Dianxianning improved amyloid β-induced pathological characteristics partially through DAF-2/DAF-16 insulin like pathway in transgenic C. elegans

Dianxianning (DXN) is a traditional Chinese formula, and has been approved in China for treating epilepsy since 1996. Here anti-Alzheimer’s disease activity of DXN has been reported. DXN improved AD-like symptoms of paralysis and 5-HT sensitivity of transgenic Aβ1-42 C. elegans. In worms, DXN significantly increased Aβ monomers and decreased the toxic Aβ oligomers, thus reducing Aβ toxicity. DXN significantly suppressed the expression of hsp-16.2 induced by juglone, and up-regulated sod-3 expression. These results indicated that DXN increased stress resistance and protected C. elegans against oxidative stress. Furthermore, DXN could significantly promote DAF-16 nuclear translocation, but it did not activate SKN-1. The inhibitory effect of DXN on the Aβ toxicity was significantly reverted by daf-16 RNAi, rather than skn-1 RNAi or hsf-1 RNAi. These results indicated that DAF-16 is at least partially required for the anti-AD effect of DXN. In conclusion, DXN improved Aβ-induced pathological characteristics partially through DAF-2/DAF-16 insulin like pathway in transgenic worms. Together with our data obtained by Morris water maze test, the results showed that DXN markedly ameliorated cognitive performance impairment induced by scopolamine in mice. All the results support that DXN is a potential drug candidate to treat Alzheimer’s diseases.

. The effects of DXN on AD-like symptoms in Aβ transgenic C. elegans. (A) DXN delayed AD worm paralysis induced by Aβ over-expression in muscle tissues in smg-1 (cc546)I temperature sensitive Aβ strain CL4176. Data are the average of three replicates with about 180 worms in each group. ***Indicated that there was significant difference between treatment group and control group at P < 0.001. (B) DXN ameliorated 5-HT hypersensitivity induced by Aβ over-expression in nerve cells in snb-1/Aβ1-42 strain CL2355. Worm strain CL2122 without Aβ expression in nerve cells was used as a transgenic control. Data are the average of three replicates with about 90 worms in each group. There is significant difference among these groups when symbols are different (P < 0.05).
paralysis. C. elegans snb-1/Aβ 1-42 strain CL2355 shows serotonin hypersensitivity when Aβ are over-expressed in nerve system 22,23 . Here we investigated whether DXN treatment could reverse serotonin hypersensitivity induced by the toxicity of Aβ to nerve cells in the transgenic snb-1/Aβ 1-42 worms. Our results showed that snb-1/Aβ  worms were hypersensitive to exogenous 5-HT, but CL2122 without Aβ expression in nerve cells were not. DXN significantly alleviated the paralysis symptom of hypersensitive response to 5-HT in a dose-dependent manner (Fig. 1B).
Scopolamine is an anti-cholinergic agent to induce aspects of the memory loss and cognitive impairment, which resemble the hallmark symptoms observed in AD associated with cholinergic dysfunction at least on some degree. The scopolamine model has been widely used for screening possible AD symptom therapeutic agents to improve cognitive performances 24 . In present work, this model was used to evaluate the effects of DXN on learning and memory impairment by the Morris water maze test. Compared with the normal control group, the memory training results indicated that the mean escape latencies were significantly lengthened in scopolamine treated mice on the 5 th day of the experiment. DXN improved learning and memory performance of mice as piracetam (Table 1). In probe test, the swimming time in the target quadrant of DXN treated animals significantly lengthened in contrast to the model group ( Fig. 2A). The platform frequency remarkably increased in DXN and piracetam treated groups (Fig. 2B). Compared to model group, the length of swimming path between drug treatment groups and normal control group were not significant in probe test (Fig. 2C). These results suggested that DXN indeed promoted learning and memory performance in scopolamine treated mice.

DXN exerted anti-AD like action by its principal component Valeriana jatamansi Rhizoma et
Radix. DXN is a kind of traditional Chinese medicine, and it is composed of eight herbs. Previous works have suggested that three herbs of Valerianae jatamansi Rhizoma et Radix (Vj), Rhizoma Acori tatarinowii (Ra) and Ramulus Uncariae cum Uncis (Ru) in DXN have anti-AD activity 1,[14][15][16][17] . In present work, Vj, Ra and Ru at an equivalent dose in DXN were firstly used to tested whether they can ameliorate AD-like symptom of delaying paralysis in worms. We further tried to investigate that whether DXN can be reduced to one or more herbs which can function as the principal component, and finally some effective compounds based anti-Aβ therapy for treating AD can be extracted from the right principal herbs.
The results showed that Vj delayed worm paralysis as DXN complete formula to a similar degree. Ra and Ru significantly alleviated the AD-like symptom in worms to a much lesser degree (Fig. 3A). These results supported that Vj was the principal component of DXN. It is deserved to notice that Vj alone also significantly shortened worm body length, but DXN did not (Fig. 3B). In contrast to drug treatment from eggs, worms were administrated with DXN or Vj when incubation temperature was upshifted to induce Aβ expression, and at that time animals were on the same development stage. The results showed that DXN and Vj did delay worm paralysis, but to a less degree (Fig. 3C). Of course, shorter dosing time and less DXN and Vj entry into worms led the decline in anti-AD activity of DXN and Vj for delaying paralysis. Anyway, the effect of Vj largely decreased than that of DXN did (Fig. 3A,C). Taken together, those results supported that complete DXN was a rational formula as it is and it should not be reduced. Additionally, transgenic worm smg-1[myo-3::GFP] was further used to test whether DXN treatment non-specifically affected gene expressions, the results showed that neither DXN nor Vj treatment altered exogenous gfp gene expression (Fig. 3D,E), suggesting that DXN and Vj treatment possible inhibiting non-specific protein expression could be rule out.
Iridoids and sesquiterpenoid derived from Vj has been reported to have neuroprotective activity 25 . Here, we further evaluated compound 1-15 extracted from Vj (Fig. 4), and treated worms with each compound at relative final concentration due to their different maximum saturated solubility in DMSO. In contrast to Vj, almost all of them delayed worm paralysis induced by over-expression of Aβ to a much lesser degree, except that 100 μM of compound 15 did not have any anti-AD activity, 250 μM of compound 11 significantly promoted worm paralysis (Fig. 5). It is deserved to notice that compound 6 is baldrinal, which is regarded as the marker substance of Vj and DXN, and its content is 0.930 mg per gram of DXN or Vj coarse powder. In Fig. 5, 160 µM baldrinal was at a dosage higher by 2.6-fold than that contained in Vj used. Previous work showed that an iridoid, 10-O-trans-p-coumaroylcatalpol at similar dose or less significantly inhibits α-synuclein aggregation in worms 26 . So that, it cannot be excluded that more active compound will be isolated from Vj in the future work. Because relative large amount of baldrinal is contained in Vj, its anti-AD activity is still dissecting in our lab. At present, it is reasonable to believe that DXN needs complete Vj. DXN reduced Aβ oligomers rather than amyloid deposits in transgenic C. elegans. To explore whether the DXN affects the aggregation of the large Aβ species, unc-54/human Aβ 1-42 worm strain CL2006 that constituently over expresses human Aβ 1-42 in muscle cells was used. Although memantine is clinically used as a selected NMDA receptor antagonist, evidence recently indicates that it lowers the Aβ level and decreases Aβ plaques to support memantine may be a disease-modifying drug for treating AD [27][28][29][30] . Therefore, memantine has been used as a positive control. We found the DXN had no remarkable effect on the deposits of the Aβ 1-42 peptide, suggesting that the DXN could not decrease the aggregation of the Aβ species significantly (Fig. 6A,B). As expected, memantine significantly lowered Aβ deposits (Fig. 6A,B). Considering the current evidence shows that soluble Aβ oligomers, rather than senile plaques (SPs) or Aβ fibrils, are the toxic species to link to    neurodegeneration and cognitive decline 31, 32 , we performed the Western blotting analyses to test whether the DXN can reduce the toxic Aβ oligomers. Primary antibody of 6E10 against Aβ was used in western blotting analysis, and the relative densities of Aβ monomer of 4 kDa band and higher molecular weight oligomer of 20 kDa band were quantified 22 . The result showed that DXN significantly reduced the level of Aβ oligomers and increased the level of Aβ monomers (Fig. 6C,D). It indicated that DXN could significantly change the Aβ oligomers with high neurotoxicity into Aβ monomers, which is much less toxic than Aβ oligomers.

DXN delaying paralysis of AD worms was partially mediated by DAF-16 activation. DAF-16
is a crucial component of DAF-2 insulin like pathway to regulate many gene expressions for defending against various stresses including oxidative stress and proteotoxicity 33 , its activation has been proved to be benefit for delaying paralysis induced by human Aβ excessive expression in the muscle in C. elegans 34 . To examine whether DAF-16 is required for DXN protecting worms from severe stress of aberrant Aβ generation and accumulation, we knocked down expression of the transcription factor DAF-16 by RNAi, and found that the inhibitory effect of DXN on worm paralysis significantly decreased (Fig. 7A). We further constructed a daf-16 mutant strain under genetic background of the temperature sensitive Aβ strain CL4176 by standard genetic crosses. Treating the AD worms with daf-16 mutation, a similar result has been obtained (Fig. 7B). These results indicated that DXN reduced Aβ toxicity at least partially through a DAF-16-based mechanism. After DXN treatment, DAF-16 was indeed activated and displayed nuclear translocation from the cytosol to the nucleus (Fig. 7C,D). Further, its typical downstream target SOD-3 was markedly up-regulated (Fig. 8A,B). According to our knowledge, DAF-16 activation increases stress resistance and then leads to lengthen lifespan. Expectedly, the result of lifespan assay on unc-54/human Aβ 1-42 worm strain CL2006 treated with DXN showed that DXN did significantly extend worm lifespan (Fig. 3S).
As a chaperon, HSP-16.2 can be induced by the abnormal Aβ proteins to increase their sequestration, degradation, and refolding 35 . In our experiment, the expression of hsp-16.2 was examined after DXN pretreatment by using TJ375. The transgenic worm strain TJ375 has a GFP reporter gene controlled by the promoter of hsp-16.2, it exhibits an excessive GFP expression in the pharynx after exposure to oxidative stress such as juglone. The results showed that DXN alone did not significantly affect the hsp-16.2::gfp expression. However, when C. elegans were pretreated with DXN and further administrated with juglone, the fluorescence density in the pharynx of worms under juglone-induced oxidative stress significantly decreased compared to the DXN untreated group (Fig. 8C,D). We further treated AD worms with RNAi for sod-3 and hsp-16.2 (Fig. 8E). The results showed that sod-3 RNAi significantly promoted worm paralysis suggesting that SOD-3 is required for battling against Aβ toxicity. It is seemed to find that DXN delayed AD-like symptoms to the same degree as DXN treatment alone without sod-3 RNAi, but it did not indicate that SOD-3 was not required for DXN-mediated neuroprotection. Possibly, the compensatory upregulation of other sod genes or related detoxifying signaling pathway counteracted the effect of sod-3 RNAi on DXN-mediated delaying worm paralysis 36 . As for hsp16.2 RNAi, no obvious effect was observed. It might be explained by that hsp-16.2 is only expressed under stress 37 .
Transcription factor SKN-1 modulates genes that fight against oxidative and xenobiotic stress and genes that regulate protein homeostasis 33 . After skn-1 RNAi, the Aβ-induced paralysis phenotype in the temperature sensitive Aβ strain CL4176 was still significantly delayed by DXN treatment (Fig. 1S-A), indicating that SKN-1 was not involved in the protective effect of DXN. It was further validated by SKN-1 did not appear nuclear translocation at all (Fig. 1S-B).
HSF-1 regulates many gene expressions in response to stress 38 , and has been recognized to control disaggregation and degradation of Aβ 34 . In present work, after hsf-1 RNAi, the Aβ-induced paralysis phenotype in the temperature sensitive Aβ strain CL4176 was still significantly delayed by DXN treatment (Fig. 1S-C), indicating that HSF-1 was not involved in the protective effect of DXN. In present work, RNAi for hsf-1 and skn-1 significantly promoted worm thermal stress sensitivity (Fig. 1S-D). Therefore, the lack of effect of RNAi on worm paralysis after DXN treatment could not result from that the RNAi treatment did not work.

Discussion
DXN has been clinically used to treat epilepsy in China since 1990s. In present study, it strongly ameliorated Aβ-induced paralysis and 5-HT hypersensitivity in the transgenic C. elegans in a dose dependent manner (Fig. 1). Furthermore, DXN significantly alleviated learning and memory impairment induced by scopolamine in mice (Fig. 2). Those results indicated that DXN was potential to serve as a drug candidate for treating AD. In fact, we also tested the anti-AD activity of other two traditional Chinese formulas of Dianxianping tablet (DXP) and Guishaozhenxian tablet (GS), these two formulas also have been approved by CFDA for treating epilepsy. Our results showed that DXP and GS are less effective in comparison to DXN at their concentration available (Fig. 2S), and both of them could not be used on a larger dosage due to that they have severely delayed worm growth and development at their concentration in present work. Therefore, although there is close link between epilepsy and AD, whether an anti-epileptic drug can be used for treating AD still needs to be carefully investigated in details.
In present work, Vj has been proved to exert its delaying paralysis of AD-like symptom as a principal component (Fig. 3A). Further, compounds of 1-15 isolated from Vj significantly delayed worm paralysis and have anti-AD activity on a certain degree, except that compound 15 did not affect worm paralysis and compound 11 even aggravated AD-like symptom. In comparison with the complete herb of Vj, they only exhibited moderate protective effect (Fig. 5). Similarly, coffee extract can protect worms from Aβ toxicity, whereas caffeine does only ***Indicated that there was significant difference between treatment group and control group at P < 0.001. There is significant difference among these groups when symbols are different (P < 0.05).
to a much lesser degree 19 . Previously, Vj has been demonstrated to have anti-inflammatory and anti-oxidant activity 14,15 . Several other compounds of iridoids, sesquiterpenoids extracted from this species and other related species of genus Valeriana have been shown to have neuroprotective effects 25,39 . It is reasonable to conclude that these compounds can exert additive or synergistic effect for anti-AD. Of course, it cannot be excluded that more active compound will be isolated from Vj in the future work.
DXN, as a kind of traditional Chinese medicine, is composed of eight herbs. Among the herbs, besides Vj, Ra and Ru have anti-AD activity 1, 14-17 . Our results show that Ra and Ru did not ameliorate AD-like symptom in worms additively or synergistically (Fig. 3A), but complete DXN did not delay nematode growth any more (Fig. 3B), indicating that DXN is less toxic than Vj alone. It is especially important for a potential anti-AD drug to act effectively to be used just at asymptomatic and preclinical stage, and the disease manifests its clinical symptoms after a decade or more, thus anti-AD therapy being a long term prevention 40 . The present evidence supported that DXN could not be reduced. Further work needs to investigate on the actions of these herbs through a feasible AD model in vivo or in vitro based on other respects of the pathogenesis of this devastating disease.
After DXN treatment, Aβ oligomers significantly reduced, but Aβ monomers increased in AD worms (Fig. 6C,D). It is well recognized that Aβ oligomers induce the death of neurons and are responsible for AD-related memory loss. Moreover, Aβ oligomers are closely related to severity of dementia in AD patients 1,39 . Accumulating evidence suggests that Aβ oligomers but not Aβ monomers or Aβ deposits are correlated with Aβ toxicity 41,42 . Our results supported that DXN strongly ameliorating Aβ toxicity in the transgenic C. elegans overexpressed human Aβ 1-42 might be a consequence of its directly or indirectly promoting the shift from the toxic Aβ oligomer to less-toxic monomer form.
Small HSPs are Low-molecular-weight Heat Shock Proteins (12-43 kD) in response to a series of injuries including thermal stress and oxidative stress 43 . Moreover, small HSPs also participate in preventing the accumulation of several different types of toxic proteins such as Aβ and ployQ 44 . HSP16.2 can suppress Aβ toxicity by assisting abnormal protein sequestration, degradation, and refolding in AD C. elegans 35 . Here, DXN treatment did not increase the expression of hsp-16.2 (Fig. 8C,D), indicating that toxic Aβ oligomer reduction after DXN treatment was not mediated by molecular chaperon HSP-16.2. Then whether DXN can directly bind to Aβ proteins to inhibit their aggregation and increase their less toxic monomers needs further investigated. However, DXN counteracted the up-regulation of hsp-16.2 induced by juglone (Fig. 8C,D). Since juglone is a reactive oxygen species generator 45 , it gave us a notion that DXN could reduce Aβ toxicity through anti-oxidant activity. It agrees with the previous work that components of DXN, Vj and Rhizoma Acori tatarinowii possess high antioxidant activity 14, 46, 47 . ***Indicated that there was significant difference between treatment groups and control group at P < 0.001. SOD-3 protects C. elegans from oxidative stress and participates in suppressing the Aβ toxicity 48,49 . In present work, our results showed that DXN indeed promoted the expression of sod-3 (Fig. 8A,B). It supported that DXN increased organism stress resistance and armed C. elegans to defend against oxidative stress induced by aberrant Aβ proteins 50 . SOD-3 locates downstream of DAF-16 in C. elegans 51,52 . By turns, DAF-16 is a FOXO transcription factor and acts as a crucial component in the DAF-2/DAF-16 insulin like signaling pathway 34,53 . The activation of DAF-16 will lead to elevated expression of genes in responses to a wide range of stressors 54,55 . Our results showed that DXN could significantly activate DAF-16 to translocate from cytosol to nucleus (Fig. 7C,D), and DAF-16 is required for the delaying paralysis effect of DXN (Fig. 7A,B). HSF-1 and SKN-1 are transcriptional factors in parallel to DAF-16, which also can promote the expression of stress responsive genes to inhibit the toxicity of Aβ 33,34,38 . In our experiment, hsf-1 and skn-1 RNAi did not affect the DXN delaying paralysis of anti-AD action (Fig. 1SA,B). Additionally, SKN-1 did not exhibit nuclear translocation after DXN treatment (Fig. 1SC). In conclusion, as shown in Fig. 9, DXN suppressing Aβ-induced pathological behaviors was at least partially mediated by DAF-16 activation in worms. DXN also directly or indirectly reduced Aβ oligomers and increased less toxic Aβ monomers. Paralysis assay. Transgenic C. elegans of temperature sensitive Aβ strain (CL4176) maintained at 16 °C were egg-synchronized onto the NGM plates containing with 0, 1 mg/mL, 5 mg/mL, 15 mg/mL DXN. Other principal herbs of Vj, Ra and Ru were used at an equivalent dose as they were in DXN, and 3.2 mg/mL Vj, 3.7 mg/mL Ra and 0.6 mg/mL Ru was used, respectively. For compounds of 1-15 extracted from Vj, they were all diluted at indicated final concentrations in 0.1%DMSO according to their maximum saturated solubility. Compounds of 1-7 were relatively used at a final concentration of 50 μM, 100 μM, 20 μM, 20 μM, 20 μM, 160 μM, 120 μM, compounds of 8-11 were used at 250 μM and compounds of 12-15 were used at 100 μM. Worms were induced to express human Aβ 1-42 till they were at L3 stage larvae by up-shifting the incubation temperature from 16 °C to 25 °C for 34 h. The transgenic worms were scored at 2 h intervals for paralysis till all animals in negative control group were Figure 9. Schematic diagram of DXN action on suppressing Aβ toxicity. DXN directly or indirectly reduced Aβ oligomers and increased less toxic Aβ monomers, further, DXN induced DAF-16 activation to regulate its downstream stress responsive gene expressions, such as sod-3, hsp-16.2, at least partially to exert its anti-AD activity.

Preparation
paralyzed. Each worm was gently touched with a platinum loop to identify as the paralysis, nematodes were considered to be paralyzed if they did not move at all or only moved their heads 58 .
Fluorescence staining of Aβ deposits. Transgenic C. elegans with unc-54/human Aβ 1-42 maintained at 20 °C were egg-synchronized onto NGM plates and grown until the L4 stage, worms were then moved to fresh NGM plates containing 0, 1 mg/mL, 5 mg/mL, 15 mg/mL DXN. Two days later, Thioflavine-S (ThS) staining was performed as described previously 22 . Briefly, worms were collected by washing with M9 and were fixed in 4% paraformaldehyde in PBS, pH7.4, at 4 °C for 24 h. The fixative solution was replaced by permeabilization solution (5% fresh β-mercaptoethanol, 1% Triton X-100, 125 mM Tris-HCl, pH7.4), and incubated in a 37 °C incubator for another 24 h. The animals were washed three times in PBS-T (PBS plus 0.1% Triton X-100), stained with 0.125% ThS (Sigma) in 50% ethanol for 2 min, and the samples were destained with 50% ethanol twice for 2 min. Stained samples were resuspended in 100 μL PBS, and finally mounted and observed by fluorescence microscopy (BX53; Olympus, Japan). Amyloid deposits in the anterior area of each animal were quantified by scoring the number of ThS reactive deposits. C. elegans N2 were used as amyloid deposit control.
Exogenous serotonin sensitivity assay. After egg-synchronized, the transgenic snb-1/Aβ 1-42 worm were placed at 20 °C on NGM plates seeded with OP50 for 48 h. The worms were treated with 0, 1 mg/mL, 5 mg/mL, 15 mg/mL DXN for another 48 h, respectively. 30 worms in each group were washed with M9 buffer for three times and were transferred into 200 μL M9 buffer containing 1 mg serotonin (Sigma), paralyzed worms were scored after 5 min. Animals were considered to be paralyzed if they did not move at all within 5 sec. Worm strain CL2122 without Aβ expression in nerve cells was used as a transgenic control.
Western blotting. Human Aβ 1-42 in the temperature sensitive Aβ strain (CL4176) was identified by immunoblotting on a Tris-Tricine gel by the standard Western blotting assay. Wild type N2 was used as negative control of the expression of Aβ. Worms were treated as paralysis assay with or without 15 mg/mL DXN. Nematodes were then washed three times in M9 to remove the bacteria after incubation temperature up-shifting for 40 h and finally harvested in distilled water with protease inhibitor cocktail (Sigma, P2714) added, quickly frozen in liquid nitrogen and stored at -80 °C. Worms were boiled in sample lysis buffer (1× protease inhibitor cocktail, 62 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 4% β-mercaptoethanol) for 10 min, then seated on ice to cool, centrifuged at 14,000 g for 5 min. Total protein content in the supernatant was measured by Lowry method. Proteins were boiled for 5 min in sample loading buffer (62 mM Tris-HCl, pH6.8, 2% SDS, 10% glycerol, 4% β-mercaptoethanol, 0.0005% bromophenol blue). Samples were performed electrophoresis on the Tris-Tricine gel with 70 μg total protein in each lane. After the gel was transferred to 0.22μm PVDF membrane, the membrane was boiled in PBS for 15 min before blocking with a solution of 5% milk in TBS-Tween (100 mM Tris-HCl, pH7.5, 150 mM NaCl, 0.1% Tween-20) for 1 h at room temperature. Blotting was carried out overnight at 4 °C. Aβ proteins were detected by 6E10 (Biolegend, 803001) at 1:1000 dilution. Internal control of α-tubulin was probed by polyclonal antibody against it (Sigma, T6199) at 1:1000 dilution. Goat anti-mouse HRP-labelled IgG was used as a secondary antibody.
Nuclear localization of DAF-16. Worm daf-16::gfp (TJ356) expressing GFP as a reporter was used for detecting the intracellular distribution of DAF-16 59 . Synchronized L1 larvae worms were fed with or without 15 mg/mL DXN for 72 h. Subsequent to this treatment, worms were mounted on glass slides and their nuclear localization of DAF-16 were observed under a fluorescence microscope (BX53; Olympus, Japan). Worms are scored based on two categories of cytosolic and nuclear distribution with respect to the major localization of the DAF-16::GFP fusion protein. DAF-16 intracellular location was indicated by the ratio of nuclear translocation in the experimental population.

RNA interference (RNAi).
RNAi gene expression clones were constructed as described by Fraser et al. 60 .
Briefly, PCR products were synthesized using Q5 High-Fidelity DNA Polymerase (New England Biolabs, USA) with genomic DNA as template by relative primers listed in Table 2 Table 2. List of primers for amplifying target genes.
strain using a standard method. The desired positive colonies were verified by sequencing. The sequences of target genes were available from the web site http://www.wormbase.org/#01-23-6. HT115 (DE3) bacteria containing the target genes and expressing double stranded RNAs (dsRNA) that inactivate specified genes were used as previously described 58,61 . Target gene expressions were knocked down by feeding the worms with dsRNA-containing E. coli HT115 strains from egg to adulthood at 16 °C. Except for skn-1, worms were exposed to RNAi bacteria from L3 stage larvae. The next generation of these worms were transferred to another RNAi plates and used for paralysis assay as above.
elegans strain CF1553 expressing GFP as a reporter for inducible sod-3 expression was used in our study. Age-synchronized L1 stage transgenic C. elegans were treated with or without 15 mg/mL DXN for 72 h, the expression of sod-3::gfp was evaluated through measuring the GFP fluorescence intensity of each whole worm. At least 25 randomly selected worms from each group were observed and photographed by fluorescence microscopy on a glass slide. To quantitatively measure the GFP fluorescence intensity in each whole worm, the Image J software (NIH, Bethesda, MD, USA) was used.
Quantification of Phsp-16.2::gfp expression via fluorescence microscopy. In the transgenic strain TJ375, GFP is fused to hsp-16.2 promoter for reporting the expression of HSP-16.2. Age-synchronized L1 stage Phsp-16.2::gfp worms were treated with or without 15 mg/mL DXN for 48 h, followed by exposure to 20 mM juglone for 24 h. The expression of hsp-16.2::gfp was evaluated through measuring the fluorescence intensity of the GFP reporter. At least 25 randomly selected worms from each group were measured by fluorescence microscopy on a glass slide. To quantitatively measure the GFP fluorescence intensity in the area anterior of the pharyngeal bulb in individual worms, Image J software (NIH, Bethesda, MD, USA) was used.
Morris water maze test. The Morris water maze test was performed as described by Morris (1981) 62 . The apparatus was a circular pool of 120 cm diameter filled with water, and maintained at 22.0 ± 0.5 °C by a calefaction stick. A transparent platform of 11 cm diameter 1 cm below the water surface was placed at a fixed point of one quadrant. 4-6 weeks old Kunming mice were provided by the Experiment Animal Center of Lanzhou University. Animals were treated with DXN at dose of 0.39 g/kg and 2.34 g/kg by i.g. for 21 days, mice in positive control group administrated piracetam at dose of 0.5 g/kg by i.g. Amnesia was induced by scopolamine at dose of 3 mg/ kg by i.p. 30 min after DXN or piracetam administration. Animals in control group were only received normal saline by i.g.. One day prior to the experiment, animals were trained to swim for 60 sec in the absence of the platform. Mice were performed single trial per day for consecutive 5 days, and escape latency of the time mice taken to swim to the platform was recorded by an automated tracking system (Ethovision XT software). 24 h after the experiment, a probe trial was performed to test for spatial memory.
All experiments were approved by the Animal Care and Use Committee of Lanzhou University, and the methods were carried out in accordance to the Guidance Suggestions for the Care and Use of Laboratory Animals, formulated by the Ministry of Science and Technology of China.
Statistical analysis. Statistical analyses were performed by SPSS 17.0 software. The statistical significances of the results were analyzed by using one-way analysis of variance (ANOVA). Except for paralysis assays, log rank survival test was carried out to compare the significance level among treatments. The P value of 0.05 or less was considered to be significant statistically.