Abnormal Lipid Droplets Accumulation Induced Cognitive De cits in Obstructive Sleep Apnea Syndrome Mice via JNK/SREBP/ACC Pathway but Not Through PDP1/PDC Pathway

Dongze Li Binzhou Medical University Yan Yu Binzhou Medical University Na Xu Binzhou Medical University Wanting Li Binzhou Medical University Yanyan Hou Binzhou Medical University Xi Wang Binzhou Medical University Yeying Sun Binzhou Medical University Wenxue Lu Binzhou Medical University Guiwu Qu Binzhou Medical University Changjun Lv Binzhou Medical University Fang Han (  Hanfangtuandui@163.com ) Binzhou Medical University

Nevertheless, the role of LDs accumulation in CIH-induced neuro dysfunction still needs to be determined.
Under physiological conditions, the irreversible oxidative decarboxylation plays a central role in lipid homeostasis. Dephosphorylation of pyruvate dehydrogenase phosphatase 1 (PDP1) boosts the acetylation status of pyruvate dehydrogenase complex E1 subunit (PDHA1) and PDP1, which contributes to the subsequent pyruvate dehydrogenase complex (PDC) activation (Fan et al,2014). Then, activated PDC converts pyruvate to Acetyl-CoA (Walther et al,2017). As an important molecule in the metabolism processes of the human body, Acetyl-CoA is the raw material for de novo lipogenesis (Walther et al,2017). ACC, the rate-controlling enzyme in the pathway of lipogenesis, catalyzes the carboxylation of acetyl-CoA to malonyl-CoA (a highly regulated molecule in fatty acid synthesis) (Hunkeler et al,2018). As the production of CIH, ROS is closely related to variety of physiological processes including lipid metabolism (Prabhakar et al,2012). Abnormal accumulation of ROS could regulate the expression of genes implicated in lipid metabolism, including SREBP, ACC, CPT-1, SCD-1 (Zhao et al,2018). However, the potential mechanisms under abnormal lipid metabolism and its functional implication in diseases remains unknown.
SMND-309 is a major potent component extracted from salvia miltiorrhiza. It can be detected in the rat brain as a novel metabolite of salvianolic acid B after administration (Tian et al,2008). Research has proven that SMND-309 promotes neuron survival through antiapoptotic, anti-in ammatory and antioxidative effects (Yang et al,2010). In this study, SMND-309 was intraperitoneally injected into mice to alleviate CIH-induced neuro damage. This might be helpful to develop a potential drug therapy target for OSAS.

Animals
Six-week-old male wild-type (WT) mice on a C57BL/6 background (Jinan Pengyue Experimental Animal Breeding Co., Ltd., China) were used for this study. All mice were maintained at 23 ± 1°C with a 12 h/12 h light/dark cycle. Food and water were available ad libitum. Before the tests, 40 mice were randomly divided into 4 groups (WT + RA, WT + CIH, SMND-309, SMND-309 + CIH) / (WT + RA, WT + CIH, 3-FP, 3-FP + CIH), and each group had 10 mice. Every mouse was given a unique number. During the whole testing process, the researchers were blinded to the treatment. All the mice were fed and used according to the NIH guidelines, and this study was approved by the Ethics Committees on Animal Experimentation of Binzhou Medical University (Permit no. SCXK20160006).

Cell culture
The neuron cell line HT22 (JNO-02001) was purchased from Guangzhou Jinio Biotechnology Co., Ltd.

CIH exposure
The mice were exposed to CIH/room air (RA) conditions from 8:00 AM-4:00 PM per day for 12 weeks. During the exposure, the mice in the WT + CIH group and the SMND-309 + CIH (3-FP + CIH) group were placed into the CIH chamber (BioSpherix OxyClycler A84, USA), the WT + RA group and the SMND-309 (3-FP) group were placed into the RA chamber (BioSpherix OxyClycler A84, USA). The CIH program was performed as described previously (Li et al,2020). Firstly, the chamber was lled with N 2 for 85-95 s, and the oxygen level was reduced from 21% ± 1% (normal) to 7% ± 1% (hypoxia). The oxygen level was then maintained at 7% ± 1% for 15-20 s. Finally, the level was recovered to 21% ± 1% in 45-50 s and sustained for 15-20 s. A cycle of CIH lasted approximately 180 s, and there were approximately 20 cycles per hour. The oxygen level in the RA chamber was always maintained at 21% ± 1%.

IH exposure
IH was induced by culturing cells in an oxygen control cabinet (Biospherix Oxycycler C42) mounted within an incubator and equipped with oxygen sensor for continuous oxygen level monitoring. A mixture of nitrogen oxygen and 5% CO 2 was infused and oxygen levels consisted of a 30-minute hypoxic period (3% O2), followed by 30 minutes of reoxygenation (21% O 2 ). Actual O 2 saturation was kept at 3% for 10 minutes, at each 1-hour cycle (Dyugovskaya et al,2012). This IH treatment was performed for 3 day. To examine cognitive function impairment due to CIH treatment, the three-chamber social test, Morris water maze (MWM) test and fear conditioning test were performed. The three-chamber social test was conducted as previously described (Moy et al,2004). The chamber used in the test (60 cm L × 26 cm W × 30 cm H) was divided into three sections: center, left and right (20 cm L × 26 cm W). The test mouse had free access to the different sides of the chamber through the door (11 cm L × 11 cm W) on the dividing wall. The test consisted of 3 phases with 20 min intervals. Each phase allowed the test mouse to explore the chamber for 10 min. During the phase, the video system tracked the behavior of the mouse. At phase 1, a test mouse was placed in the center chamber, and two empty wire cages (10 cm D × 14 cm H) were placed into the left and right sides. During phase 2, an unfamiliar (stranger) mouse was placed in the left wire cage, and an empty wire cage was placed on the right side. Another stranger mouse was then placed in the right wire cage in phase 3. Mice were tested from approximately 8:00 AM to 12:00 AM and received CIH/RA treatment from 12:00 AM to 4:30 PM.

MWM
The MWM test was conducted as previously described (Lee et al,2019). The system contained a circular pool (122 cm D × 51 cm H) and an escape platform (12 cm D × 34 cm H) (ZS Dichuang, Beijing, China). The pool was divided into four equal quadrants and lled with 35 cm of deep water. The water was dyed with TiO 2 and kept at 23 ± 1 ℃. The escape platform was positioned 1 cm below the water surface in quadrant two. The same discriminate landmarks were placed around the maze during the test. The test consisted of two stages, including four days of the training stage and one day of the probe trial stage.
The day before the test, all the mice were acclimated to the pool without the platform for 60 s. During the training stage, the test mouse performed four trials for 4 consecutive days. In the rst trial, a mouse was put into the water facing the pool wall at quadrant one. If the mouse found the escape platform in 60 s, the tracker system would be automatically closed; otherwise, the mouse would be guided to the platform and forced to stay on it for 15 s. With a 15 min interval, the mouse received a second trial and started at quadrant two. After trials 3 and 4, the mouse was dried with a towel and placed back into the home cage. At the probe trial stage (day 5), the mouse was put into the water at quadrant four and forced to swim without the platform for 60 s. Several performance parameters of the test mouse were recorded, including the total swimming distance, the number of platform crossings, the duration and the distance traveled in the target quadrant. The tests were performed from 8:00 AM to 12:00 AM. After the tests, the mice were exposed to CIH/RA conditions from 12:00 AM to 4:30 PM.

Fear conditioning test
The equipment consisted of a conditioning chamber with a sound-attenuating wall and a video monitoring system (ZS Dichuang, Beijing, China). An acrylic box (35 cm L × 35 cm W × 35 cm H) and a stainless-steel grid oor were placed in the chamber. The oor was connected to a device to deliver the footshock, and tests were performed as previously described (Shoji et al,2014). On day 1, after acclimating the chamber (black acrylic box with a jasmine smell) for 2 min, the test mouse underwent 30 s of sound (65 dB, 3 kHz) three times (separated by 30-second intervals), and the mouse received a foot shock (0.7 mA) during the last 2 s of each sound. On day 2, the mice were exposed to the same environment (black acrylic box with a jasmine smell) for 5 min with no sound and no foot shock. The fear behavior (freezing) of the mouse, which referred to contextual memory, was recorded. On day 3, after 2 min of acclimatization to the new environment (blue acrylic box with a lemon smell), the mouse was exposed to 30 s of sound (65 dB, 3 kHz) three times (separated by 30-second intervals). During this stage, the fear behavior (freezing), which refers to cued memory, was measured.

Tissue preparation
Five mice from each group were deeply anesthetized by 4% choral hydrate through intraperitoneal injection. They were then perfused with physiological saline via the left ventricle of the heart, and xed with 4% ice-cold paraformaldehyde. The xed brains were embedded in optimal cutting temperature compound, frozen in liquid nitrogen and stored at -80°C. For the other mice, the hippocampi were harvested and stored in liquid nitrogen.

Transmission electron microscopy (TEM)
The hippocampi of each group were xed in 2.5% ice-cold glutaraldehyde in 0.1 M PBS at pH 7.4 and osmium tetroxide. After gradient dehydration with ethyl alcohol and acetone, the hippocampi were embedded in epoxy resin Epon 812. Tissues were then cut into ultrathin sections (50 nm) by a CM1900 microtome (Leica, Germany) and stained with uranyl acetate and lead citrate. Samples were viewed in a JEOL JEM 1010 TEM at 80 kV and captured through an AMT XR-16 mid-mount 16 mega-pixel digital camera (Hu et al,2011).

Hematoxylin and eosin (H&E) staining
The brains of each group were cut into 5 µm sections and stained with H&E. After dehydration with ethanol, the sections were observed by an Invitrogen EVOS M5000 imaging system (Thermo Fisher, USA) (Ballok et al,2004).

Nile red staining
To detected the accumulation of LDs, the frozen slices were washed three times in PBS and incubated with Nile red (1:1000) for 10 min. After washing three times with PBS, the slices were mounted with Vectashield with DAPI (Vector Labs, USA) and imaged with an Invitrogen EVOS M5000 imaging system (Thermo Fisher, USA). The Nile red powder was dissolved in methyl alcohol at a concentration of 1 mg/mL.

Measure the levels of PDC activity, Acetyl-CoA and ROS
Pyruvate Dehydrogenase Activity Assay Kit (MAK183) was purchased from Sigma. Acetyl-Coenzyme A Assay Kit (Sigma, MAK183) was used to determine the Acetyl-CoA level. The contents of ROS in the hippocampus were detected with an Aconitase Activity Assay Kit (Sigma-Aldrich, MAK051) (Liu et al,2015). All the protocols were conducted in accordance with the manufacturer's instructions. The optical density of the wells was detected by the Ultra Multitask Ascent (Epoch, BioTek, USA).

Statistical analysis
All experiments were repeated three times. The experimenter was unaware of the animal's group during experimentation and removed the mice that were in poor condition (they did not like to exercise). SPSS version 24.0 (IBM Crop, Chicago, USA) was used for statistical analysis. Data are displayed as the mean ± SEM. One-way analysis of variance (ANOVA) and a Bonferroni post hoc test were used to evaluate the results of the three-chamber social tests (Xu et al,2019). The results of escape latency in the MWM test were examined through three-way mixed ANOVA (Champagne et al,2002). For the total traveling distance, platform crossing counts, the duration and distance in the target zone and the fear conditioning tests were assessed with two-way ANOVA (Van Dam et al,2000). Other data were assessed with two-way ANOVA. Statistical signi cance was accepted at p < 0.05.

The exposure of CIH
The mice were exposed to RA/CIH condition for 12 weeks (shown in Fig. 1. B). During the last week of the exposure, the mice were intraperitoneal injected with SMND-309 for once daily at 25 mg/kg of body weight. The three-chamber social test, the MWM test and the fear conditioning test were conducted in sequence as soon as the treatment was nished (shown in Fig. 1. A). 3.2. CIH exposure caused the social damage, SMND-309 could attenuate the damage At phase 1, there was no obvious side preference in the 4 groups (shown in Fig. 2. A, D). During phase 2, all the mice preferred to stay with the stranger mouse than at the inanimate side (p < 0.05, shown in Fig. 2. B, E). At phase 3, WT + RA mice (F (2,27) = 55.839, T = 6.827, p < 0.05) and SMND-309 mice (F (2,27) = 71.527, T = 6717, p < 0.05) preferred to stay with the new stranger mouse than the familiar one (shown in Fig. 2. C). However, WT + CIH mice spent similar time at the two sides (F (2,27) = 18.716, p > 0.05) (shown in  Fig. 2. C). The traveling distance of phase 3 showed same pattern with spending time for all groups (shown in Fig. 2. F). These results showed that CIH treatment damaged the social novelty cognition of mice and SMND-309 could alleviate the damage.

CIH-induced pathological changes in the hippocampus
After CIH treatment, severe histological changes have been revealed, neurons are loosely arranged, irregular sized, with fuzzy outline. Meanwhile, the number of neuroblasts (NBs) in the hippocampal subgranular zone (SGZ) was signi cantly decreased (shown in Fig. 3. A). This injury was relieved by SMND-309 treatment (shown in Fig. 3. A). As seen with TEM, the neuron ultrastructure was completed, and the blood-brain barrier (BBB) was thin and intact in the WT hippocampus (shown in Fig. 3. B). However, neurons exhibited cytoplasmic vacuolization and mitochondrial disintegration under CIH conditions. Moreover, a large number of LDs appeared in the damaged neurons (shown in Fig. 3. B). For astrocytes, severe edema could be observed, especially in the area surrounding the BBB (shown in Fig. 3. B).
To determine the damage of NBs in the SGZ by LDs accumulation, the NB marker DCX and the apoptosis marker cleave-caspase 3 was stained with the LDs dye BODIPY (483/503). The number of DCX-positive cells in the DG area was decreased 80% after CIH treatment (p < 0.05) (shown in Fig. 6). In addition, the quantity for cleaved caspase 3-positive cell increased 85% with BODIPY (483/503) spots in the DG area of WT + CIH mice (p < 0.05) (shown in Fig. 6). Notably, apoptosis NB signi cantly reduced in SMND-309 + CIH group than WT + CIH group (p < 0.05) (shown in Fig. 6).

3.7.
Affecting the pathway of PDP1/PDHA1 did not improve the damage caused by CIH exposure The mice were exposed to RA/CIH condition for 12 weeks,and received intraperitoneal injection with 3-FP every 3 days at the last 3 weeks of the exposure(shown in Fig. 7. A). The fear conditioning test were conducted as soon as the treatment was nished (shown in Fig. 7. A).
In the context memory test, WT + CIH mice presented with a lower level of freezing compared with WT + RA group than mice. It should be noted that this change was not improved after treatment with PDHA1 inhibitor 3-FP (shown in Fig. 7. C, *p < 0.05). After 48h retention delay, the cued fear conditioning testing were determined, the freezing level was similar to the trend of context memory test (shown in Fig. 7. D, *p < 0.05). Taken together, these results provided evidence that CIH exposure contributed to the long-term memory dysfunction, and PDHA1 inhibitor 3-FP could not improve the damage.
The H&E stained of hippocampus section in 4 groups were shown in Fig. 7E. Compared to the RA group, severe histological changes had been revealed in CIH group. Hippocampal granular neurons were loosely arranged,some neurons had showed a swollen and vacuolated cytoplasm and the neural progenitor cells in the DG area were signi cantly reduced. Signi cantly, this injury was not relieved by PDHA1 inhibitor treatment.

PDP1/PDHA1 did not signi cantly promote the increase of LDs induced by CIH in hippocampus
The uorescence of Nile red showed that the CIH group and CIH + 3-FP group had sporadic super cial, round or elongated lipid-containing structures in the hippocampus, indicating abnormal accumulation of LDs (shown in Fig. 8. A, B). Real time RT-PCR analysis and Western blot analysis were performed to analyze the expression level of PDP1 and PDHA1. Quantitatively, the gene expression and protein levels of PDP1 and PDHA1 showed no signi cant differences among 4 groups (shown in Fig. 8. C-G). The protein level of phosphorylated PDHA1 (p-PDHA1) (shown in Fig. 8. H), the activity of PDC (shown in Fig. 8. I) and the content of Acetyl-CoA (shown in Fig. 8. J) were tested to further explore the role of PDP1/PDHA1 related lipid synthesis on CIH condition. All these results showed no statistically differences between groups, which suggested CIH exposure almost had no impact on the Acetyl-CoA biosynthesis regulated by PDP1/PDHA1. In vitro experiments supported this conclusion (shown in Fig.  Sp. 1).
Experimental data showed that the PDP1/PDHA1 may not be the main source of the abnormal increased lipids after CIH exposure. Therefore, other reasonable mechanisms of lipids synthesis disorder in neurocytes under stress needed to be interpreted.
3.9. ROS-triggered JNK/SREBP/ACC pathway activation and lipid peroxidation, SMND-309 could decrease To con rm the level of ROS in the hippocampus under CIH conditions, an aconitase activity assay kit was used. The results showed that the activity for aconitase decreased 50% in WT + CIH group (p < 0.05) (shown in Fig. 9. G). Furthermore, the expression of phosphorylated JNK increased nearly 1.5-fold (p < 0.05), and this effect activated SREBP and upregulated the expression of ACC for 3-fold (p < 0.05) (shown in Fig. 9. A-F). Treated with SMND-309 eliminated ROS (p < 0.05) (shown in Fig. 9. G) and inhibited the activation of the JNK/SREBP/ACC pathway (shown in Fig. 9. A-F).

Discussion
As a common sleep disorder disease, OSAS is characterized by sudden pauses of breathing during sleep (Li et al,2018). Resulting from repeated obstructions of the pharyngeal airway, CIH is a cardinal feature of OSAS, which induces degrease of cognitive performance. Our research found that the ability of social memory and spatial memory was damaged in C57BL/6 mouse after CIH treatment. Further study discovered that the hippocampal neurocytes of those mice were severely injured. It is thought that hippocampus is crucial for encoding new memory (Suh et al,2011). Therefore, the injury of hippocampus by CIH nally gave rise to behavioral de cits. Until now, researchers consider that the mechanism of those neuro damage is related to the in ammation and the oxidative stress. In response to CIH, immune Recently, LDs accumulation is closely related to a variety of human diseases (Max eld & Tabas,2005). In peripheral system, LDs-associated protein cell death-inducing DFF45-like effector (CIDE), which is crucial for the formation and fusion of LDs, regulates the occurrence of type II diabetes (Dahlman et al,2005). The depletion of perilipin 2 prevents hepatic steatosis via downregulating triglyceride synthesis and LDs accumulation (Carr et al,2014). In nervous system, LDs accumulation enhances the formation of oligomeric α-synuclein, a major component of the pathological hallmarks in Parkinson's disease (Ruipérez et al,2010). Knocking out HSP-related genes leads to LDs accumulation in neurons (Klemm et al,2013). LDs accumulation also disrupts energy homeostasis (Konige et al,2014), impairs the folding and clearance of proteins in neurons (Inloes et al,2018), and breaks neuron-glia metabolic coupling (Schmitt et al,2014). All these studies have shown that LDs accumulation is closely related to neurodegeneration diseases. Contradictorily, some studies support that LDs accumulation is helpful for neural protection. In the glial niche of Drosophila larvae, LDs accumulation keeps neural stem cells away from oxidative damage (Moldavski et al,2015). Some fatty acids are vulnerable to peroxidation, and they are diverted into LDs to protect from ROS under hypoxia condition (Welte & Gould,2017). Therefore, exploring the molecular mechanisms of altered lipid metabolism in brain injury, will help to reveal the cause of neurodegenerative changes, including CIH-induced cognitive dysfunction. PDP1/PDHA1 pathway is an essential regulatory for de novo lipid synthesis. PDHA1 is activated by dephosphorylation of PDP1 and phosphorylation of pyruvate dehydrogenase kinases (PDK) (Shan et al,2014). PDHA1 is one of the most important components of the PDC (Zhong et al,2017), which oxidates pyruvate to Acetyl-CoA. As a major and central precursor in metabolism, Acetyl-CoA candidates in the synthesis and decomposition of biomacromolecules, especially for lipid biosynthesis (Kuerschner et al,2008). It has been reported that PDP1/PDHA1 was related to a variety of diseases by affecting lipid metabolism. Inactivation of PDHA1 suppresses tumourigenesis by decreasing Acetyl-CoA levels in prostate cancer. However, knocking out PDK4 alleviates the hepatic steatosis by regulating the activity of PDC in nonalcoholic steatohepatitis mouse models (Zhang et al,2018). In this study, experimental data showed that the activity of PDC and the production of Acetyl-CoA did not noticeable change after CIH exposure, which suggested the de novo lipid synthesis regulated by PDP1/PDHA1 might not be the main source of the abnormal increased lipids and LDs after CIH exposure. Therefore, the relevant mechanism still needs to be further studied.
ROS production and oxidative stress participate in neuro injuries and neurodegeneration. In Parkinson disease, ROS induces missense mutation by damaging DNA and inevitably causing neural cell damage (Pignataro et al,2017). ROS also evokes Alzheimer's disease through active NLRP3, which promotes IL-1βmediated in ammation (Pignataro et al,2017). Recently, studies found that ROS is capable to promote lipid synthesis and LDs accumulation (Liu et al,2015). Increased level of ROS exerts harmful effects by causing oxidative damage to biological macromolecules and disrupting various signaling pathways including the lipid metabolism (Fransen et al,2012). In the development of fatty liver, redox cellular state especially the high level of ROS activates lipid biosynthesis gene SREBP and speeds up the disease process (Pan et al,2017). Reports show that ROS triggers SREBP activity in fruit y neurons leading to LDs accumulation (Liu et al,2015). SREBP regulates the expression of several genes, such as ACC and fatty acid synthase (FAS), which is the candidate of cellular fatty acid synthesis (Nogalska et al,2005).
ACC is a rate-limiting enzyme in de novo fatty acid synthesis, catalyzing ATP-dependent carboxylation of Acetyl-CoA to form malonyl-CoA (an intermediate in fatty acid biosynthesis) (Hunkeler et al,2018). We found that high level of ROS triggered JNK/SREBP/ACC pathway activated in neurons, thus more Acetyl-CoA were converted into fatty acid after CIH exposure. The excessive increase of lipid synthesis promoted abnormal LDs accumulation, which severely injured neurocytes in hippocampus. In addition, lipid peroxidation further aggravated neuro damage.
SMND-309 is a degradation production of Salvia miltiorrhiza, which has been widely used for neuroprotection (Su et al,2015). SMND-309 inhibits apoptosis by upregulating the ratio of Bcl-2/Bax (Yang et al,2010) and promotes neuron survival by increasing the content of brain-derived neurotrophic factor via activating the phosphatidylinositol 3-kinase/Akt/cAMP-response element-binding (CREB) signaling pathway (Wang et al,2016). Consistent with the results of previous studies, SMND-309 treatment improved the behavioral performance of CIH mice by reducing the accumulation of LDs in neurocytes of the DG area. These ndings might be helpful to provide a novel potential neuroprotective therapy.
In this study, CIH-induced hippocampal damaging was triggering by LDs accumulation in NBs, neurons and glia cells. The generation of LDs could be regulated via JNK/SREBP/ACC pathway. And these damages were alleviated by SMND-309 treatment. Until now, the role of LDs in neurocyte is controversial.
The mechanism of lipid synthesis disorder and LDs abnormal accumulation is also unclear. All these questions need to be further investigated.

Conclusion
Present study found that cognitive function was severely damaged in CIH mice because of neurocyte injuries in the hippocampal DG area. After CIH exposure, the expression of PDP1/PDHA1, the activity of PDC and the level of cellular Acetyl-CoA have barely changed. Noticeably, ROS triggered JNK/SREBP/ACC pathway, and then more Acetyl-CoA were converted into fat acids. The redundant lipids led to aberrant LDs accumulation, which contributed to the neuro injuries. What's more, lipid peroxidation as a result of excessive ROS also promoted the damage in neurocytes. Remarkably, all of these damages induced by lipid metabolic disorders were relieved by SMND-309 treatment.  The process of research and the structure of SMND-309. (A) The study work ow. After 7 days of acclimation, the mice were exposed to the CIH condition for 12 weeks, and they were treated with SMND-309 at the last week. Behavioral tests were conducted as soon as CIH exposure nished. (B) The CIH treatment program. The red line represents the set point. The blue line represents the actual oxygen level. (C) The structure of SMND-309.    Affecting the pathway of PDP1/PDHA1 did not improve the damage caused by CIH exposure (A) The study work ow. After 7 days of acclimation, the mice were exposed to the CIH condition for 12 weeks, and they were treated with 3-FP at the last 3 weeks. Fear test were conducted as soon as CIH exposure