Diammonium Glycyrrhizinate Upregulates PGC-1α and Protects against Aβ1–42-Induced Neurotoxicity

Mitochondrial dysfunction is a hallmark of beta-amyloid (Aβ)-induced neurotoxicity in Alzheimer's disease (AD), and is considered an early event in AD pathology. Diammonium glycyrrhizinate (DG), the salt form of Glycyrrhizin, is known for its anti-inflammatory effects, resistance to biologic oxidation and membranous protection. In the present study, the neuroprotective effects of DG on Aβ1–42-induced toxicity and its potential mechanisms in primary cortical neurons were investigated. Exposure of neurons to 2 µM Aβ1–42 resulted in significant viability loss and cell apoptosis. Accumulation of reactive oxygen species (ROS), decreased mitochondrial membrane potential, and activation of caspase-9 and caspase-3 were also observed after Aβ1–42 exposure. All these effects induced by Aβ1–42 were markedly reversed by DG treatment. In addition, DG could alleviate lipid peroxidation and partially restore the mitochondrial function in Aβ1–42-induced AD mice. DG also significantly increased the PGC-1α expression in vivo and in vitro, while knocking down PGC-1α partially blocked the protective effects, which indicated that PGC-1α contributed to the neuroprotective effects of DG. Furthermore, DG significantly decreased the escape latency and search distance and increased the target crossing times of Aβ1–42-induced AD mice in the Morris water maze test. Therefore, these results demonstrated that DG could attenuate Aβ1–42-induced neuronal injury by preventing mitochondrial dysfunction and oxidative stress and improved cognitive impairment in Aβ1–42-induced AD mice, indicating that DG exerted potential beneficial effects on AD.


Introduction
Alzheimer's disease (AD), with typical pathological abnormalities including amyloid plaques, neurofibrillary tangles and neuron death, is the most prevalent neurodegenerative disease [1]. Beta-Amyloid (Ab) is the primary component of senile plaques and Abinduced oxidative stress and neuronal apoptosis play an important role in the pathogenesis of AD [2,3,4]. Persuasive evidence indicates that Ab leads to the mitochondrial dysfunction partially by causing an imbalance of mitochondrial fission/fusion and impairing the mitochondrial biogenesis [5,6,7]. Moreover, Ab was demonstrated to interact with Ab-binding alcohol dehydrogenase (ABAD), which caused the release of reactive oxygen species (ROS), diminished cytochrome c activity and ATP depletion in AD patients and transgenic mice [8,9,10]. Thus, one promising preventive or therapeutic strategy for treatment of AD may be to attenuate or suppress Ab-mediated oxidative stress and mitochondrial dysfunction.
The peroxisome proliferator-activated receptor gamma coactivator 1 (PGC-1) are a small family of transcriptional coactivators which play a critical role in the control of glucose, lipid, and energy metabolism [11]. There are three known isoforms of PGC-1: PGC-1a, PGC-1b and PGC-1-related coactivator. The physiological significance of PGC-1 in mitochondrial energy metabolism has been well demonstrated [12,13]. Several groups including our studies have demonstrated that PGC-1a exerted neuroprotective effects in multiple neurological diseases [14,15,16,17]. An intriguing finding in these studies was that PGC-1a null mice developed spongiform neurodegeneration in selective brain areas, which indicated the direct role of PGC-1a in the neurodegeneration [13]. PGC-1a was a direct target of cyclic AMP (cAMP) response element binding (CREB) in vivo and CREB-dependent gene expression played critical roles in the neuroplasticity associated with cognitive function [18].
Glycyrrhizin (GL), which is extracted in liquorice root, has a wide range of pharmacological actions including anti-virus, antiallergenic and anti-immune-mediated cytotoxicity [19,20,21,22]. Diammonium glycyrrhizinate (DG), which is also extracted and purified from liquorices, is more stable, soluble and has more significant bioactivities than GL. DG has been used for treatment of hepatitis for many years in Asian countries because of its antiinflammatory effect, resistance to biologic oxidation and membranous protection [23,24]. This study demonstrated that DG suppressed Ab 1-42 -induced oxidative stress and mitochondrial dysfunction partially via induction of PGC-1a and alleviated Ab 1-42 -induced cognitive impairment, suggesting DG might be developed into a promising drug for treatment of AD.

Ab 1-42 induced AD mice model
The Ab  (Millipore, CA, USA) was dissolved in 1% NH 3 ?H 2 O at a concentration of 1 mg/ml and incubated at 37uC for 5 days to allow for fibril formation. DG was purchased from Jiangsu Chia-Tai Tianqing Pharmacy Company. The male ICR mice (weight range: 15-20 g) were anesthetized and Ab 1-42 (4 mg, i.c.v) was injected to bilateral hippocampus by infusion cannulae. DG was co-injected intraperitoneally with Ab 1-42 . The mice were randomly assigned into four groups: the normal mice with saline or DG (10 mg/kg/day, i.p. for 14 days), and Ab 1-42 -induced AD mice with saline or DG (10 mg/kg/day, i.p. for 14 days). All animal experiments were approved by the Animal Care Committee in Nanjing University and performed according to institutional guidelines. We made every effort to minimize the number of mice used and their suffering.

Cell culture and treatment
Primary cortical neurons were prepared from E15-17 mouse embryo. Cortexes were dissected and plated at 4610 5 cells/ml on poly-D-lysine-coated plates. Cells were maintained in Neurobasal media supplemented with B27 (Invitrogen, Carlsbad, California, USA) and 25 nM glutamine at 37uC in a humidified 5% CO 2 incubator. The purity of neurons was over 95%. The cells at day 8 were incubated with 2 mM Ab 1-42 with DG or saline for 24 h.
HEK293T, BV-2 and RAW264.7 cells were obtained from American Type Culture Collection (ATCC) and maintained in DMEM containing 10% of heat-inactivated fetal bovine serum (FBS), 2 mmol/L of L-glutamine, 100 U/ml of penicillin, and 100 mg/ml of streptomycin at 37uC in a humidified 5% CO 2 incubator.
The oligonucleotides were synthesized by Biocolor BioScience and Technology Company (Invitrogen, USA). shRNAs were transfected into neurons using Lipofectamine 2000 (Invitrogen, CA, USA) according to the manufacturer's instructions. Cells were harvested for RT-PCR and western blotting at 24 h after the transfection.

Apoptotic assay by flow cytometry
Apoptosis was determined by Annexin V-FITC apoptosis detection kit (KeyGen Biotech, Nanjing, China). After treatment, the cells were rinsed with PBS twice, centrifuged at 600 g for 10 min and resuspended in 0.5 ml binding buffer containing 5 ml Annexin V and 5 ml propidium iodide (PI), and then incubated for 15 min at 37uC in the dark. The apoptotic rate was examined by flow cytometry.

MTT assay
Cell viability was determined using the conventional MTT assay. After treatment, primary cortical neurons were treated with 0.5 mg/ml MTT for 4 h at 37uC. The formazan crystals were dissolved in 100 ml of DMSO and the absorbance was measured at 570 nm in a plate reader. Cell survival rates were expressed as percentages of the value of normal cells.

LDH assay
LDH is the most widely used marker in cytotoxicity study. At the end of incubation, the supernatant was collected from plates and the LDH content was determined using an LDH assay kit according to the manufacturer's instructions (Nanjing Institute of Jianchen Biological Engineering, China). LDH cytotoxicity was calculated by the following formula: LDH cytotoxicity = (sample OD2blank OD)/(standard solution OD2blank standard solution OD )62000.

Measurement of mitochondrial membrane potential
Change of the mitochondrial transmembrane potential in neurons was quantified by JC-1 (Beyotime, Nanjing, China). Briefly, neuronal cells were centrifuged at 600 g for 10 min, and resuspended in 0.5 ml medium containing 5 mM JC-1. After 20 min of incubation at 37uC in the dark, the cells were washed with PBS twice and resuspended in 0.5 ml PBS. Samples were analyzed by flow cytometry.

Measurement of intracellular ROS
To monitor intracellular accumulation of ROS, flow cytometry was used with commercial kit (Beyotime, Nanjing, China) according to the manufacturer's instructions. After treatment, the cells were harvested, rinsed with PBS twice, centrifuged at 600 g for 10 min, and then resuspended in 10 mM DCFH-DA solutions. After 20 min of incubation at 37uC, cells were washed with PBS twice and resuspended in 0.5 ml PBS. Samples were analyzed by flow cytometry.

Measurement of 4-hydroxy-2-trans-nonenal (4-HNE)
The levels of 4-HNE from hippocampus and serum were measured by the ELISA kits (Genmed Scientifics Inc, USA) according to the manufacturer's instruction. Briefly, supernatant from hippocampus or serum were added into the 96-well plate coated with purified anti-4-HNE antibody, and then HRP-labeled 4-HNE antibody was added. The absorbance was measured at 450 nm and the concentration of 4-HNE was determined by comparing the O.D. of the sample to the standard curve.

Cytochrome c detection
For measurement of cytochrome c release, the mitochondrial and cytosol fractions were prepared according to the manufacturer's instructions (Beyotime, Nanjing, China). Briefly, mice hippocampus were washed twice with cold PBS, resuspended in fresh cytosolic extract buffer and incubated for 30 min on ice with frequent tube tapping. Tissues were homogenized on ice, and then nuclei, unbroken cells, and cell debris were pelleted at 600 g for 10 min at 4uC. The supernatant was spun again at 13,000 g for 20 min at 4uC. The supernatant was carefully transferred and the final pellet was used as the mitochondrial fraction. The cytochrome c levels were determined using a monoclonal antibody to cytochrome c by western blotting as described below.

Caspase-9 and -3 activity assay
Caspase-3 and caspase-9 activities of primary cortical neurons were measured by means of colorimetric assay kits (Keygen BioTech, Nanjing, China), according to the manufacturer's instructions. In brief, harvested cells were incubated with 50 ml lysis buffer on ice for 30 min, followed by centrifugation at 10,000 g for 1 min at 4uC. Then, cells were suspended in 50 ml 26reaction buffer and 5 ml caspase-3 or caspase-9 substrate incubating for 4 h at 37uC. Later, the absorbance was read in a microplate reader at 400 nm.
The PCR products were analyzed on 1.5% agarose gels and visualized by ethidium bromide. The gel was visualized with UVtransilluminator and photographed.

Luciferase reporter activity assays
The promoter regions of mouse PGC-1a (23000 to 0 bp) were amplified using PCR, DNAs of primary cortical neurons as templates, and specific primers with MluI and BglII restriction enzyme (Fermentas Inc., USA) cut sites engineered on the ends (Forward: 59-ATAAACGCGTAATGTGTGGCCGAACACAC-TGT-39, Reverse: 59-CGCCGAGATCTAAAGCTATTAAAAA-GTAGGCT-39) to facilitate directional cloning. The PCR products were cloned into the pGL3 basic in sense orientation (designated as p-PGC-3K). The truncated constructs were made using the following primers: All transfection experiments in this study were performed with Lipofectamine 2000 (Invitrogen) following the manufacturer's instructions. pPGCs and phRL-CMV Renilla were cotransfected to cells followed by DG treatment for 24 h. The Luciferase activity was assayed by using the Promega Bright-N-Glo system as previously described [25]. All data points were the averages of at least four independent transfections.

Morris water maze test
The Morris water maze test was conducted as previously described [26]. Briefly, mice were trained to find a transparent Plexiglas platform in the pool placed 2 cm below the water surface in the middle of one quadrant. The position of the platform was unchanged during the training trials. Four time training trails per day were conducted for four consecutive days from 14 days after the injection. In each trial, the latency to escape on the platform was recorded for 1 min. Data of each mice behavior were collected by a video camera linked to a computer through an image analyzer. The total sum of latency and searching distance for the platform in four trials of each mouse was counted for all tested mice per group per day. At the end of the training period, mice were tested on a spatial probe trial in which the platform was removed from the pool, and each mouse was allowed to swim freely for 1 min. During the probe trial, the number of platform crossings was recorded. The recorded data were used to analyze mice performance.

Statistical analysis
The data were expressed as means 6 SEM and analyzed by SPSS12.0 statistical analytical software (SPSS, Chicago, IL, USA). Group differences in the escape latency, searching distance and swimming speed during the Morris water maze test were analyzed using two-way analysis of variance (ANOVA) with repeated measures followed by Bonferroni post hoc test with day and treatment as the sources of variation. Otherwise comparison between two groups was statistically evaluated by Student's t-test and multiple group comparisons were analyzed by one-way ANOVA followed by Tukey post hoc test. Values of P,0.05 were considered statistically significant.

DG protects neurons from Ab 1-42 -induced neurotoxicity in vitro
To investigate whether DG could suppress the cellular toxicity induced by Ab, the primary cortical neurons were incubated with Ab 1-42 (2 mM) and different concentrations of DG or saline. As expected, the viability of cortical neurons exposed to Ab 1-42 was reduced by 34.1% in comparison with the control group (P,0.01, Figure 1A) and DG significantly enhanced neuron viability (P,0.05, Figure 1A).
To further confirm the neuroprotection of DG on Ab 1-42mediated toxicity, neuronal apoptosis was detected by Annexin V and PI staining. As shown in Figure 1B and 1C, cell apoptosis was demonstrated in Ab 1-42 -treated neurons compared with the control, while significantly attenuated after treatment of DG. In addition, the inhibition of Ab 1-42 -induced neuronal death in presence of DG was confirmed by the LDH assay ( Figure 1D).

DG decreases the oxidative damage induced by Ab 1-42
Emerging evidence suggests that mitochondrial dysfunction and oxidative stress are involved in Ab 1-42 -induced neurotoxicity. Thus, this study surmised that DG might be able to reduce Ab 1-42mediated mitochondrial dysfunction and excessive production of ROS which was mainly produced by mitochondria. As shown in Figure 2A and 2B, exposure of cortical neurons to Ab 1-42 led to an increase in ROS production; whereas the effect was significantly decreased by treatment with DG (1365.67667.52 vs. 705.67651.87, P,0.01). To further investigate DG's ability to inhibit Ab-induced oxidative stress, a marker of lipid peroxidation, 4-HNE was examined. As shown in Figure 2C and 2D, 4-HNE levels in the serum and hippocampus of Ab 1-42 -induced AD mice were significantly increased by 40.4% and 67.3% compared to control mice respectively (P,0.05), while 4-HNE levels were reduced 24.2% and 33.2% in the serum and hippocampus after DG treatment respectively (P,0.05).

DG prevents mitochondrial dysfunction mediated by Ab 1-42
Mitochondrial membrane potential (Dy) is widely recognized as an indicator of mitochondrial functionality, which is measured by JC-1, a cationic lipophilic fluorescent. The results showed that there was a significant loss of Dy in neurons treated with Ab 1-42 ( Figure 3A and 3B, P,0.01). However, the decrease of Dy induced by Ab 1-42 was greatly alleviated after DG treatment ( Figure 3A and 3B, P,0.01), indicating that DG protected mitochondrial against Ab 1-42 -induced injury. Meanwhile, the activities of caspase -9 and caspase-3, were assessed. As shown in Figure 3C and 3D, the activities of caspase-9 and caspase-3 were significantly increased by 44.80% and 68.17% of the control group in Ab 1-42 -treated neurons, while DG-treated neurons exhibited lower caspase-9 and caspase-3 activities compared to Ab 1-42treated neurons (P,0.05). To further explore the role of DG against mitochondrial dysfunction induced by Ab in vivo, the release of cytochrome c from the mitochondrial membrane as well as the subsequent activation of caspase-9 and caspase-3 was investigated by western blotting. The levels of cytosolic cytochrome c expression in Ab 1-42induced AD mice were significantly increased, which were significantly reversed by the treatment with DG ( Figure 3E and 3F). In addition, DG could inhibit the activation of caspase-9 and caspase-3 in Ab 1-42 -induced AD mice ( Figure 3E and 3F).   Figure 3E. Results were shown as the mean6 SEM and represented at least three independent experiments. * P,0.05 and ** P,0.01 for one-way ANOVA followed by Tukey post hoc test compared with control, respectively; # P,0.05 and ## P,0.01 for oneway ANOVA followed by Tukey post hoc test compared with Ab 1-42 -treated, respectively. n = 4 mice per group. doi:10.1371/journal.pone.0035823.g003 treatment could significantly increase its expression ( Figure 4D and 4E). To demonstrate whether PGC-1a contributed to the neuroprotection of DG, endogenous PGC-1a was knocked down by shRNAs (Figure 5A to 5D) and the results of MTT revealed that PGC1a-shRNA could partially block neuroprotective effects by DG in Ab 1-42 -treated neurons ( Figure 5E).
To explore whether DG could induce the transcriptional activity of PGC-1a, p-PGC-3K and five truncated plasmids were constructed. As shown in Figure 6A, DG treatment significantly up-regulated the transcriptional activity of PGC-1a by 6.23-fold at a concentration of 0.001 mg/ml, and 5.13-fold at a concentration of 0.005 mg/ml in neurons (P,0.01). Also DG treatment significantly up-regulated the transcriptional activity of PGC-1a by 4.08-fold in HEK293T cells, 1.98-fold in BV-2 cells and 1.62-fold in RAW264.7 cells at a concentration of 0.001 mg/ml, indicating that induction of transcriptional activity of PGC-1a by DG may not be cell type specific ( Figure 6B). Interestingly, DG significantly downregulated the transcriptional activity of p-PGC-2.5K while increasing the transcriptional activity of p-PGC-3K (P,0.01), which indicated that 2500-0 bp in the promoter of PGC-1a might be involved in the protective effects of DG ( Figure 6C). To explore whether the CREB binding sequence (284-77 bp) was essential for the DG-induced transcriptional activity of PGC-1a, 2100-0 bp of the PGC-1a promoter and the CREB binding sites mutated/deleted sequences were constructed. As shown in Figure 6D, mutation or deletion of the CREB binding sequence completely abolished the induction of the transcriptional activity of PGC-1a, which suggested that CREB might play an important role in the DG-induced transcriptional activity of PGC-1a.

DG improves cognitive impairment in Ab 1-42 -induced AD mice
To explore whether DG could improve cognitive impairment in Ab 1-42 -induced AD mice, Morris water maze test was employed. Escape latency reflects the ability of learning and remembering the relationships between multiple distal cues and the platform location to escape the water, which is a hippocampus-dependent task. As shown in Figure 7A, DG could decrease the mean latency reaching to the submerged platform of AD mice (two-way ANOVA with repeated measures; groups: F(3, 31) = 6.123, P = 0.002; days: F(3, 93) = 3.620, P = 0.016; group x day: F(9, 93) = 0.915, P = 0.516). In addition, the search distance was also significantly decreased by DG treatment compared to AD mice (two-way ANOVA with repeated measures; groups: F(3,  Figure 7B). On the fifth day, the platform was removed and the probe trail was conducted. AD mice had fewer times crossing the previous platform position than the normal mice, while those under the treatment of DG significantly improved their performance (P,0.05, Figure 7C). No speed differences appeared among these four groups ( Figure 7D). It suggested that DG could alleviate the deficits of spatial learning and memory in Ab 1-42 -induced AD mice.

Discussion
In the Ab 1-42 -induced AD model in vitro and in vivo, this study for the first time shows: 1) DG exerts neuroprotective effects and improves cognitive impairment; 2) DG rescues mitochondrial dysfunction and inhibits oxidative stress; and 3) DG increases the expression of PGC-1a, which might contribute to the neuroprotection of DG.
Mitochondrial dysfunction is a hallmark of Ab 1-42 -induced neuronal toxicity, and is considered as an early event in AD pathology. Several evidences indicated that Ab triggered mitochondrial dysfunction through a number of pathways such as increase of ROS, interaction with ABAD, impaired mitochondrial biogenesis, and alteration of mitochondrial dynamics [5,6,7,10]. DG is the salt form of glycyrrhizin, a major active constituent isolated from licorice. Licorice and glycyrrhizin have anti-oxygenic and anti-inflammatory action in bile acid-induced apoptosis and necrosis [22] [27,28]. The route of administration of DG was selected because of good bioavailability reported previously [29], and our results had shown that DG did not attenuate cognitive impairment in AD mice at high dose (50-100 mg/kg) (data not shown). This study indicated that DG could decrease the accumulation of ROS, rescue the mitochondrial membrane potential loss and activation of caspase-9 and caspase-3 in Ab 1-42 -treated neurons. In addition, DG decreased lipid peroxidation and release of cytochrome c from the mitochondria, and the activation of caspase-9 and caspase-3 in Ab 1-42 -induced AD mice. Furthermore, this anti-oxidation function of DG could refrain neurotoxicity mediated by Ab 1-42 , that is, increased cell viability, decreased apoptosis and LDH release in Ab 1-42 -treated neurons.
Regardless of the possible mechanisms of DG restraining oxidative stress, it is clear that PGC-1a is a major regulator of mitochondrial biogenesis and is protective against oxidative damage [30]. In addition, PGC-1, which could bind to and activate many other transcription factors, played an essential role in physiological signaling transduction and gene expression [31]. However, the role of PGC-1a in the neurological diseases was not extensive studied until recently. PGC-1a null mice were much more sensitive to the neurodegenerative effects of MPTP and kainic acid, which were oxidative stressors. Increasing PGC-1a levels dramatically protected neurons from oxidative-stressormediated death [17]. Resveratrol was an ideal compound for treating neurodegenerative diseases by increased the activity of numerous proteins, including PGC-1a [32]. Activation or overexpression of PGC-1a could be used to compensate for neuronal mitochondrial loss [16]. Expression levels of PGC-1a were significantly decreased in both AD hippocampus and M17 cells stably expressing human Swedish mutation APP695 [5] . Consistent with these reports, our data showed that DG increased the expression of PGC-1a and that knocking down PGC-1a by shRNAs could block the neuroprotection of DG, suggesting that PGC-1a contributed to the neuroprotective effect of DG in Ab 1-42 -treated neurons.
It is intriguing that DG treatment increased the expression of PGC-1a in Ab 1-42 -treated neurons and CREB might play an important role in induction of the transcriptional activity of PGC-1a. PGC-1a was a direct target of CREB induction of gluconeogenesis in vivo [33]. CREB was also essential for long- Figure 6. DG might induce the transcriptional activity of PGC-1a through CREB. (A) pPGC-3K and phRL-CMV Renilla were cotransfected to neurons followed by DG treatment at the indicated concentration and the luciferase activity was assayed at 24 h. Results are shown as the mean6 SEM and represent at least four independent experiments. ** P,0.01 for one-way ANOVA followed by Tukey post hoc test compared with control. (B) pPGC-3K and phRL-CMV Renilla were cotransfected to neurons, HEK293T, BV-2, or RAW264.7 cells followed by DG treatment (0.001 mg/ml) and the luciferase activity was assayed at 24 h. ** P,0.01 for Student's t-test compared with pGL-3. (C) pPGCs and phRL-CMV Renilla were cotransfected to HEK293T cells followed by DG treatment (0.001 mg/ml) and the luciferase activity was assayed at 24 h. (D) pPGC-100 bp or mutated pPGC-100 bp or deleted pPGC-100 bp were cotransfected to HEK293T cells with phRL-CMV Renilla followed by DG treatment (0.001 mg/ml) and the luciferase activity was assayed at 24 h. Results are shown as the mean6 SEM and represent at least four independent experiments. * P,0.05 and ** P,0.01 for one-way ANOVA followed by Tukey post hoc test compared with pGL-3, respectively; ## P,0.01 for one-way ANOVA followed by Tukey post hoc test compared with pPGC-100 bp. doi:10.1371/journal.pone.0035823.g006 lasting changes in synaptic plasticity that mediates the conversion of short-term memory to long-term memory. Ab altered hippocampal-dependent synaptic plasticity and memory storage and mediated synapse loss through the CREB signaling pathway, which suggested a crucial role of CREB signaling in cognitive dysfunction [34]. Therefore, further studies are needed to demonstrate the exact role of CREB to the transcriptional activity of PGC-1a and neuroprotective effects of DG.
Taken together, DG exerted neuroprotective effects against Ab 1-42 -induced toxicity in vitro and in vivo. DG significantly increased the viability of Ab 1-42 -treated neurons by inhibiting oxidative stress and reversing mitochondrial dysfunction. Furthermore, PGC-1a upregulated by DG treatment might play an important role against Ab 1-42 -induced neurotoxicity. Findings of current study revealed new function and mechanism of DG on neurotoxicity induced by Ab 1-42 , suggesting that DG may be developed into a new drug for treatment of AD. Escape latency for escape to a submerged platform in the training trials. (B) Searching distance for escape to a submerged platform in the training trials. * P,0.05 and ** P,0.01 for two-way ANOVA with repeated measures followed by Bonferroni post hoc test compared with Ab 1-42 -treated, respectively. (C) 24 h after the training trials platform crossing times were recorded. * P,0.05 for one-way ANOVA followed by Tukey post hoc test compared with control; # P,0.05 for one-way ANOVA followed by Tukey post hoc test compared with Ab 1-42 -treated. (D) Swimming speed in the training trials. con: normal mice; DG: normal mice with DG (10 mg/kg/day, i.p. for 14 days); Ab: Ab 1-42 -induced AD mice; Ab+DG: Ab 1-42 -induced AD mice with DG (10 mg/kg/day, i.p. for 14 days). n = 10 mice per group. doi:10.1371/journal.pone.0035823.g007