(e)).

3.5. Aerobic Exercise Activated the AMPK/SIRT1 Pathway in T2DM Mice

To evaluate the effect of exercise on AMPK/SIRT1 pathway, we performed the western blotting. And the SIRT1 and phospho-AMPK expression evidently declined in the T2DM mice compared to the Con mice. In the Con+Exe and T2DM+Exe groups, exercise remarkably elevated SIRT1 and phospho-AMPK expression (P < 0.05, Figures 6(e)–6(g)). In addition, the same tendency of phospho-AMPK was observed in immunohistochemical staining, SIRT1 in immunofluorescence staining, and SIRT1 activity (P < 0.05, Figures 6(a)–6(d) and 6(h)).

3.6. Aerobic Exercise-Induced Improvement in the Learning and Memory Ability of T2DM Mice Was Dependent upon JAK2/STAT3 and AMPK/SIRT1

The JAK2 activator RO8191 was used to enhance the activation of JAK2/STAT3. The AMPK inhibitor Compound C was used to inhibit the activation of AMPK/SIRT1 (Supplementary Figure S1). From the representative swimming trajectories on the first day of navigation training and latency period, the difference among the three groups was insignificant (Figures 7(a) and 7(c)). However, escape latency and path length revealed an increase on the fifth day in the T2DM+Exe+RO8191 group compared to the T2DM+Exe+Vehicle group (P < 0.05, Figures 7(b)–7(d)). Compared to the T2DM+Exe group, the escape latency and path length of the T2DM+Exe+Compound C group remarkably increased (P < 0.05, Figures 7(b)–7(d)). During the probe test, the number of platform crossings and the time spent in the target quadrant of the T2DM+Exe+RO8191 group remarkably decreased in comparison with the T2DM+Exe+Vehicle group (P < 0.05, Figures 7(e) and 7(f)). And the T2DM+Exe+Compound C mice displayed a lower crossing number and less time staying in the target quadrant compared with the T2DM+Exe+Vehicle group (P < 0.05, Figures 7(e) and 7(f)). There was no significant difference among the three groups for total swimming distances (P > 0.05, Figure 7(g)).

3.7. Aerobic Exercise-Induced Improvement in Nerve Cell Damage and Protection in Synaptic Damage of T2DM Mice Was Dependent upon JAK2/STAT3 and AMPK/SIRT1

We observed the damage of hippocampal nerve cells in the CA1 region after the application of RO8191 or Compound C. The damaged nerve cells of RO8191-treated mice were significantly more than those of the vehicle mice (P < 0.05, Figures 8(a) and 8(e)). In the Compound C-treated mice, it was observed that more nuclei had become pyknotic (contracted and dark, arrowheads in Figure 8(a)) in comparison with the T2DM+Exe+Vehicle group (P < 0.05, Figures 8(a) and 8(h)). Further, the positive expression of ADPN, BDNF, SYN1, PSD95, and NMDAR1 significantly declined within the T2DM+Exe+RO8191 group in comparison with the T2DM+Exe+Vehicle group (P < 0.05, Figures 8(b), 8(c), 8(f), and 8(g)). Respectively, the positive expression of ADPN, BDNF, SYN1, PSD95, and NMDAR1 evidently decreased after the injection of Compound C (P < 0.05, Figures 8(b), 8(d), 8(i), and 8(j)).

4. Discussion

In the present study, we found that the 12 weeks of moderate-intensity treadmill training could improve the abnormal glycemia and insulin levels in db/db mice. Additionally, exercise remitted impaired glucose tolerance and insulin resistance. Emerging evidence has demonstrated that cognitive impairments occur in T2DM [24]. We found that cognitive dysfunction related to the function of the hippocampus existed in the T2DM mice, and we revealed that aerobic exercise improved it. In a subsequent investigation, we sought the possible mechanism responsible for the aforementioned phenomenon of recovery of cognitive function via exercise. We observed that phospho-JAK2 and phospho-STAT3 were increased, and the phospho-AMPK/SIRT1 was reduced in T2DM. However, aerobic exercise intervention over a long period improves this situation. Furthermore, JAK2 activator, RO8191, and AMPK inhibitor, Compound C, partly blocked the beneficial effects of aerobic exercise on cognitive function in T2DM mice. These results indicate that JAK2/STAT3 and AMPK/SIRT1 are potential mechanisms by which aerobic exercise improves cognitive decline mediated by T2DM.

Cognitive impairment caused by diabetes can be explained by the loss of synapses caused by insulin resistance (IR). Synaptic dysfunction is the main pathophysiological sign of neurodegenerative diseases [25]. Further, previous studies have indicated that synaptic dysfunction is caused by IR in the brain [26], leading to cognitive decline. However, studies have shown that aerobic exercise could attenuate synaptic loss in models of neurodegenerative disorders, such as Alzheimer’s disease [7]. In this study, aerobic exercise mitigated the changes in classical synaptic protein levels observed in T2DM. This is because exercise benefits glucose homeostasis in mice with IR. Glucose homeostasis affects the normal physiological activities of neurons and is able to induce synaptic proteins changes [27]. In this study, our data support that aerobic exercise efficiently improved the impaired cognition caused by diabetes. The expression of three classical synaptic proteins, PSD95, NMDAR1, and SYN1, decreased in the hippocampal CA1 area of diabetic mice, indicating that hippocampal synapse formation in diabetic mice was impaired. The aerobic exercise intervention enhanced hippocampal synaptic formation in T2DM mice. Additionally, we found more abnormal structure neurons in db/db mice, consistent with previous research results [23]. Aberrant insulin modulation in diabetic mice can cause abnormal nerve cell formation. Furthermore, our investigation suggested improvement of neuron shape within hippocampal CA1 region in db/db mice after exercise. Therefore, our research indicates that aerobic exercise has a protective effect on the synapses and neurons in diabetes.

We found that cognitive impairment in T2DM manifests as synaptic disorders and neuronal damage. Recent studies have shown that the JAK2/STAT3 pathway is related to synaptic disorders [28]. JAK2 has a negative impact on synapse formation [29] and insulin sensitivity in diabetes [30]. In addition, the activation of STAT3 has been shown to exacerbate neuronal damage [31]. However, to date, there is a lack of evidence to indicate that aerobic exercise can improve this situation through the JAK2/STAT3 pathway. In our experiments, our data confirmed that phospho-STAT3 and phospho-JAK2 levels were elevated in the mice model of T2DM. In contrast, prolonged aerobic exercise successfully reversed the increased signal of this pathway. Therefore, the data showed that the JAK2/STAT3 pathway is one of the potential mechanisms by which aerobic exercise reduces insulin resistance and protect synapse loss. To verify it, we used RO8191, an activator of JAK2, to activate JAK2/STAT3. We found that the positive effect of exercise on cognitive impairment is reversed after activating JAK2. This indicated that exercise inhibited the JAK2/STAT3 pathway, which is likely one of the mechanisms by which aerobic exercise improves the symptoms of diabetes.

With regard to the AMPK/SIRT1 pathway, evidence has indicated that under pathological conditions, AMPK plays a crucial role in controlling synaptic plasticity and neuronal cell survival [32]. In hippocampal neurons of T2DM mice, the decrease in AMPK phosphorylation is related to the damaged of hippocampal neurons [33], and the activation of phospho-AMPK increases the expression of proteins involved in synaptic plasticity [34]. Moreover, SIRT1 has an important function in regulating neuronal injury [35]. In line with these results, we found that the phospho-AMPK in T2DM was significantly reduced. As confirmed in this study, aerobic exercise upregulated phospho-AMPK and SIRT1 in the hippocampus. Furthermore, we used Compound C, the AMPK inhibitor. Additionally, our data confirmed that the cognitive function improvement by exercise of T2DM mice was partly blocked after suppressing AMPK. Therefore, we believe that aerobic exercise may also prevent synapse formation disorders and improve symptoms of T2DM by activating the AMPK/SIRT1 pathway.

In summary, this study demonstrated that prolonged aerobic exercise can reverse the cognitive decline caused by T2DM, which may be related to the activity of AMPK/SIRT1 and the inhibition of JAK2/STAT3 signalling in T2DM mice. Aerobic exercise may also be used to prevent synapse formation disorders and cognitive decline through enhancing the AMPK/SIRT1 and inhibiting the JAK2/STAT3.

Ethical Approval

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of Wenzhou Medical University (protocol number: wxdw 2016-0266).

Conflicts of Interest

The authors have no relevant financial or nonfinancial interests to disclose.

Authors’ Contributions

All authors contributed to the study’s conception and design. Material preparation, data collection, and the first draft of the manuscript were performed by Lili Lin and Yonghua Wang; analysis was performed by Xiaoyong Zhao and Wenli Xu; Jinrong Hu contributed to manuscript revision; Chaolu Huang and Xixi Chen contributed to the conception of the study; Xinhuang Lv helped perform the analysis with constructive discussions; Yuelin Qin contributed to manuscript preparation; Haiyan Li contributed significantly to experimental plan and revised the manuscript. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Lili Lin and Yonghua Wang contributed equally to the manuscript.