Diabetes mellitus is a significant global public health problem depicting a rising prevalence worldwide. As a serious complication of diabetes, diabetes-associated cognitive decline is attracting increasing attention. However, the underlying mechanisms are yet to be fully determined. Both endoplasmic reticulum (ER) stress and autophagy have been reported to modulate neuronal survival and death and be associated with several neurodegenerative diseases. Here, a streptozotocin-induced diabetic mouse model and primary cultured mouse hippocampal neurons were employed to investigate the possible role of ER stress and autophagy in diabetes-induced neuronal apoptosis and cognitive impairments, and further explore the potential molecular mechanisms. ER stress markers GRP78 and CHOP were both enhanced in diabetic mice, as was phosphorylation of PERK, IRE1α, and JNK. In addition, the results indicated an elevated level of autophagy in diabetic mice, as demonstrated by up-regulated expressions of autophagy markers LC3-II, beclin 1 and down-regulated level of p62, and increased formation of autophagic vacuoles and LC3-II aggregates. Meanwhile, we found that these effects could be abolished by ER stress inhibitor 4-phenylbutyrate or JNK inhibitor SP600125 in vitro. Furthermore, neuronal apoptosis of diabetic mice was attenuated by pretreatment with 4-phenylbutyrate, while aggravated by application of inhibitor of autophagy bafilomycin A1 in vitro. These results suggest that ER stress pathway may be involved in diabetes-mediated neurotoxicity and promote the following cognitive impairments. More important, autophagy was induced by diabetes possibly through ER stress-mediated JNK pathway, which may protect neurons against ER stress-associated cell damages.
An increasing body of evidence supports the view that diabetic patients present a high risk of developing cognitive disorders [1,2]. Nowadays, diabetes-associated cognitive decline is considered as an accepted chronic complication of diabetes, which has been demonstrated both in animal models  and epidemiological studies [2,4]. It not only reduces self-care ability and quality of life of patients, but also increases the risk of morbidity and mortality . In recent years, cognitive impairments in diabetes has become a common and serious health problem and attracted an increasing attention from the scientific community. However, the mechanisms underlying the pathogenesis of this disorder remain unclear.
Endoplasmic reticulum (ER), an organelle with multiple functions in cells, is entrusted with biosynthesis of lipids and synthesis, folding and maturation of proteins, as well as intracellular calcium storage and signaling. Perturbation of ER homeostasis due to extracellular or intracellular factors, such as disorder of calcium, and accumulation of unfolded or misfolded protein and hypoxia, leads to ER stress and activation of unfolded protein response (UPR), which serves as a protective mechanism to preserve ER homeostasis . In the process of UPR, the chaperone GRP78 in the ER lumen is released to bind to three main sensor in the ER membrane, including protein kinase RNA-like ER kinase, activating transcription factor 6 and inositol-requiring enzyme 1, then activates a serial of transduction signal proteins to restore cellular homeostasis. However, under conditions of prolonged and strong ER stress, the adaptive responses mediated by UPR are not sufficient to restore normal cellular function. Subsequently, apoptosis signaling pathways are activated, including proapoptotic pathway of C/EBP-homologous protein (CHOP), apoptosis signal-regulating kinase 1/c-Jun NH2-terminal kinase (JNK), and activation of the ER localized cysteine protease, caspase-12. ER stress has been implicated in the pathogenesis of several neurodegenerative diseases, such as Parkinson’s disease, Alzheimer’s disease (AD), amyotrophic lateral sclerosis, and multiple sclerosis [7,8]. Hence, we hypothesized that ER stress pathways play a role in the mechanisms of diabetes-induced cognitive dysfunction.
It is of interest that ER stress is also a potent trigger of autophagy . Autophagy is a major intracellular degradation process that targets primarily cytosolic organelles, long-lived proteins as well as misfolded protein aggregates, and is essential for maintaining cellular energy and metabolic homeostasis. It is known to include three main forms: chaperone-mediated autophagy, microautophagy, and macroautophagy. Macroautophagy hereafter referred to as autophagy is the primary mechanism by which cells “self-eat” delivering sequestered cytosolic cargo to lysosomes for proteolytic degradation by lysosomal proteases . In recent years, autophagy has aroused significant interest due to its important role in regulating a variety of clinical diseases. It is revealed that autophagy may play a critical role in neurodegenerative diseases [11–13]. In animal models for AD, it is reported that activation of autophagy contributed to the clearance of intracellular accumulated amyloid β (Aβ) protein , while disruption of autophagy promoted neurodegeneration . Both ER stress and autophagy have been reported to modulate cellular survival and death and be related to several neurodegenerative diseases. Nevertheless, the role of autophagy in the pathogenesis of diabetes-associated cognitive deficits and its potential link to ER stress are yet far from clear.
In the present study, we investigated the possible role of ER stress and autophagy in diabetes-induced neuronal apoptosis and cognitive impairments, and further explore the potential molecular mechanisms using a streptozotocin (STZ)-induced diabetic mouse model and primary cultured mouse hippocampal neurons.
Materials and methods
Animals and study design
All animals were treated in accordance with the guidelines of the Guide for the Care and Use of Laboratory Animals (United States National Institutes of Health). All animal experimental procedures and protocols were approved by the Institutional Animal Care and Use Committee of Tongji University (Shanghai, China). The animals were housed in a room with controlled temperature (21 ± 2°C), humidity (40–60%), and light cycle (12/12-h light/dark) throughout the experiments and were fed with a nutritious standardized diet and with unrestricted access to the distilled water. All efforts were made to minimize the number of animals used and their suffering.
Six-week-old male clean healthy C57BL/6 mice weighing 18–20 g were purchased from Shanghai Laboratory Animal Center and employed for the present study. After a week of acclimation, all animals were fasted overnight, and then three-sixths of the mice were randomly selected for a single intraperitoneal injection of STZ (150 mg/kg; Sigma, St Louis, MO, U.S.A.) at a volume of 10 mg/ml in 0.1 M citrate buffer (pH 4.5) using a 29 G insulin needle to induce diabetes as experiment group [14,15]. STZ-injected mice were given 5% (w/v) glucose during the following 24 h after STZ administration. The remaining mice were injected with an equal volume of sodium citrate buffer without STZ as control group. Three days later, animals with nonfasting blood glucose in a tail–vein sample more than 16.7 mmol/l were considered as diabetic and were used in the following experiment. Three STZ-injected mice died and three failed to be models during this period. Then the rest of the mice in the experiment group were further randomly divided into two groups: diabetes mellitus (DM) and diabetes mellitus combined with 4-phenylbutyrate (4-PBA, Sigma-Aldrich Corp., St. Louis, MO, U.S.A.) (DM + PBA) groups. Those mice in the control group were randomly divided into two groups: control (Con) and control + 4-PBA (PBA) groups.
Cell culture and treatments
Hippocampal tissues from embryonic day 18 mouse embryos were dissected, gently minced and trypsinized (Trypsin, 0.05%), and the digestion was terminated by DMEM combined with 10% heat inactivated FBS . Cells were then seeded at a concentration of 1.5 × 105 per milliliter into six-well plates and maintained in serum-free Neurobasal medium supplemented with 2% B27 and 0.5 mM L-glutamine at 37°C in a humidified atmosphere containing 95% O2 and 5% CO2. Half of the medium was replaced with fresh medium at 2–3 days. All culture reagents were purchased were from GIBCO (Carlsbad, CA, U.S.A.). The cells were cultured till the 6th day, then treated with high glucose Neurobasal medium (contained 100 mM glucose) for 24 h as the high glucose group (Glu) or cultured with the normal Neurobasal medium (contained basal 25 mM glucose) as the control group (Con). The concentration was chosen based on our previous study . Meanwhile, to further explore the potential molecular mechanism, relevant chemical inhibitors 4-PBA, SP600125, and bafilomycin A1 (Baf A1) were purchased from Sigma-Aldrich Corp. (St. Louis, MO, U.S.A.) and used.
Open-field test was performed to detect the emotional and locomotor activity impairments. As previously described , the open-field device consists of an opaque acrylic container, a square arena with dimension of 40 × 40 × 40 cm divided into 25 square marks measuring 8 × 8 cm, and a video camera, which was fixed 1 m above the area to track the mice’ movement. The computerized tracking system was used to analyze the images and measure the distance of the mice’ movements. Each mouse was released in the center of the arena. Activity was measured as the total distance (meters) the mouse traveled in 10 min. At the end of each test, the arena was carefully cleaned with 75% alcohol to avoid the presence of olfactory cues.
Morris Water Maze (MWM) test
The Morris Water Maze test was performed as previously described [21–23]. A round pool (diameter, 150 cm; depth, 50 cm) was filled with warm (24°C) opaque water to a height of 1.5 cm above the top of the movable clear 15-cm-diameter platform in the third quadrant. A video tracking system recorded the swimming motions of animals, and the data were analyzed using motion-detection software for the Morris Water Maze (Actimetrics Software, Evanston, IL, U.S.A.). After every trial, each mouse was wiped dry and kept warm before returning to its regular cage, where it had free access to food.
Place trials were performed for 4 days to determine the mice’ ability to obtain spatial information. A dark black curtain surrounded the pool to prevent confounding visual cues. All mice received four trials per day in each of the four quadrants of the swimming pool. On each trial, mice were placed in a fixed position into the swimming pool facing the wall. They were allowed 60 s to find the platform in the third quadrant upon which they sat for 20 s before being removed from the pool. If a mouse did not find the platform within 60 s, the mouse was gently guided to the platform and allowed to remain there for 20 s. For all training trials, swimming speed and the time to reach the platform (escape latency) were recorded. The less time it took a mouse to reach the platform, the better the learning ability. We took the average of four trials as the escape latency each day. Probe trials were conducted immediately after the 4-day period to evaluate memory retention capabilities. The probe trials involved removing the submerged platform in the third quadrant from the pool and allowing the rats to swim for 60 s in any of the four quadrants of the swimming pool. The number of original platform crossings and time spent in each quadrant were recorded.
Hematoxylin and eosin (HE) staining
After behavior study, all of the mice were anesthetized by intraperitoneal injection of overdose Nembutal and killed. Seven out of fifteen mice in each group were killed to harvest the brain for morphological analysis, while the rest of animals were subjected to protein analysis.
The brain was preserved in Bouin fixative for 24 h, washed with 10 vol ddH2O, and stored in 70% EtOH. Then the brain was transected perpendicular to the long axis of the disc, processed, and paraffin embedded. Sections (5-μm thick) were cut and mounted onto silane-coated glass slides. The tissue slides were processed according to a standard HE staining technique. The morphology of neurons in hippocampus and cerebral cortical was observed under light microscopy (BX50, Olympus Corporation, Tokyo, Japan).
Bouin-fixed brain from each group was processed, paraffin embedded, and coronal 5 μm sections were prepared and subjected to Nissl staining. These sections were examined by an observer blinded to the group assignment of the sections. Three sections from each animal were selected at random and the number of positive cells (neurons) in hippocampus and cerebral cortical under a high magnification field (×400) in five visual fields/per section was counted. The densities of neurons were measured quantitatively using Image-Pro Plus 6.0 software (Media Cybernetics, Silver Spring, MD, U.S.A.).
Double immunofluorescence labeling
Bouin-fixed brain from each group was processed, paraffin embedded, and sectioned at 5 μm for GRP78, CHOP, and LC3 immunofluorescence labeling. Sections were dewaxed in xylene and hydrated using a graded series of ethanol. Antigenic retrieval was performed by immersing mounted tissue sections in 0.01 mM sodium citrate buffer (pH 6.0) and heating in an autoclave (121°C) for 5 min. Deparaffinized sections were blocked for 1 h in normal goat serum, followed by treatment with either rabbit anti-GRP78 polyclonal antibody (1:100; Abcam, Cambridge, U.K.), rabbit anti-CHOP polyclonal antibody (1:100; Abcam, Cambridge, U.K.) or rabbit anti-LC3 monoclonal antibody (1:200; Cell Signaling Technology, Beverly, MA, U.S.A.) with mouse anti-neuron-specific nuclear protein (NeuN) polyclonal antibody (1:100; Millipore, Billerica, MA, U.S.A.) overnight at 4°C. Subsequently, the tissue slides were incubated with fluorescein (1:100; FITC) and rhodamine (1:200; TRITC) conjugated F(ab’)2 secondary antibodies (Jackson Immuno Research Laboratories Inc, West Grove, PA, U.S.A.) for 1 h at room temperature. The sections were then treated with 4′,6-diamidino-2-phenylindole (DAPI) for 5 min and analyzed using an Olympus BX51 fluorescence microscope (Olympus Optical, Tokyo, Japan).
Transmission electron microscopy
For transmission electron microscopy (TEM) analysis, the hippocampal CA1 region were dissected into 1 mm3 pieces for good penetration of the fixative, and then immersed in 2.5% glutaraldehyde overnight at 4°C. The samples were postfixed in 1% osmium tetroxide for 1 h and dehydrated with ascending grades of alcohol. The tissue block was then infiltrated and embedded in Spurr resin. Ultrathin sections were obtained using Leica EM UC7 (Leica, Wetzetlar, Germany), and stained with uranyl acetate and lead citrate. Ultrastructural changes of synapse and mitochondria were observed and photographed under a transmission electron microscope (Hitachi Model H-7650, Tokyo, Japan). Electron microscope photographs were analyzed using Image-Pro Plus 6.0 software (Media Cybernetics, Silver Spring, MD, U.S.A.). The number of synapse and the area of the postsynaptic density were calculated as described previously .
Western blot analysis
Hippocampus tissues were harvested and were homogenized on ice in the presence of protease and phosphatase inhibitors. Homogenates were centrifuged at 12000 g at 4°C for 15 min. Protein concentration in supernatants was quantified by the BCA method using bovine serum albumin as the standard. Proteins were analyzed by 10–12% SDS/PAGE and transferred to PVDF membranes that were incubated in 5% non-fat milk at room temperature for 1 h, followed by incubated with appropriate primary and secondary antibodies (Cell Signaling, Beverly, MA, U.S.A. or Abcam, Cambridge, U.K.). Membranes were washed and proteins detected by enhanced chemiluminescence using a LAS-4000 lumino-image analyzer (Fuji Film, Tokyo, Japan). Bands were digitally scanned and analyzed using ImageJ software (NIH Image, National Institutes of Health, Bethesda, MD, U.S.A.).
All data were calculated as means ± SD, and checked using the Kolmogorov–Smirnov (KS) test before further analysis. Statistical significance between two datasets was assessed using the Student’s t test. Multiple groups were compared using one-way ANOVA followed by Tukey multiple comparison testing. Specifically, behavioral studies (place trials of MWM) were analyzed by a two-way repeated measures ANOVA (treatment as between-groups and time as repeated measures factors) followed by Bonferroni multiple comparison testing. A P value of <0.05 was considered statistically significant. All statistical tests were performed using GraphPad Prism Version 6.0 (GraphPad Prism Software, Inc. CA, U.S.A.).
Successful induction of diabetic animal model
Three days after STZ or citrate buffer injection, blood glucose levels of the STZ-injected mice were about three folds compared with those of con mice (P<0.05, Table 1), indicating that a mouse model of Type 1 diabetes induced by STZ was established successfully. In addition, during the experiment, diabetic mice exhibited reduced activity and body weight, increased fluid and food intake, damp padding material, poor mental state, and messy dull hair compared with the control mice.
|.||Blood glucose before injection (mmol/l) .||Blood glucose 72 h after injection (mmol/l) .|
|con group||6.32 ± 0.97||6.30 ± 1.38|
|STZ group||6.28 ± 0.86||20.29 ± 3.13*|
|.||Blood glucose before injection (mmol/l) .||Blood glucose 72 h after injection (mmol/l) .|
|con group||6.32 ± 0.97||6.30 ± 1.38|
|STZ group||6.28 ± 0.86||20.29 ± 3.13*|
Data are expressed as the mean ± SD values (n=30). *P<0.05 vs con group.
Hippocampal neurons of diabetic mice are damaged
Representative graphs of neuronal morphology of the two groups were shown in Figure 1A. It revealed that obviously pathological changes were observed in the neurons of hippocampal CA1 region in diabetic mice, as manifested by irregular arrangement, widened intercellular space, reduced cell volume with condensation of nucleus, and blurring of nucleolus. Moreover, a significantly decreased neuronal density in the hippocampal CA1 region was also detected in DM mice compared with Con mice (P<0.05, Figure 1B and C). However, no significant changes were observed in neuronal morphology or neuronal density in the region of cerebral cortex between the two groups. Additionally, we examined the protein levels of cleaved caspase-3, bax, and bcl-2 (Figure 1D), which are well-recognized indicators of apoptosis. The results revealed that the expressions of apoptosis-related proteins were obviously elevated in DM mice, as indicated by reduced level of bcl-2 and increased expressions of cleaved caspase-3 and bax (P<0.01, Figure 1E–G).
Hippocampal neurons are damaged in diabetic mice
Ultrastructure of synapse and mitochondria changes dynamically in diabetic mice
As shown in Figure 2A, TEM revealed that chemical synapses of the mice in the Con group were normal with many synaptic vesicles containing neurotransmitters, centralized postsynaptic densities, and an inerratic synaptic cleft. While in the DM group, the number of the synapse was decreased remarkably and the structure of synapse was irregular compared with the Con group, with reduced synaptic vesicles, decreased postsynaptic region, deformed anterior region, a decreased postsynaptic density (PSD) area, and a widened synaptic cleft (P<0.05, Figure 2B and C). Meanwhile, as molecular markers of synaptic function, the protein expressions of synaptophysin and PSD95 were decreased significantly in the mice of DM group compared with that of Con group (P<0.01, Figure 2D–F). Furthermore, in the DM group, mitochondrial crista of some neurons disappeared and the vacuoles in the swollen mitochondrial matrix increased, which is sometimes assumed to be the characteristics of both apoptosis and neurodegeneration (Figure 2G).
Ultrastructure of synapse and mitochondria are impaired in diabetic mice
Autophagy is activated in diabetic mice
The results of Western blotting revealed that autophagy was induced in DM mice, with enhanced level of autophagic marker microtubule-associated protein 1 light chain 3 (LC3)-II and up-regulated expression of autophagy-associated protein beclin 1 (P<0.05, Figure 3A–C). And the decrease in p62 protein expression, an adaptor protein which serves as an autophagy receptor targeting ubiquitin proteins to autophagosomes for degradation, further indicated an enhanced autophagic flux (P<0.01, Figure 3D). Furthermore, TEM revealed that the number of autophagic vacuoles was increased in DM mice, which were identified as double-membrane structures with engulfed damaged organelles (P<0.05, Figure 3E and F). Double immunofluorescence labeling of LC3 and general neuronal marker NeuN was conducted to further identify autophagy. A higher magnification indicated that the LC3/NeuN/DAPI double-labeled neurons exhibited green punctate LC3 dots in the cytoplasm of red neurons with a blue nucleus (Figure 3G). The result showed that the percentage of colocalized LC3 and NeuN cells to total NeuN cells in the hippocampal CA1 region of DM mice increased significantly compared with that of Con mice (P<0.05, Figure 4H).
Autophagy is promoted in diabetic mice
ER stress is activated in diabetic mice
ER stress pathway is triggered in diabetic mice and contributes to the induction of apoptosis and autophagy in hippocampal primary neurons
As shown in Figure 4, the expression level of ER stress markers GRP78 was significantly enriched in DM mice (P<0.05, Figure 4A and B), as well as phosphorylation of PERK and IRE1α (P<0.05, Figure 4C and D). And as important components of the ER stress-mediated apoptosis pathway, CHOP and phosphorylation of JNK protein levels of DM group were significantly increased in contrast with that of the Con group (P<0.01, Figure 4E and F). Moreover, double immunofluorescence labeling of GRP78 or CHOP and NeuN results showed that the percentage of colocalized GRP78 and NeuN cells to total NeuN cells in the hippocampal CA1 region of DM mice enhanced remarkably compared with that of the Con mice (P<0.01, Figure 4G and H), as well as the change in CHOP protein (P<0.01, Figure 4I and J).
Furthermore, to assess the potential involvement of ER stress pathway in the activation of autophagy and induction of hippocampal neurons injuries, the inhibitor of ER stress 4-PBA (a type of chemical chaperone with functions of improving protein folding capacity, stabilizing protein conformation, and transporting mutant protein to alleviate ER stress) and a selective JNK inhibitor SP600125 were used in hippocampal primary neurons in vitro. The results showed that both 4-PBA and SP600125 exhibited an inhibitory effect on the high glucose-induced increase in LC3-II expression (P<0.01, Figure 5A and B), indicating that ER stress-associated JNK pathway may be partially involved in regulation of autophagy. In addition, activation of apoptosis-related proteins induced by high glucose treatment was also attenuated by 4-PBA pretreatment (P<0.01, Figure 5C–F), which suggests that ER stress-mediated apoptosis pathways may be responsible for the high glucose-induced neurotoxicity.
Inhibition of ER stress attenuates high glucose-induced hippocampal neurons apoptosis and autophagy
Inhibition of autophagy aggravates high glucose-induced ER stress and apoptosis in hippocampal primary neurons
To determine the role of autophagy in hippocampal neurons injuries, autophagy inhibitor bafilomycin A1 (a specific inhibitor of vacuolar-type H+-ATPase that used to suppress vesicle acidification and inhibit autophagy) was used in hippocampal primary neurons in vitro. As shown in Figure 6, the increased expressions of ER stress markers, GRP78 and CHOP, induced by high glucose were further exacerbated following bafilomycin A1 pretreatment (P<0.01, Figure 6A–C), and the high glucose-induced enhancement in the levels of cleaved caspase-3 was even more pronounced (P<0.01, Figure 6D).
Inhibition of autophagy aggravates high glucose-induced ER stress and apoptosis in hippocampal primary neurons
Attenuation of ER stress protects against diabetes-induced cognitive impairments in mice
The procedure of behavior study was indicated in Figure 7A. The results of the open-field test revealed that no statistical significance was found in the traveled distance among the four groups (Figure 7B). In the place trials, the swimming speeds of the four groups had no significant differences (Figure 7C), which suggested that mice of all groups did not develop any motor impairment. The escape latency of mice in all groups showed a downward trend in the place trials. However, mice in DM group spent significantly more time to find the submerged platform at trial day 3 and trial day 4 than those of the Con group, while pretreatment with 4-PBA mitigated the poor performance of diabetic mice in escape latency (P<0.05, Figure 7D). Additionally, in the probe test, significantly less time spending in the third quadrant where the platform was located and less number of crossings over the former platform location were detected in the animals of DM group, and this was reversed by 4-PBA treatment (P<0.05, Figure 7E and F).
Attenuation of ER stress protects against diabetes-induced cognitive impairments in mice
In the present study, we investigated the role of ER stress and autophagy in the pathogenesis of cognitive dysfunction using STZ-induced diabetic mice. Meanwhile, potential mechanisms were also explored through primary cultured mouse hippocampal neurons. Our results demonstrated that diabetes-induced cognitive decline was accompanied by up-regulation of neuronal apoptosis and structural alteration of synapses at the hippocampal level, which may be attributed to the activation of ER stress signaling pathway. More important, autophagy was promoted by diabetes possibly due to the activation of ER stress-mediated JNK pathway, which may play a protective role in ER stress-induced neuronal damages.
Diabetes mellitus (DM), a common metabolic disease characterized by elevated blood glucose level, is increasing at an alarming rate and has become a global challenge. In addition to leading to micro- and macrovascular complications , diabetes, both in Type 1 diabetes mellitus (T1DM) and T2DM, has been linked to complications in the central nervous system, including cognitive impairments and an increased risk of developing neurodegenerative disorders [26,27].
A growing body of evidence from animal models has demonstrated that diabetes, in different degrees, impaired cognitive function, including learning and memory, psychomotor efficiency and attention, and enhanced the vulnerability to depression and anxiety [28–31]. And clinical data has also indicated that diabetes in midlife was associated with a 19% greater cognitive decline over 20 years compared with those without diabetes . Diabetic patients exhibited deficits in different cognitive domains compared with non-diabetic controls, including reduced performance in information processing speed, impaired memory, attention and executive function, slower mental speed, and reduced mental flexibility [4,33–35]. Furthermore, increasing evidence supports the view that diabetes predisposes to cognitive decline leading to dementia in both animal models and humans with both T1DM and T2DM [36,37]. It is reported that diabetes was significantly associated with an approximately 60% increased risk of dementia . Diabetes doubled the risk of a patient developing Alzheimer’s disease (AD) and vascular dementia , and aggravated learning and memory deficits in AD and vascular dementia [39,40]. With the growing epidemic of diabetes, central nervous system-related complications of diabetes are expected to rise and could have challenging future public health implications.
As one of the chronic complications of diabetes, cognitive decline has attracted wide attention due to its serious and extensive impact on diabetes individuals. In the present study, we evaluated cognitive function using MWM test. The MWM protocol is a spatial discriminative learning and memory model, which is related to the structure and function of hippocampus. The cognitive components reflected by MWM performances involve acquisition of spatial information and retention of spatial location memory . Consistent with previous studies , our results showed that diabetic mice displayed deficits in spatial learning and memory capabilities as indicated by the longer escape latency to find the platform, the fewer times of original platform crossing and the less time spent in the target quadrant in the MWM test. Additionally, there was no significant difference in the traveled distance or swimming speeds, excluding the possible influence of physical or emotional depression or motor disturbances on the cognitive impairments observed in our study. These results provide new preclinical evidence to further confirm this disorder in diabetes. Synaptic plasticity has been regarded as a critical mechanism of learning and memory [43–45]. As expected, the damage in synaptic plasticity was observed in mice with diabetes, as characterized by decreased number of synapse, widened synaptic cleft, as well as reduced thickness of postsynaptic density, all of which contribute to changes in synaptic structural plasticity that is closely related to synaptic function. Meanwhile, mice with diabetes exhibited hippocampal neurons injuries. Taken together, our results indicated that changes of synaptic structural plasticity in the hippocampus mediated by neuron injury may be a potential mechanism which further led to impairments of synaptic function and consequent cognitive decline.
ER stress has been reported to be related to a variety of diseases, such as diabetes [46,47], neurodegenerative disease , ischemia/reperfusion injury , and spinal cord injury . Consistently, our results demonstrated a remarkable increase in the expression levels of ER stress marker GRP78, phosphorylation of PERK and IRE1α, indicating ER stress was triggered that attempt to alleviate glucose toxicity stress. Adaptive ER stress restores cellular function, whereas severe ER stress will trigger apoptosis pathways to induce cell apoptosis. Previous studies have indicated the involvement of ER stress-induced apoptosis in diabetes-induced cognitive deficits [19,51]. Similarly, CHOP and phosphorylation of JNK, markers of ER stress-induced apoptosis, were increased significantly in mice with diabetes. These results suggested that ER stress and ER stress-mediated apoptosis pathways may be contribute to diabetes-induced hippocampal neuronal toxicity, eventually leading to cognitive impairments. To further confirm this idea, ER stress inhibitor 4-PBA was used in STZ-induced diabetic mice in vivo and primary cultured mouse hippocampal neurons in vitro. Inhibition of ER stress significantly ameliorated diabetes-induced cognitive impairments in mice, and obviously alleviated high glucose-induced enhancement in the expression of cleaved caspase-3 and bax, and reduction in the level of bcl-2, which further confirmed the involvement of ER stress signaling pathway in the neuronal apoptosis and subsequent cognitive decline mediated by diabetes.
Interestingly, previous studies have demonstrated the essential role of autophagy in neuronal cell health and survival, and neurodegeneration [52,53]. Activation of autophagy alleviates neurodegeneration and vice versa , suggesting autophagy may play an important neuroprotective function in these settings. Additionally, the involvement of autophagy in the etiology of diabetes has been reported, which appears to play a protective role . Nevertheless, scientific evidence for the role of autophagy in cognition of DM is insufficient. Therefore, our study was conducted to investigate this issue. The results showed a remarkable enhancement in the level of autophagy. Furthermore, in agreement with the protective role in DM, inhibition of autophagy strongly correlated with exacerbated ER stress and ER-stress-induced apoptosis induced by high glucose in hippocampal primary neurons in vitro. These results suggest a potential neuroprotective function of autophagy in response to ER stress-associated neuronal injury in diabetes. In this scenario, autophagy modulation could be a possible preventive strategy to consider. However, studies have demonstrated the dual role of autophagy in deciding cell fate, to survival or death . Although many studies suggested the prosurvival mechanism, evidence also indicate the role of autophagy in the cell death, named autophagic cell death . It has been showed that at different stages of AD, alteration of autophagy exhibited two faces, amelioration or exacerbation neurodegeneration. These results indicated that the effect of autophagy could be different depending on types of tissues or cells, animal or disease model, detection stage, and technical or analytical method.
Autophagy and apoptosis are two important molecular processes that attempt to resolve ER stress or initiate programmed cell death to maintain intracellular environment homeostasis . It is clear that apoptosis mediated by ER stress is related to diabetes-induced cognitive decline [19,51]; however, autophagy has not received adequate attention yet. Autophagy that induced due to homeostasis perturbation of ER is considered as a cytoprotective mechanism in response to ER stress [57,58]. Our results displayed activation of ER stress, as expected, up-regulated level of autophagy was also detected in our study. In addition, the IREI–tumor necrosis factor receptor associated factor 2–JNK pathway  and PERK–eukaryotic translation initiation factor 2 subunit α signaling  are two important pathways linking ER stress to autophagy. To further determine the possible signaling pathway involved in the diabetes-induced autophagy, we used inhibitors of ER stress and JNK in hippocampal primary neurons in vitro. Consistently, our results revealed that autophagy induced by diabetes was blocked both by 4-PBA and SP600125, indicating that the ER stress-mediated JNK pathway did correlate with the activation of autophagy in diabetes.
The findings provided here are translationally important in that they determined whether the ER stress/autophagy pathway involves in neuronal apoptosis and cognitive decline induced by diabetes in animals. However, there are several limitations in our current work. First, DM is a complex metabolic disorder accompanied with various symptoms, although there are many methods already exist to establish the experimental animal models of diabetes with their own characters individually, STZ-induced diabetes model is a classical and stable animal model, which indicates that the diabetes model we used may not fully simulate the complex clinical setting of diabetes. However, previous studies have successfully used this experimental diabetes model to investigate cognitive impairments and changes in behavior of diabetic animals [3,61]. Second, a further increase in the level of autophagy might also alleviate learning and memory impairment induced by diabetes, which were not explored in the present study.
In summary, the present study indicates that diabetes induces neuronal injuries and following damage of synaptic structure and function to impair the learning and memory in an experimental mouse model of diabetes, which may be associated with the activation of ER stress pathway. Furthermore, activation of ER stress-mediated JNK pathway may be responsible for the elevated level of autophagy induced by diabetes, which may play a protective role in ER stress-associated cell damages. Our results provide new evidence to further confirm the cognitive impairments in diabetes and suggest the involvement of ER stress and autophagy pathways at the molecular level. More sophisticated studies are needed to focus on this issue, and a better understanding would help to develop more effective treatment or new therapeutic target for this disorder.
Diabetes induces neuronal injuries and following damage of synaptic structure and function to impair the learning and memory in an experimental mouse model of diabetes.
ER stress pathway may be involved in diabetes-mediated neurotoxicity and promote the following cognitive impairments.
Activation of ER stress-mediated JNK pathway may be responsible for the elevated level of autophagy induced by diabetes, which may play a protective role in ER stress-associated cell damages.
F.-J.K. and S.Q. conceived and designed the experiments. F.-J.K., L.-L.M., and J.-J.G. conducted the study. Y.L. and L.-H.X. analyzed the data. F.-J.K. and L.-L.M. wrote the paper. S.Q. revised the manuscript.
This work was supported by National Natural Science Foundation of China (No. 81401633).
The authors declare that there are no competing interests associated with the manuscript.
The authors contributed equally to this work.