Role of Adiponectin‐Notch pathway in cognitive dysfunction associated with depression and in the therapeutic effect of physical exercise

Abstract A substantial percentage of late‐life depression patients also have an cognitive impairment, which severely affects the life quality, while the co‐occurring mechanisms are still unclear. Physical exercise can ameliorate both depressive behaviors and cognitive dysfunction, but the molecular mechanisms underlying its beneficial effects remain elusive. In this study, we uncover a novel adipose tissue to hippocampus crosstalk mediated by Adiponectin‐Notch pathway, with an impact on hippocampal neurogenesis and cognitive function. Adiponectin, an adipocyte‐derived hormone, could activate Notch signaling in the hippocampus through upregulating ADAM10 and Notch1, two key molecules in the Notch signaling. Chronic stress inhibits the Adiponectin‐Notch pathway and induces impaired hippocampal neurogenesis and cognitive dysfunction, which can be rescued by AdipoRon and running. Inhibition Notch signaling by DAPT mimics the adverse effects of chronic stress on hippocampal neurogenesis and cognitive function. Adiponectin knockout mice display depressive‐like behaviors, associated with inhibited Notch signaling, impaired hippocampal neurogenesis and cognitive dysfunction. Physical exercise could activate Adiponectin‐Notch pathway, and improve hippocampal neurogenesis and cognitive function, while deleting adiponectin gene or inhibiting Notch signaling blocks its beneficial effects. Together, our data not only suggest that Adiponectin‐Notch pathway is involved in the pathogenesis of cognitive dysfunction associated with depression, but also contributes to the therapeutic effect of physical exercise. This work helps to decipher the etiology of cognitive impairment associated with depression and hence will provide a potential innovative therapeutic target for these patients.


| INTRODUC TI ON
Cognitive impairment is prevalent in late-life depression and often persists even after remission of mood symptoms (Culpepper et al., 2017;Morimoto et al., 2015). The occurrence of depression in mild cognitive impairment (MCI) accelerates the progression to dementia (Rosenberg et al., 2013). However, little is known about the joint or individual mechanisms of co-occurring depression and cognitive impairment. Previous studies have reported that physical exercise ameliorates depressive behaviors, enhances hippocampal neurogenesis, and improves hippocampal-dependent learning and memory (Duzel et al., 2016;Yau et al., 2014). However, the mechanisms that mediate these effects of physical exercise remain largely unknown.
Moreover, it is unclear which components of exercise programs are therapeutic.
Adiponectin (APN), a hormone secreted predominantly by adipocytes and playing critical roles in body energy homeostasis (Katsimpardi et al., 2020), decreases under chronic stress (Guo et al., 2017) and increases after exercise (Yau et al., 2014) in the circulation. Adiponectin could cross the blood-brain barrier (Neumeier et al., 2007), exert neuroprotective and antidepressant properties through binding its receptors (Thundyil et al., 2012), AdipoR1, and AdipoR2, which are expressed in many brain regions. It has recently been found that adiponectin mimics many of the ameliorative effects of physical exercise on metabolism, hippocampal neurogenesis, depression and cognitive dysfunction (Greenhill, 2015;Liu et al., 2020;Nicolas et al., 2015;Yau et al., 2014). Adiponectin deficiency has been found to induce decreased hippocampal neurogenesis and cognitive dysfunction, and increase susceptibility to developing depressive behaviors under stress (Liu et al., 2012;Ng et al., 2016;Zhang et al., 2016). These studies indicate that adiponectin may be a candidate molecule involved in both depression and cognitive impairment induced by stress and may also be a therapeutic component of exercise programs.
In this study, we showed that the modulation of adiponectin mediates the beneficial effects of physical exercise and adverse effects Morris water maze of chronic restraint stress on hippocampal neurogenesis and cognitive functions. Furthermore, we discovered a novel mechanism that adi-  (Thundyil et al., 2012) and exerts similar biologic activity as full-length adiponectin after release by the enzyme leukocyte elastase (Fruebis et al., 2001;Waki et al., 2005). In our study, the exon 3 of adiponectin gene, containing 521bp coding sequence for globular domain and part of collagenous domain, was knocked out using CRISPR/Cas9 technology ( Figure 1a). As shown in Figure 2b, we didn't detect adiponectin in the serum of the knockout Previous studies have shown that adiponectin deficiency in middle-aged mice leads to spatial learning and memory impairments, in which the used APN-KO mice lines are different from ours (Bloemer et al., 2019;Ng et al., 2016  showing that adiponectin deficiency also didn't affect the depressive level in middle-aged mice (Figure 1f,g). The results of the Y-maze test showed that the total percentage of correct spontaneous alterations percentage (SAP) did not differ significantly between APN-KO mice and WT littermates (Figure 1h). In the NOR test, APN-KO mice spent less time on novel object investigation than WT littermates ( Figure 1i). The results of the Morris water maze test showed that the escape latency of APN-KO mice was significantly increased compared with that of WT littermates on day 3-day 5 (Figure 1j, left) and that APN-KO mice spent less time in the target quadrant than WT littermates on the test day ( Figure 1j, lower middle). The numbers of target annulus crossovers revealed a decreasing but nonsignificant trend for APN-KO mice compared with WT littermates (Figure 1j, right). Consistent with previous reports, our 12-month-old APN-KO mice also displayed cognitive dysfunction, as evaluated by the NOR and Morris water maze tests.

| Adiponectin deficiency leads to attenuated hippocampal neurogenesis in middle-aged mice
Studies have indicated that hippocampal neurogenesis plays a key role in learning and memory (Alam et al., 2018  Collectively, these results suggested that adiponectin was required for the beneficial effects of physical exercise on neurogenesis and cognition in middle-aged mice.

| Decreased serum adiponectin level associated with impairments in hippocampal neurogenesis and cognitive function in aged mice
As cognitive impairment and dementia are age-related, we next determine whether adiponectin level was correlated with cognitive function in aged mice. The serum adiponectin level was significantly decreased in aged mice (24 months), but not in middle-aged mice (12 months;

| Adiponectin is required for physical exerciseinduced activation of the Notch signaling pathway in the hippocampus
We next explored the molecular mechanisms underlying the decreased neurogenesis in middle-aged APN-KO mice and adiponectin-induced neurogenesis after running. Notch signaling plays an important role in adult hippocampal neurogenesis (Ables et al., 2010;Breunig et al., 2007), but there is no information on whether adiponectin can regulate Notch signaling. A previous report has shown that Osmotin, plant homolog of adiponectin, can increase the expression of ADAM10 and ADAM17, two ratelimiting S2 enzymes for Notch cleavage, in the hippocampus of APP/PS1 mice (Shah et al., 2017), providing an indirect clue to F I G U R E 3 Adiponectin decreases in aged mice displaying decreased hippocampal Notch signaling, impaired neurogenesis, and cognitive dysfunction. (a) Representative immunoblots and quantification of hippocampal (including Notch1, NICD, ADAM10, ADAM17) and serum (adiponectin) protein levels. 2 months, n = 6; 12 months, n = 7; 24 months, n = 7. *p < 0.05, **p < 0.01, ***p < 0.001.

| The Notch signaling mediates the adiponectin-dependent beneficial effects of physical exercise on hippocampal neurogenesis and cognitive function
We next determined whether the Notch signaling was involved in adiponectin-dependent improvements in hippocampal neuro-

| Adiponectin-Notch pathway was involved in both cognitive dysfunction associated with depression and the therapeutic effect of physical exercise
Given the importance of Adiponectin-Notch pathway in hippocampal neurogenesis and cognitive function under basal conditions and during physical exercise, we next assessed whether Adiponectin-Notch pathway was also involved in cognitive dysfunction associated with depression induced by chronic stress in middle-aged mice.
We thus first determined whether Adiponectin-Notch pathway was inhibited in mice with chronic stress-induced depression. In our study, chronic restraint stress-induced depressive-like behaviors, as evaluated by the sucrose preference test (SPT; Figure 6b) and Immunofluorescence results showed that chronic restraint stress in-

| DISCUSS ION
Dementia and depression, both common disorders in the elderly, impact the quality of life for patients and relatives and involve substantial health-care service and social benefit costs (Barnes et al., 2006;Leonard, 2007). The relationship between depression and dementia is complex with depression having been reported to be both a risk factor and a prodrome for Alzheimer's disease and other dementia's, and also be a common complication of dementia at all stages (Bennett & Thomas, 2014;Panza et al., 2010). The mechanisms of comorbidity of these diseases are still unclear. Research drawing more confident conclusions about the underlying neurobiologic pathways, may pave the way for more effective treatments of both depression and dementia. In this study, we found that adiponectin deficiency is associated with decreased hippocampal neurogenesis and cognitive impairment in both middle-aged APN-KO mice and middle-aged depression mice models. Previous studies showed that adiponectin deficiency in middle-aged mice leads to learning and memory impairments (Bloemer et al., 2019;Ng et al., 2016), while we revealed that Adiponectin level was correlated with both depressive behaviors and cognitive function in middle-aged depression mice model induced by chronic stress, suggesting Adiponectin was a potential candidate responsible for both cognitive impairment and depression in elderly. Furthermore, our data suggested that adiponectin was also required for physical exercise-induced improvements in cognitive function.
Impairment in hippocampal neurogenesis is linked to cognitive dysfunction in both major depressive disorder (MDD) and Alzheimer's disease (AD; Berger et al., 2020;Clelland et al., 2009).

Previous studies have suggested the indispensable role of hippocam-
pal neurogenesis in hippocampus-dependent learning and memory learning (Thuret et al., 2009). A previous report shows that adiponectin knockout doesn't influence basal hippocampal neurogenesis (Yau et al., 2014). However, there is also other study that shows adiponectin deficiency reduces hippocampal neurogenesis, which may be due to the difference in the APN-KO mice lines (Zhang et al., 2016 1, 3, and 7). These observations suggest that impaired hippocampal neurogenesis regulated by adiponectin may be the pathogenesis of cognitive dysfunction associated with depression.
Adiponectin, an adipose-specific cytokine, could cross the bloodbrain barrier (BBB) from the blood into cerebrospinal fluid (Neumeier et al., 2007). Plenty of studies have demonstrated the beneficial effects of adiponectin on adult neurogenesis (Nicolas et al., 2015;Yau et al., 2018;Zhang et al., 2016) and cognitive function (De Franciscis et al., 2017;Rizzo et al., 2020). However, the mechanism underlying those effects of adiponectin has not been fully elucidated.
In this study, we uncovered a novel pathway, Adiponectin-Notch pathway, which mediated the interaction between adipose and brain. Adiponectin could increase the expression of two key molecules in the Notch pathway, ADAM10 and Notch1, which exerted the beneficial effects on hippocampal neurogenesis and cognitive function (Figures 4 and 5). Moreover, Notch signaling decreased in the hippocampus of both middle-aged APN-KO mice and aged mice which displayed impaired hippocampal neurogenesis and cognitive dysfunction (Figures 1-3), and inhibition Notch signaling by DAPT blocked the beneficial effects of AdipoRon on hippocampal neurogenesis and cognitive function (Figures 5 and 7). Collectively, our results suggest that Notch signaling mediates the effect of adiponectin on hippocampal neurogenesis and cognitive function.
Physical exercise is considered an effective therapeutic alternative to improve cognition in patients suffering from MDD (Olson et al., 2017) or AD (Jia et al., 2019). We expected to find whether Adiponectin-Notch not only is involved in pathophysiology of cognitive impairment associated with depression, but also contributes to the therapeutic effect of physical exercise on cognitive dysfunction. In our study, physical exercise enhanced Adiponectin-Notch pathway (Figure 2b,c), increased hippocampal neurogenesis (Figures 2A and 5B), and improved learning and memory ability (Figures 1i,j and 5e,f), showing a correlation between activation of Adiponectin-Notch pathway and improved cognition by physical exercise. These results were consistent with the previous report (Yau et al., 2014). Furthermore, our data showed that physical exercise could reverse the decreased Adiponectin-Notch signaling induced by chronic restraint stress (Figure 6), and ameliorate the F I G U R E 6 Physical exercise reverses the decreased Notch signaling pathway in the hippocampus induced by chronic restraint stress. Furthermore, we revealed the molecular mechanism by which adiponectin activated the Notch signaling in hippocampus. Limited studies give us some indirect clues that PPARα and JNK are involved in the expression regulation of ADAM10 (Corbett et al., 2015) and Notch1 (Xie et al., 2017) respectively.
In this study, we demonstrated that Adiponectin upregulated the expression of ADAM10 and Notch1 through PPARα and JNK respectively.
In conclusion, we revealed a novel mechanism that adiponectin increases hippocampal neurogenesis through activating Notch signaling. In addition, our work suggests that the Adiponectin-Notch pathway may be involved in chronic stress-induced hippocampal

| Chronic restraint stress and physical exercise
For the chronic restraint stress procedure, male C57BL/6J experi-

| Co-Immunoprecipitation (Co-IP)
Mice were decapitated rapidly, Hippocampus was collected and transferred to a 2 ml tube (Hippocampus of two mice/one tube).
Then added 1.2 ml ice-cold IP lysis buffer and homogenized, centrifuged at 10,000 g for 5 min. Transferring the supernatant to a

| Chromatin Immunoprecipitation (CHIP)
Mice were decapitated rapidly, Hippocampus was collected and transferred to a 2 ml tube (Hippocampus of two mice/one tube).
Then added 1.2 ml ice-cold PBS containing 1% Formaldehyde and added 75 µl Glycine solution (2 M) 15 min later. Centrifuged and got rid of the supernatant, and washed the precipitate with ice-cold PBS.
Then added 1.2 ml nuclear lysis buffer and homogenized, centrifuged at 10,000 g for 5 min. Transferring the supernatant to a 1.5 ml tube and getting DNA fragments using ultrasonication. Transferred 400 µl supernatant to a 5 ml tube, and added 4 ml CHIP dilution buffer. Then performed CHIP with PPARα and RXR antibody. qPCR was performed as described previously (Corbett et al., 2015).

| Sucrose preference test
This task is used to assess anhedonia in depression which is based on the animal's natural preference for sweets. Before beginning testing, Mice were habituated to the presence of two drinking bottles for one week. On an experimental day, water was deprived for three hours. After lights off during the dark cycle, mice have the free choice of either drinking 1% sucrose solution or water for 2 h.
Sucrose and water consumption were determined by measuring the weight changes. Sucrose preference was calculated as the ratio of the mass of sucrose consumed versus the total mass of sucrose and water consumed during the test.

| Forced swim test
This task is used for assessing the behavioral despair in depression by measuring the immobility time when mice were immersed in a plexiglas cylinder filled with water. On an experimental day, the plexiglas cylinder (25 cm height × 10 cm diameter) was filled with water at a 15 cm depth (24°C ± 1°C). Each mouse was tested for 6 min and video was recorded by a camera directly above. The latency to immobility at the first 2 min and the duration of immobility during the last 4 min were measured. Immobility was defined as no movements except those that maintain their head above water for respiration.

| Tail suspension test
This task is used for assessing the behavioral despair in depression by measuring the immobility time when mice were suspended by their tails. On the experimental day, each mouse was suspended within a three-walled compartment (50 height × 15 width × 15 cm depth) and video was recorded by a camera for 6 min. The degree of depression was assessed by calculating the duration of immobility during the 6 min.

| Light-Dark test
This test is based on the conflict between innate aversion of light and spontaneous exploratory behavior in the novel environment which could be used to evaluate the anxiogenic-like activity in mice (Bourin & Hascoët, 2003). The apparatus consisted of a polypropylene cage (45 × 27 × 30 cm) and was separated into two compartments, one third for the dark compartment and two thirds for the light compartment.
There was an opening between the two compartments (7 × 7 cm).
When conducted this test, each mouse was placed in the center of the dark compartment facing away from the opening and video was recorded by a camera for 5 min. The time spent in the light compartment and the number of entries into the light compartment were recorded.

| Elevated plus maze
This test was used to measure the anxiety-like behavior in mice

| Locomotor
This task is used to assess locomotor activity which was per- formed in SuperFlex open field cages (40 × 40 × 30 cm, Omnitech Electronics Inc.), and mice were allowed 30 min free exploration under illuminated conditions. The total distance traveled was quantified using Fusion version 6.5M software (Omnitech Electronics Inc.).

| Open field
This test was performed in an arena (60 × 60 × 40 cm) with even illumination. Mice were allowed free movement for 10 min that was recorded by a camera. The distance traveled in the central zone and the total distance traveled in the arena were analysed using Any-maze software (Stoelting). The arena was divided into nine squares (3 × 3 grid), and the central square was defined as the central zone.

| Novel object recognition (NOR)
Novel object recognition test was used to assess short-term spatial memory of mice and performed with a slightly modified protocol as described previously (Antunes & Biala, 2012;Liu et al., 2020). Mice received 2 days of habituation in a 45 × 45 cm square arena, and on the third day, they were allowed to explore two identical objects for 10 min (training trial). After 2 hr, one object was replaced by a novel one and the mice were allowed to explore for another 10 min (testing trial). The time spent on each object was then calculated as a percentage of total object exploration.

| Y maze
The Y-maze test was performed with a slightly modified protocol as previously described (Chiba et al., 2009). The apparatus for Y maze was a symmetrical Y Maze (3 arms, 40 × 9 cm with 12 cm-high walls).
The three arms were connected at an angle of 120°. Mice were individually placed at the end of an arm and allowed to explore the maze freely for 10 min. The total arm entries and spontaneous alternation percentage (SAP) were measured. Overlapping triplets of 3-arm visits were counted as one 'successful choice'. SAP was defined as a ratio of the number of 'successful choice' to the number of total choices (total entry minus two).

| Morris water maze (MWM)
The Morris water maze test was performed as previously described (Barnhart et al., 2015;Vorhees & Williams, 2006). The water maze of 150 cm in diameter and 50 cm in height was filled with water (25 ± 0.5°C) to maintained the water surface 1.00 cm higher than the platform (10 cm in diameter). Water was dyed white and the tank was divided into four quadrants and the platform was placed at the center of the designated quadrant. In the acquisition phase (4 trials/ day for 5 consecutive days), mice were put into the water from four points in random order every day until they found the platform and stayed for 10 s within 1 min. If the mice cannot find the platform within 1 min, they were guided to the platform. During the retention phase, the platform was removed from the pool, and the mice were placed in water from the opposite quadrant of the platform and tested for 1 min. Videos were recorded and analysed by Any-maze software (Stoelting).

| Statistical analyses
Statistical analysis was performed with graphpad prism software.
Results are presented as mean ± standard error of mean (SEM).
Shapiro-Wilk test and F test were used to test the normality and equal variance assumptions, respectively. For normally distributed data, two-tailed t-tests were used to assess differences between two experimental groups with equal variance. For a two-sample comparison of means with unequal variances, two-tailed t-tests with Welch's correction were used. One-way analyses of variance (ANOVAs) followed by Tukey's multiple comparisons test were used for analysis of three or more groups. For non-normally distributed data, Mann-Whitney U-tests were performed to compare two groups. For analysis of three or more groups with non-normally distribution, the Kruskal-Wallis test followed by Dunn's multiple comparisons test was used. For multiple groups, two way or two-way repeatedmeasures ANOVAs followed by Tukey's multiple comparisons test were used. p < 0.05 was considered statistically significant.