Berberine Promotes TREM2-Dependent Phagocytosis of Amyloid-β By Microglia

Background: The production and accumulation of amyloid-β (Aβ) is the most important pathological feature of Alzheimer’s Disease (AD), and the deciency of Aβ clearance contributes to the progression of AD. TREM2-dependent microglial activation may be the key to Aβ clearance. BBR plays the neuroprotective role in the progression of AD by inhibiting Aβ production and promoting Aβ degradation. However, the specic relationship between BBR and microglial activation remains unclear. Thus, we aimed to investigate whether BBR can inhibit the pathological progression of Aβ in AD by changing the phenotype of microglia. Methods: Western blot and Immunouorescence staining were applied to detect the effects of BBR on the transformation of resting microglia to different phenotypes. ELISA, Immunohistochemistry and Immunouorescence were used to detect the effect of BBR on microglial phagocytosis of Aβ. Morris water maze (MWM) test was applied to test the effect of BBR on the spatial learning and memory of experimental animals. Results: Firstly, BBR promoted the phagocytosis of Aβ 1-42 by BV2 cells. Secondly, BBR promoted the changes of microglia to phenotypes M2 and DAM in vivo and in vitro, which were in close proximity to Aβ and reduced Aβ aggregation. Finally, BBR ameliorated spatial learning and memory impairment in APP/PS1 mice. Conclusion: BBR could enhance the phagocytosis of microglia, which decreased Aβ level and improved the spatial learning and memory of APP/PS1 mice.

biomarkers of classical M1 include iNOS and CD32, which can release a large number of proin ammatory cytokines, such as IL-1β, IL-6, and TNF-α, etc., and damage CNS. The corresponding biomarkers of M2 include Arg1 and CD206, which are capable of anti-in ammatory cytokines such as IL-10 and TGF-β, promoting nerve tissue repair and nerve regeneration [7]. Moreover, a recent single-cell RNA-seq study has revealed a unique microglia type in a mouse model of AD, named "disease-associated microglia" (DAM, marker CD11c), that is actively involved in the breakdown and digestion of amyloid plaques [8].
Triggering receptor expressed on myeloid cells 2 (TREM2) is a member of the TREM family of innate immune receptors that are expressed in microglia, dendritic cells, macrophages, and osteoclasts. TREM2 could bind phospholipids and other polyanionic ligands and transmit intracellular signals through the related adaptor DNAX-activation protein 12 (DAP12), promoting the survival, proliferation, phagocytosis, and secretion of cytokines and chemokines [9,10]. Previous convincing studies have identi ed the rare mutation (R47H) in TREM2 as a genetic risk factor for AD [11,12]. Studies in AD mice (APP/PS1) have found that the TREM2 R47H variant increases the risk of AD by impairing the function of TREM2 and enhancing neurodystrophy around plaques [13]. Another study in 5×FAD mice and humans with AD found that TREM2 haplode ciency may disrupt the formation of neuroprotective microglial barriers that regulate amyloid compaction and insulation [14]. Similarly, in PS2APP AD mouse models, TREM2 de ciency reduced the accumulation of late-stage amyloid plaques, increased the Aβ42 to Aβ40 ratio, and exacerbated axonal dystrophy and dendritic spine loss [15].
Berberine (BBR) is a natural chemical found in a variety of plants, including Coptis chinensis, European barberry, Tree turmeric, Goldenseal, Phellodendron and so on [16]. These plants have historically been used to treat a wide range of ailments, including intestinal infections, diarrhea and ulcers. Current research has found that BBR could be used to help treat diabetes [17], high cholesterol [18], obesity [19], cardiovascular disease [20], musculoskeletal disorders [21] and even gynecological cancers [22]. In addition, BBR shows great therapeutic potential in neurodegenerative diseases such as Alzheimer's, Parkinson's and Huntington's diseases [23].
In AD pathologic progression, BBR not only alleviates Aβ pathology by inhibiting β/γ secretases, but also weakens neuronal damage caused by Aβ, and improves glial hyperplasia and cognitive impairment [24][25][26]. Studies have shown that autophagy is defective in AD and cannot remove cellular garbage normally [27]. However, BBR intervention can alleviate cognitive decline in AD mice by inhibiting Aβ production, Tau hyperphosphorylation and autophagy clearance [28,29]. Our previous study found that BBR ameliorates ribosylation induced Aβ pathology by inhibiting mTOR/p70S6K signaling and promoting autophagy, thereby improving spatial learning and memory in APP/PS1 mice [30].
In this study, we hypothesized that BBR promotes microglial phenotypic changes, thereby regulating TREM2-dependent phagocytosis of Aβ by microglia. To test this hypothesis, we performed multiple experiments in BV2 cells and transgenic APP/PS1 transgenic AD model mice.

Morris water maze test
MWM is carried out in a circular tank with a diameter of 1.2 m and a depth of 0.4 m, with water temperature maintained at 25 ± 1°C. A circular escape platform (8 cm diameter) was xed and hidden 1 cm in the middle of the northeast quadrant below the water surface. The adaptive mice were trained 4 times per day for 6 consecutive days as acquisition trials. During the trail, mice were placed into the different quadrants and allowed to swim freely for 60 s to climb up and stay on the platform for 5 s. If a mouse failed to reach the platform within 60 s, it was guided to the platform for an extra 30 s. The time the mice took to reach the platform was considered its escape latency and was recorded along with the swim trajectory by ANY-maze video tracking software (Stoelting Co., USA). For the probe test, mice were placed in the opposite quadrant from the original platform quadrant and tested for 60 s without a platform. The time spent in the four quadrants, the distance to the platform, the distance in the original platform quadrant and the number of platform crossings were observed and recorded.
Cell proliferation and apoptosis assay Cell Counting Kit 8 (CCK-8, abs50003, Absin, Shanghai, China) was used to assess the proliferation of BV2 cells according to the manufacturer's protocol. In brief, BV2 cells were cultured in 96-well plates with 5.0×10 3 cells/well. After intervention with different reagents, BV2 cells were cultured for 0, 12, 24 or 48 h, followed by adding CCK-8 solution to each well. After further incubation for 4 h, the absorbance of each well was detected at 450 nm using Multiskan FC (Thermo Scienti c). Cell apoptosis of BV2 cells was assayed by Hoechst Staining Kit (C0003, Beyotime, Shanghai, China). BV2 cells were seeded on the coverslips in 24-well plates with 1. Western blot analysis BV2 cells or brain tissue of APP/PS1 mice were lysed with RIPA Lysis Buffer containing protease inhibitor PMSF (Beyotime), and the total protein extracted was detected by BCA Protein Assay Kit (Beyotime). Equivalent amount of protein were separated on the polypropylene gels with corresponding concentration and transferred to PVDF membranes (Bio-Rad, Hercules, CA, USA). Blocked membranes were incubated with indicated primary antibodies overnight at 4°C. Membrances were then incubated with corresponding secondary antibodies for 1 h at room temperature. Protein bands were visualized using Immobilon Western Chemiluminescent HRP Substrate (Millipore Corporation, MA, USA) and Tanon-5200 multi chemiluminescence analysis system (Tanon, Shanghai, China).

Histology and Immunohistochemical staining
After the behavioral test, the brains of APP/PS1 mice were removed and xed with 4% paraformaldehyde for 48 h. Para n sections with a thickness of 4 µm were cut along the long axis of the sagittal plane by a Rotary Microtome (HM 340E, Thermo Scienti c). After depara nization and rehydration, antigen-retrieval was performed by microwave oven with citric acid buffer. Then para n sections were blocked with 10% goat serum and incubated with indicated primary antibodies overnight at 4°C. After incubation with the secondary antibodies, sections were detected by DAB staining kit, counterstained with hematoxylin, and the target images were observed and captured with a bright eld NEXCOPE microscope (NE900, USA).
Immuno uorescence staining BV2 cells cultured on glass coverslips were washed with PBS and xed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100. 4 µm para n sections or BV2 cells were blocked with 5% goat serum albumin, and incubated with indicated primary antibodies overnight at 4°C. After incubation with secondary uorescent antibodies Alexa Fluor® 488 or Alexa Fluor® 594 antibodies (ZSGB-BIO, Beijing, China) for 1 h in the dark, the nucleuses were stained with DAPI (Beyotime). The coverslips were mounted onto glass slides using Antifade Mounting Medium (Beyotime), and the imaging was performed using NEXCOPE microscope (NE900, USA).

Statistical analyses
All the experimental data was analysed using GraphPad Software. ANOVA and Independent-sample t tests were applied to compare the differences of measurement data between each group. The data was represented as means ± standard error of mean (SEM) and p < 0.05 was considered statistically signi cant.

BBR promoted phagocytosis of Aβ 1−42 by BV-2 cells
Western blot analysis revealed that BBR treatment could signi cantly increase the protein expression level of TREM2 compared with the control group in BV2 cells (Fig. 1a). To investigate the pathophysiological function of TREM2, siRNA interference was used to knock down TREM2. Western blot results showed that the expression of TREM2 in three different siRNA-targeted interferences was signi cantly lower than that in the blank control group (Fig. 1b). Based on this result, siTREM2-1 was selected for following knockdown experiments. Then, BV2 cells were stained with Hoechst 33258, a special uorescent dye that can distinguish apoptotic cells from normal cells, under a uorescence microscope, and the cell morphology was observed. As shown in Fig. 1c, after BBR treatment, the nuclear morphology did not change, showing diffuse homogeneous blue uorescence as in normal nuclei. When treated with si-TREM2 or Aβ 1−42 , the cells show apoptotic cell-like nucleus fragmentation, chromosome condensation and strong blue uorescence. However, the addition of BBR in combination with si-TREM2 or Aβ 1−42 treatment can restore the nuclear state of apoptosis. The effect of TREM2 on BV2 cell viability was assessed using the CCK8 assay. As shown in Fig. 1d, the cell viability of BBR treatment group decreased slightly by 11.8%, whereas si-TREM2 and Aβ 1−42 treatment signi cantly decreased by 53.8% and 62.4%. Compared with si-TREM2 group and Aβ 1−42 group, BBR supplementation reversed cell viability by 58.1% and 94.3%. These results indicated that BBR has no signi cant effect on cell viability, and can rescue the decreased cell viability caused by si-TREM2 and Aβ 1−42 . ELISA assay was used to detect the effect of BBR on the Aβ phagocytosis of BV2 cells. In BV2 cell contents, Aβ 1−42 treatment signi cantly increased Aβ 1−42 content compared with the control group, while the addition of BBR signi cantly reversed this effect (Fig. 1e Left). Similarly, in the supernatant of BV2 cells, Aβ 1−42 treatment signi cantly increased Aβ 1−42 content, while BBR signi cantly reversed this result (Fig. 1e Right). These data suggest that BBR promotes Aβ phagocytosis in BV2 cells.
BBR promotes the phenotype changes of microglia in vitro.
Microglia cells have different phenotypes. Activated microglia cells include neuro-damaging M1, neuroprotective M2, and a new type of DAM that can signi cantly limit the progression of AD. Therefore, the role of microglia cells in AD is a double-edged sword. We should take advantage of its favorable side and inhibit its harmful side to achieve the effect of treating AD. In this study, we found that BBR inhibits the transition of resting BV2 cells to M1 and promotes the transition of resting BV2 cells to M2 by regulating TREM2 (Fig. 2a). As shown in Fig. 2b-c, compared with the control group, BBR signi cantly decreased the expression of M1 markers CD32 and IL-1β, and signi cantly increased the expression of M2 markers CD206 and IL-10, while si-TREM2 signi cantly promoted the expression of CD32 and IL-1β, and signi cantly inhibited the expression of CD206 and IL-10. However, compared with si-TREM2 group, BBR supplementation signi cantly decreased the expression of CD32 and IL-1β, and signi cantly increased CD206 and IL-10 expression. In addition, we evaluated the expression of CD32 (Fig. 2d) and CD206 (Fig. 2e) in microglia (IBA1) using immuno uorescence staining, which was consistent with Western blot results. In short, our results prove that BBR could inhibit the transformation of resting microglia to M1 and promote the transformation of resting microglia to M2. Then, immuno uorescence was used to detect BBR's effect on microglial transformation to DAM cells. As shown in Fig. 3a-b, BBR signi cantly promoted DAM marker CD11c expression and si-TREM2 signi cantly promoted CD11c expression compared with the control group in microglia (IBA1). However, BBR combined with si-TREM2 treatment did not restore CD11c expression compared with si-TREM2 group, suggesting that BBR may promote microglial conversion to DAM through TREM2. In conclusion, these data demonstrate that BBR promotes BV2 cells transformation to neuroprotective M2 and DAM and inhibit BV2 cells transformation to neurotoxic M1.

BBR promotes the phenotype changes of microglia in vivo.
To further con rm the effect of BBR on microglia phenotypic changes, wild-type (WT) mice, AD control mice (Saline treatment) and AD experimental mice (BBR treatment) were used as animal models. In vivo experiments indicated that BBR inhibits the transition of resting microglia to M1 and promotes the transition of resting microglia to M2 and DAM, which was in accordance with in vitro results (Fig. 4a). Western blot assay showed that CD32 and IL-1β expression in the Saline group was signi cantly higher than that in the WT group. Compared with the Saline group, the expression of CD32 and IL-1β in AD mice after BBR treatment was signi cantly decreased, while the expression of CD206 and IL-10 was signi cantly increased (Fig. 4b-c). Immuno uorescence results showed that the CD11c uorescence intensity in microglia (IBA1) of AD mice treated with BBR was signi cantly increased compared with that of Saline group (Fig. 4d). These results suggest that BBR can not only inhibit the transformation of microglia into neurotoxic M1, but also promote the transformation of microglia into neuroprotective M2 and DAM in mice.
BBR promotes phenotypic altered microglia to surround Aβ in vivo.
The results of both in vivo and in vitro experiments showed that BBR could promote the transformation of microglia to M2 and DAM phenotypes. Next, we studied the effects of M2 and DAM phenotypic microglia on Aβ. Here we report that BBR promotes microglial transformation of M2 and DAM to encircle Aβ and inhibit Aβ expression through TREM2 (Fig. 5a). Immuno uorescence results showed that BBR promoted the increase of M2 (CD206) in microglia (IBA1), and the increased M2 tended to surround Aβ (Fig. 5b).
Similarly, BBR promoted the signi cantly increased DAM (CD11c) in microglia (IBA1) to be in proximity to Aβ (Fig. 5c), which may have the phagocytic effect on Aβ. Both M2 with anti-in ammatory effect and DAM with phagocytic effect were close to Aβ, which may have a positive promoting effect on Aβ clearance. Immunohistochemical results showed that BBR signi cantly increased the expression of TREM2 and signi cantly reduced Aβ plaques compared with AD control group (Fig. 5d), suggesting that BBR may promote the transition of resting microglia to M2 and DAM by promoting the expression of TREM2, thereby leading Aβ clearance.
BBR regulated the spatial learning and memory of APP/PS1 mice.
The MWM test was performed to evaluate spatial learning and memory of APP/PS1 mice followed the mouse experiment process diagram (Fig. 6a). The escape latency of WT, APP/PS1 control (Saline), and APP/PS1 + BBR groups decreased in a time-dependent way during 5 consecutive days of pre-training. On day 5, APP/PS1 control group had signi cantly higher escape latency than WT group, while APP/PS1 + BBR had signi cantly lower escape latency than APP/PS1 control group (Fig. 6b). After pre-training, the platform was removed to measure the platform location time, the across platform times and the retention time in the target quadrants to test the long-term spatial memory ability of experimental mice. As shown in Fig. 6c-e, APP/PS1 control group showed worse spatial memory ability than WT group, while the platform location time, the across platform times and the retention time of APP/PS1 + BBR group were signi cantly improved compared with that of APP/PS1 control group. Typical trajectories of the mice clearly showed differences in the spatial memory abilities of the three groups (Fig. 6f). These results indicated that BBR could signi cantly improve spatial learning and memory of APP/PS1 mice.

Discussion
AD is one of the most common diseases leading to cognitive dysfunction in middle-aged and elderly people. The pathogenesis and treatment of AD have always been a problem in the world. Although an endless stream of drugs and treatments have failed to achieve ideal results, the therapies and drugs that target Aβ have always been the key point in the treatment of AD. Microglia, a type of glial cells, is the rst and most important line of immune defense in CNS and plays an important role in the damage and repair of CNS. Activated microglia, including pro-in ammatory M1, anti-in ammatory M2 and phagocytes associated with AD (DAM) could be directly or indirectly involved in the occurrence and development of AD [31,32]. Microglia is a double-edged sword in AD. On the one hand, Aβ stimulates and activates microglia to produce in ammatory factors and neurotoxins, leading to neuronal damage and even death and triggering AD. On the other hand, microglia can protect CNS by phagocytosis of Aβ [33]. In current study, we found that BBR inhibits the transition of resting microglia to M1 and promotes the transition of resting microglia to M2 and DAM, which may play a positive role in inhibiting the pathological progression of Aβ.
In recent years, researchers have turned their attention to the central role of myeloid cells in a variety of pathologies, and TREM2 has been identi ed as the primary pathology-induced immune signaling center. They observed changes in the levels of TREM2 in various contexts of neurodegenerative change, further underlining the importance of TREM2 in neurological disorders and the exciting possibilities for treatment targeting TREM2, particularly in AD [34]. Colonna et al had previously shown that microglia surround Aβ plaques, preventing the damaged area from growing [10], while microglia in mice lacking TREM2 allow these plaques to spread and damage neurons more widely [35]. Reaserch found that TREM2 de ciency impairs Aβ degradation in vitro and in vivo, which demonstrates TREM2 as a microglial Aβ receptor transducing physiological and AD-related pathological effects associated with Aβ [36]. In addition, TREM2 is a receptor required to activate DAM and is critical for DAM to phagocytosis of Aβ and improves AD pathology, which may have important signi cance for the future treatment of neurodegenerative diseases such as AD [8]. In this study, we report that BBR promotes the transition of resting microglia to M2 and DAM in a TREM2-dependent way.
BBR is a natural drug extracted from plants such as Coptis chinensis and has a broad spectrum of antiin ammatory and anti-tumor activities. The pharmacological role of BBR in neurodegenerative diseases is gradually being reported. For example, BBR could improve brain dopa/dopamine levels to ameliorate Parkinson's Disease by regulating gut microbiota [37]. In Huntington's Disease transgenic mice, BBR could enhance the autophagy function to promote the degradation of mutant Huntington protein, and effectively alleviate the motor dysfunction of model mice, prolong their survival time [38]. A systematic review of pre-clinical studies [39] shows the neuroprotective effects of BBR in AD animal models: BBR showed signi cant memory-improving activity through a variety of mechanisms including antiin ammatory, anti-oxidative stress, cholinesterase inhibition and anti-amyloid. In this study we showed that BBR promotes the enveloping and phagocytosis of Aβ by regulating microglia phenotypic changes, and ultimately reduces Aβ content and promotes the improvement of cognitive impairment in AD mice.
However, the main limitation of this study is that only Aβ pathology was studied. AD is a complex disease driven by multiple factors, and many therapeutic strategies based on reducing Aβ have failed in clinical trials. This suggests that the treatment of AD should not be based on a single cause but on a number of different pathways. Given the complexity of AD pathology and the multiple e cacy of BBR, we should further investigate the value of BBR in the treatment of AD from different approaches.

Conclusion
In conclusion, we demonstrated that BBR promotes microglial phagocytosis of Aβ in BV2 cells and APP/PS1 mice (Fig. 7). First, BBR promotes the phagocytosis of Aβ in BV2 cells by increasing the expression of TREM2. Second, BBR inhibits the transition of resting microglia to M1 and promotes the transition of resting microglia to M2 and DAM in vivo and in vitro. Thirdly, BBR promotes the increase of M2 and DAM towards Aβ and decreases the content of Aβ by increasing the expression of TREM2. Moreover, BBR improves the spatial learning and memory of APP/PS1 mice. Taken together, these ndings reveal novel molecular pathways through which BBR improves Aβ pathology and promotes Aβ reduction by altering microglial phenotypes. It is suggested that BBR may be developed as a novel agent for the treatment of AD by altering microglia phenotype to promote Aβ clearance, thus contributing to the alleviation of Aβ pathology.

Availability of data and materials
The data during the current study are available from the corresponding author on reasonable request.

Figure 5
BBR promotes phenotypic altered microglia to surround Aβ. (a) Schematic diagram depicting the mode of BBR regulating the phenotype of microglia cells and Aβ. (b) Immuno uorescence was used to detect the uorescence intensity of CD206 (red) in the cortex of WT and AD mice. DAPI (blue), Aβ (green). (c) Immuno uorescence was used to detect the uorescence intensity of CD11c (green) in the cortex of WT and AD mice. DAPI (blue), Aβ (red). Scale bar = 50 μm. (d) TREM2 and Aβ detection was performed by immunohistochemistry in the cortex of WT and AD mice. Scale bar = 50 μm. Figure 6 BBR regulated the spatial learning and memory of APP/PS1 mice. (a) Experiment and timeline procedure.
16 weeks mice were divided into 3 groups, WT group, APP/PS1+Saline group and APP/PS1+BBR group. (b) Average escape latency of the 5 days space navigation training. The time in the platform location (c), the times of across platform (d), the retention time in the target quadrants (e), and the representative mouse trajectories (f) were observed during the space exploration task after removing the platform. (**p < 0.01 and ***p < 0.001 versus WT group, ##p < 0.01 and ###p < 0.001 versus Saline group).

Figure 7
Schematic diagram of possible mechanisms of BBR regulating microglia. BBR could inhibit the transformation of resting microglia to pro-in ammatory M1 and promote the transformation of resting microglia to anti-in ammatory M2. In addition, BBR may also promote the transformation of M1 to M2.
At the same time, BBR could promote the transformation of microglia to DAM by promoting the expression of TREM2, and DAM can surround Aβ and then decrease Aβ to protect neurons.