Inhibition of Lactate-GPR81-PI3K/Akt Pathway May Exacerbate Aβ Aggregation in 3-Month-Old APP/PS1 Mice

Lactate is not only an energy metabolite for neurons, but also serves as a molecule regulator affecting neuronal activity through its receptor, G protein-coupled receptor 81 (GPR81). This receptor can trigger cellular signaling pathways, such as phosphatidylinositol 3 kinase /protein kinase B (PI3K/Akt) pathway. Particularly, lactate decit and inhibition of PI3K/Akt pathway were observed to be related with early synaptic dysfunction in Alzheimer’s disease (AD). In addition, amyloid beta (Aβ) is toxic to neurons, while in vitro lactate administration of neurons can resist against this toxicity. Hence, this work focuses on the effect of lactate deciency on Aβ production, suggesting that lactate decrease can inhibit GPR81-PI3K/Akt pathway, and then reduce deoxyribonucleic acid methyltransferase 1 (DNMT1) expression, further resulting in increase of beta-site amyloid precursor protein cleaving enzyme 1 (BACE1) and Aβ production. Based on the analysis of results, this study primarily proves that lactate reduction suppresses its downstream GPR81-PI3K/Akt pathway, which decreases the expression of DNMT1 through regulating cyclic-adenosine-monophosphate response element-binding protein (CREB)/P300. Then, it is proved that DNMT1 reduction can lead to the increase of BACE1 and Aβ accumulation in AD. At last, in vitro experiment recognizes that lactate directly activates GPR81-PI3K/Akt pathway. Thus, this study provides a novel insight in Aβ production in relation with lactate decit at early stage of AD. Particularly, it is suggested that extra addition of lactate might be protective for neurons targeting Aβ clearance in early treatment of AD.


Introduction
Alzheimer disease (AD) is a neurodegenerative disorder characterized by progressive cognitive decline and dementia [1]. About 95% of AD is sporadic form, which is characterized by a late onset, and is the consequence of the failure to clear the amyloid beta (Aβ) peptide from the interstices of the brain [2].
Genetic risk factors for sporadic AD are inherited mutations in genes that affect processing of Aβ and develop the disease at a much younger age (mean age of about 45 years) [3]. Continuous production of Aβ leads to aggregation of Aβ-containing amyloid plaques and accelerates tau-derived neuro brillary tangles (NFTs), which ultimately leads to AD dementia. [4]. Hence, cerebral accumulation of Aβ peptide is not merely an important molecular hallmark of AD, but a promising therapeutic target [5].
Aβ clearance can decrease progressive axonal degeneration [6]. Unfortunately, two key questions remain unanswered: How to effectively lower Aβ production and promote Aβ clearance? Which stage of AD would be e cacy in an Aβ-directed therapeutic approach [7]? beta-site amyloid precursor protein cleaving enzyme 1 (BACE1) is an aspartic protease which functions in the rst step of the pathway leading to the production and deposition of Aβ [8]. Therefore, increase of BACE1 expression in AD at early stage can accelerate Aβ accumulation and quicken AD progression. BACE1 expression can be regulated by deoxyribonucleic acid methyltransferase 1 (DNMT1), DNMT1 reduction leads to hypomethylation of speci c loci within the BACE1 gene promoter, which further enhances BACE1 expression [9]. Hence, decrease of DNMT1 can upregulate BACE1 expression, which may further lead to Aβ production during Page 3/17 AD progression. Otherwise, the change of DNMT1 expression in relation with the levels of BACE1 and Aβ in AD brain at early stage hasn't been recognized.
The aggregation of Aβ plaques is associated with in ammation, oxidative stress, and energy de cit [10].
Especially, suppression of cerebral glucose utilization is identi ed in individuals of familial AD before the manifestation of Aβ plaques [11,12]. At this moment, glycolytic pathway and its product, lactate, become the main energetic resource of neurons, which can further defend against Aβ-mediated impairment of mitochondrial respiration [13]. In vitro experiment has proved that nerve cells favoring the glycolytic pathway resist against Aβ toxicity [14]. Even, lactate administration rescues the death of cultured neurons [15]. Further, it is discussed that lactate is necessary for long-term memory formation and improvement of cognitive function [16]. Hence, lactate may play a protective role in neuronal damage at early stage of AD. Otherwise, in our previous work, it is recognized that cerebral lactate content is decreased in 3-month-old double-transgenic amyloid precursor protein/presenilin 1 (APP/PS1) mice [17]. the observed lactate de ciency is proved to be associated with early neuronal damage in AD brain [17,18].
Here, this study aims to explore underlying mechanism of lactate de cit in relation with the production of neuronal Aβ in AD.
Lactate, an intermediate metabolite of glucose, transports from bloodstream to blood brain barrier via monocarboxylate transporters (MCTs) [19]. In the brain, lactate is temporarily stored in glial cells. During speci c periods, such as brain development and AD, lactate can be quickly transported from glias to neurons and metabolized to sustain neuronal activity [20,21]. Therefore, lactate is normally recognized as a quickly energetic substrate of neurons. Actually, lactate is not merely an energy resource, but also recognized as a signaling molecule [22]. Lactate can bind to its receptor, G protein-coupled receptor 81 (GPR81), which activates several downstreams, including phosphatidylinositol 3 kinase /protein kinase B (PI3K/Akt), extracellular regulated protein kinases (ERK1/2) pathway, Nod-like receptor family pyrin domain-containing 3/nuclear factor-κB (NLRP3/NF-κB) in ammatory signaling pathway and so on [23].
Here, this study hypothesizes that lacate reduction results in Aβ production, its underlying mechanism is that lactate de cit leads to the inhibition of GPR81-PI3K/Akt pathway, and then downregulating DNMT1 expression through CREB/P300. Moreover, DNMT1 decrease induces the hypomethylation of speci c loci within the BACE1 gene promoter, which upregulates BACE1 expression, further aggravating Aβ production in AD, especially at early stage. The results illustrate that lactate reduction inhibits GPR81-PI3K/Akt/CREB-DNMT1-BACE1 signaling pathway, which exacerbates Aβ aggregation in AD. Hence, this study provides a novel insight into the mechanism of Aβ accumulation and provides a novel target in anti-Aβ therapy for early treatment of AD.

Materials And Methods
Animals 3-month-old heterozygous APP/PS1 mice (n = 10) and their nontransgenic littermates (wild type, n = 10) were used in this study. Animals were housed in individual cages in a controlled environment (temperature, 22 ± 1℃; humidity, 50% ± 10%; 12-hour light/12-hour dark cycle). Food and water were available ad libitum. Animals were grouped and named as Wild Type and APP/PS1.

Primary Neuron Culture and Grouping
Adult C57BL/6 mice were obtained from the Experimental Animal Center of Army Medical University (Chongqing, China). The primary culture of cerebral cortical neurons from embryos of C57BL/6 mice was performed as previously described [27]. In brief, cerebral cortex of 18-day-old embryos was dissected from brain and then cut into slices. The slices were mechanically dissociated by trituration. The dissociated cells were suspended in Eagle's minimal essential medium supplemented with 3% B27 Minus AO, 10 μg/ml insulin, 0.25 μM glutamine, 1 mM β-hydroxybutyrate, 1 mM fumarate, and 50 ng/ml sodium selenite. The cell suspension was plated onto 96-well culture plate previously coated with poly-D-lysine and bronectin. Cells were incubated at 37 °C in a humidi ed atmosphere of 95 % O 2 /5 % CO 2 for 14 days. To detect the impact of lactate on neurons, the obtained neurons were given lactate for 24 hours in a dose-dependence (5mM, 10mM, 15mM), and lactate inhibitor (sodium oxamate) administration.
Further, these cells were grouped and named as Control, 5mM LAC, 10mM LAC, 15mM LAC, and Inhibitor.

Tissue Processing
Following perfusion with phosphate buffer solution (PBS), pH 7.4, the left hemispheres of brains from APP/PS1 mice and wild type mice (n = 5, respectively) were collected and stored in -20°C for lactate measurement. Meanwhile, the right hemispheres of brains were homogenized by RIPA lysis buffer containing 1mM PMSF. Homogenates were then centrifuged for 20minutes at 14000rpm and concentrations of the supernatants were assayed using the BCA protein assay kit and adjusted to 1.5mg/mL. Finally, the samples were stored at -20°C for WB analysis. For immunohistochemical assays, APP/PS1 mice and wild type mice (n = 5, respectively) under anesthesia were perfused with saline followed by 4% paraformaldehyde in PBS and the brains were extracted and post xed with fresh 4% paraformaldehyde at 4°C. Tissues were transferred to 30% sucrose solution for 2-3 days and subsequently cut into 20 μm slices.

Measurement of Lactate Concentration
Determinations of lactate concentration were performed with a lactate assay kit. The cortex and hippocampus were homogenized in saline at 4°C for 10 minutes, then the homogenates were centrifuged for 15 minutes at a speed of 2500 rpm. Assay buffers were added to the supernatants and incubated for 10 minutes at 37°C . Finally, optical density (OD) values were recorded at 530 nm after the reaction of lactate and assay buffers. Lactate content was calculated using the following formula: Lactate content (mmol/g) = (OD measured value-OD blank value)/(OD standard value -OD blank value)×standard substance content (3 mmol/L)/protein content (g/L)

WB Analysis
Protein samples were subjected to SDS-PAGE and transferred to PVDF lter membranes. The membranes were blocked with 5% nonfat milk for 1hour at 37°C and incubated with primary antibodies for 12 hours at 4°C, including mouse anti-Aβ, mouse anti-p-PI3K, mouse anti-p85/p55, mouse anti-PI3Kp85, mouse anti-pAkt, mouse anti-Akt, mouse anti-pCREB, mouse anti-CREB. After washing with TBST, the membranes were incubated for secondary antibodies for 1 hour at 37°C and detected using ECL kit. Finally, the blots were quali ed by Image J software (NIH, Bethesda, MD).

Immuno uorescence Staining
Brain sections were probed with mouse anti-Aβ, rabbit anti-GPR81, rabbit anti-DNMT1, rabbit anti-BACE1 primary antibodies, respectively. Brain sections were incubated overnight with primary antibodies in a humidi ed chamber at 4°C. Sections were then washed 3 times with PBS for 5minutes each, followed by incubation with anti-mouse FITC secondary antibody. For double-labeling immuno uorescence, sections were incubated with the mixture of 2 primary antibodies overnight at 4°C as follows: Mouse anti-neuronal nuclei (Neu N) and rabbit anti-GPR81. Fluorescent secondary antibodies, raised in different species (FITC with green signal against mouse and TRITC with red signal against rabbit) were used to locate complexes of antigen/primary antibody. Nuclei were counterstained with DAPI for 5minutes. Images were obtained using a uorescence microscope (Olympus, Tokyo, Japan) at 400× magni cations. Positive expressions of staining pictures were analyzed by OD. OD values were calculated using Image-Pro Plus 6.0 (IPP6.0) software according to manufacturer's instructions.

Statistical Analysis
All data are expressed as mean ± standard deviation (SD). All statistical analyses were performed by Statistical Product and Service Solution (SPSS, IBM version 21) and Prism 6 (GraphPad Prism Software Inc., La Jolla, CA). Results of WB and immuno uorescence were analyzed using one-way analysis of variance (ANOVA) and, correlation analysis, and the least signi cant difference (LSD) test. Differences with *P < 0.05 and **P < 0.01 were considered statistically signi cant.

Results
Reduction of lactate and GPR81 expression in the cortex and hippocampus of 3-month-old APP/PS1 mice Lactate content and GPR81 expressions in cortex and hippocampus of 3-month-old APP/PS1 mice and wild type mice were assessed by immuno uorescence staining and WB. Lactate levels in the cortex and hippocampus of APP/PS1 mice are 5.1 ± 1.28 mmol/g and 5.29 ± 1.76 mmol/g, which are lower than lactate contents in cortex (14.8 ± 1.13 mmol/g) and hippocampus (15.91 ± 1.03 mmol/g) of wild type mice (n = 5, *p < 0.05; Fig. 1a). As arrowheads pointed, Positive staining of GPR81 in cortex and hippocampus of APP/PS1 mice is lower than that of wild type mice (Fig. 1b). In statistic, OD values of GPR81 in cortex and hippocampus of APP/PS1 are 21196 ± 2187 and 18503 ± 1895, which are decreased in comparison with wild type mice (32741 ± 3012 in cortex and 27332.3 ± 2198 in hippocampus) (n = 5, *p < 0.05; Fig. 1c).
DNMT1 expression is reduced in the cortex and hippocampus of 3-month-old APP/PS1 mice DNMT1 expressions in cortex and hippocampus were evaluated by WB and immunostaining. Positive protein bands of DNMT1 are identi ed in cortex and hippocampus of APP/PS1 mice and wild type mice.
In comparison with wild type mice, relative expressions of DNMT1 in cortex and hippocampus of APP/PS1 mice are reduced . Speci cally, relative levels of DNMT1 in cortex and hippocampus of APP/PS1 mice are 0.14 ± 0.02 and 0.08 ± 0.01. In wild type, relative expressions of DNMT1 are 0.53 ± 0.04 in cortex and 0.38 ± 0.02 in hippocampus (n = 5, **p < 0.01; Fig. 3a). In cortex and hippocampus of APP/PS1 mice and wild type mice, arrowheads points out postive expressions of DNMT1 which mainly localizes in cellular nuclei (Fig. 3b). Statistically, OD values of DNMT1 in cortex and hippocampus of APP/PS1 mice are 10057 ± 428 and 9900 ± 568. In wild type, OD values of DNMT1 are 16863 ± 1486 in cortex and 14157 ± 1521 in hippocampus. Postive expressions of DNMT1 in cortex and hippocampus of APP/PS1 mice are lower than than thoes of wild type mice (n = 5, *p < 0.05; Fig. 3c).

BACE1 and Aβ expression is increased in the cortex and hippocampus of 3-month-old APP/PS1 mice
Expressions of BACE1 and Aβ were checked by immunostaining and WB. BACE1 is postively labelled and spreads in cortex and hippocampus of APP/PS1 mice and wild type mie, pointed by arrowheads (Fig. 4a). Statistically, positive expressions of BACE1 in cortex and hippocampus of APP/PS1 mice are increased in comparison with wild type mice . In speci c, OD values of BACE1 in cortex and hippocampus of APP/PS1 mice are 38036 ± 1299 and 21917 ± 2229. In wild type mice, OD values of BACE1 are 21896 ± 1595 in cortex and 15333 ± 1052 in hippocampus (n = 5, *p < 0.05; Fig. 4b). BACE1 expressions were futher assessed by WB. There are positive protein bands of BACE1 in cortex and hippocampus of APP/PS1 mice and wild type mice (Fig. 4c). Compared with wild type mice, relative levels of BACE1 in cortex and hippocampus of APP/PS1 mice are increased. Relative expressions of BACE1 in cortex and hippocampus of APP/PS1 mice are 1.2 ± 0.05 and 0.59 ± 0.02. In cortex and hippocampus of wild type mice, relative quantities of BACE1 are 0.67 ± 0.02 and 0.19 ± 0.04 respectively (n = 5, *p < 0.05; Fig. 4c). Moreover, Aβ contents were evaluated by WB, positive bands of Aβ are mainly found in cortex and hippocampus of APP/PS1 mice (Fig. 4d). In comparison with wild type mice, Aβ levels in cortex and hippocampus of APP/PS1 mice are increased. Relative quantities of Aβ in cortex and hippocampus of APP/PS1 mice are 0.93 ± 0.05 and 0.19 ± 0.02. In wild type mice, relative levels of Aβ are 0.3 ± 0.02 in cortex and 0.02 ± 0.01 in hippocampus respectively (n = 5, *p < 0.05; Fig. 4d).
Lactate activates GPR81-PI3K/Akt signaling pathway in neurons Cultured neurons were divided into 5 groups and given different administrations. Then, expressions of GPR81, p-PI3K, p-Akt, p-CREB were assessed to identify whether lactate can directly activate GPR81-PI3K/Akt pathway. As results showed, in comparison with Control group, 10 mmol/L lactate effectively upregulates relative levels of GPR81, p-PI3K, p-Akt and p-CREB expressions. Otherwise, expressions of GPR81, p-PI3K, p-Akt and p-CREB are suppressed by 15 mmol/L lactate or sodium oxamate, an inhibitor of lactate ( Fig. 5 and Fig. 6).

Discussion
In 1930s, lactate is considered as a way to clear unwanted or toxic metabolite rather than as a process through which lactate could produce energy [28]. Until 1950s, lactate is considered as an alternative energy substrate to glucose in the case of glucose deprivation [29]. Even, in the presence of adequate glucose, lactate is a preferred substrate to sustain neuronal activity [30]. Recently, studies signify that lactate is not merely an energy resource of neurons, but plays as a molecule to regulate activities of neural cells [31]. As a signaling regulator, lactate can bind to its receptor, GPR81, which can activate PI3K/Akt, ERK1/2 and NLRP3/NF-κB pathways [23]. Also, lactate can signal by being transported into neurons through MCTs, which further modulates redox-dependent and energy-dependent mechanism[32].
Here, this study discusses lactate, as a signaling molecule, effects on the regulation of Aβ production through GPR81.
During AD progression, Aβ deposition starts with the production of insoluble Aβ brils. Hence, anti-Aβ production at early stage of AD is though to be a promising therapeutic strategy [34]. BACE1 is the ratelimiting enzyme of insoluble Aβ cleavage, which cleaves amyloid precursor protein to produce insoluble Aβ [34]. Meanwhile, promoter of BACE1 gene in AD subjects is hypomethylation, which is related with the reduction of DNMT1 [9,35]. In addition, inhibition of PI3K/Akt pathway was proved to aggravate Aβ accumulation [36]. Therefore, this work proposes that at early stage of AD, lactate de cit leads to the suppression of GPR81-PI3K/Akt pathway, which downregulates DNMT1 level, further promoting BACE1 expression and Aβ production.
Based on the assessment of 3-month-old APP/PS1 mice and wild type mice, this study primarily identi es that lactate content and GPR81 expression are reduced in cortex and hippocampus of APP/PS1 mice. It is recognized that the activity of GPR81 downstream PI3K/Akt pathway is inhibited in APP/PS1 mice. Hence, it is proved that lactate and its downstream GPR81-PI3K/Akt are suppressed at early period of AD. Otherwise, what's the direct effector of this pathway? It is showed that the inhibition of GPR81-PI3K/Akt downregulates DNMT1 expressions through CREB/P300. Moreover, DNMT1 reduction leads to hypomethylation of BACE1 promoter [9]. this is in accordance with the observed increase of BACE1 and Aβ levels in 3-month-old APP/PS1 mice. Thus, it is proves that at early stage of AD, lactate is an important molecule regulator, its reduction leads to the suppression of GPR81-PI3K/Akt pathway, which aggravates Aβ production through decreasing DNMT1 and increasing BACE1 levels.
These evidences illustrate that lactate decrease promotes Aβ production through inhibiting PI3K/Akt pathway in AD at early stage. Otherwise, the direct regulation of lactate in GPR81-PI3K/Akt pathway hasn't been recognized. Therefore, we further checked the effect of lactate and lactate inhibitor on the activation of GPR81-PI3K/Akt pathway. As results showed, lactate directly triggers GPR81 and activate PI3K/Akt pathway in the cultured neurons. Hence, this work discovers a novel mechanism of early Aβ accumulation in AD that is dominated by lactate-GPR81-PI3K/Akt signaling pathway. Meaningfully, it is suggested that lactate supplement might be a novel therapeutic strategy targeting early Aβ production of AD.