Impact of green-synthesized silver nanoparticles on cognitive function and fluorescence spectroscopy of redox status in the hippocampus

INTRODUCTION
Frequent use of silver nanoparticles (AgNPs) in medical products generates potential risks for human health [1]. The orally administered nanoparticles cross the blood–brain barrier [2, 3] and little information is available regarding their interactions with the central nervous system macromolecules [4]. The nanoparticles show a tendency to accumulate in the brain [5]. Therefore, it is essential to investigate their neurotoxicity. 
While several studies have demonstrated AgNPs toxicity on neuronal cells [6, 7], their anti-inflammatory and protective effects have also been reported [8, 9]. The noxiousness of AgNPs has been reported, but their hormesis effects have been overlooked [10]. Hormesis is a dose-dependent effect, as a low dose stimulates and a high dose inhibits certain targets [11].
Moreover, the cytotoxicity of AgNPs begins with the silver ions release [12] and induction of oxidative stress [13]. The green synthesis reduces the toxicity [9, 14] by stabilizing AgNPs [15, 16]. Myrtle (Myrtus communis L., Myrtaceae), a traditional medicine plant [17] could be used for AgNPs green synthesis [18, 19]. 
This study aimed to determine the possible cognitive alterations after intra-cerebroventricular (ICV) administration of different doses of green-synthesized silver nanoparticles (AgNPs). The working memory, passive avoidance test, novel object recognition, and the anxiety behavior were evaluated. Hippocampal redox status was evaluated via malondialdehyde (MDA) and superoxide dismutase (SOD) activity measurements. The reaction of aldehydes such as MDA with proteins or phospholipids produces lipophilic fluorescent products [20]. We have evaluated the hippocampus redox status by fluorescence spectroscopy of lipophilic fluorescent products (LFPs). The inherent fluorescence property is specific feature of LFPs which may change in different microenvironments [21]. Autofluorescence bioproducts are more sensitive than MDA[22] and a more reliable marker of tissue redox status.

MATERIAL AND METHODS 
Chemicals
The AgNO3 (CAS Number: 7761-88-8) was purchased from Sigma-Aldrich. For animal anesthesia, the Ketamine hydrochloride and Xylazine were purchased from Alfasan (Woerden, Holland). The hippocampal total protein, MDA, and SOD were measured by kit (TebPazhouhan Razi, Iran). 

Synthesis of AgNPs with the M. communis plant 
The plant was collected from the grid reference of 34.9805° N, 46.2720° E, from Bayangan Mountains in the west of Iran. After washing with the distilled water, 10 g of amassed leaves were air-dried to prepare the aqueous leaf extract. The dried leaves powder was boiled by 150 mL of deionized water at 60 ̊C. After filtration, the suspension was stored at 4˚C till biochemical analysis [23].
The AgNPs was synthesized by adding 50 mL of AgNO3 (0.001 M) to the 50 mL of leaf aqueous extract at 25 ̊C under stirred conditions for 48 h. Then, the solution was centrifuged (4000 rpm, one hour). The reduction of AgNO3 with M. communis leaf extract was confirmed by dark brown color alteration of the solution[24]. After determining the average AgNPs crystal size, the FE-SEM (Model XL30, Philips, Eindhoven) was applied to assess the morphology, and the elemental analysis of AgNPs and the FT-IR analysis was performed via spectrophotometer (Germany, Bruker, Model: ALPHA). The peak absorbance of AgNPs was obtained in the range of 300 to 600 nm via an UV-visible spectrophotometer (Tomas, UV 331). The TEM technique revealed the average diameter size of 91.68 nm. 

In-vivo method to study AgNPs effects
Animals 
Adult male rats from Wistar race (240 – 270 g, 6-7 weeks of age) were housed in 12h light/dark cycle in a standard room (21 ± 2◦C). Food and water consumption was ad libitum. The total number of animals was seventy, and ten rats in each group were randomly selected according to the previous published data [25]. Tehran University of Medical Sciences Ethics Committee (IR.TUMS.VCR.REC.1396.4033) approved completely experimental protocols.

Stereotaxic surgery
Anesthesia agents were Ketamine hydrochloride (65 mg/kg, i.p.) and xylazine (15 mg/kg, i.p.). The lateral ventricles were cannulated by a guide cannula (22G) during the stereotaxic surgery (-0.8 mm, 1.5 mm laterality, and 4 mm beneath the brain surface)[26]. A heating pad was used during surgery. The 0.5 mg/kg ketoprofen was injected after surgery to reduce pain and distress. The recovery periods were seven days. Then, an intra-ICV injections were performed by a Hamilton syringe. The ICV-route ensures that a similar dose of AgNPs reaches to the brain of all animals irrespective of variations in drug transport or elimination routes. 

Treatment Procedures
The AgNPs (0.1, 1, 10, and 100 ppm) were prepared in the normal saline. The solution was dispersed with an ultrasonic shaker prior each administration. The volume of intra-ICV injections was similar in all animals. Animals were dedicated to 7 groups (10 rats in each group) accidentally: 1) Intact group: remained intact during the experiments; 2) Sham group received normal saline (0.9 %, intra-ICV); 3) Extract group received an intra-ICV injection of leaf extract; 4) In the AgNPs groups, different doses of green-synthesized AgNPs (0.1, 1, 10 and 100 ppm, intra-ICV) were injected. 
The open field 
Motor activity of all rats was evaluated one week after surgery by the open field test. Three parameters of the number of line cross, frequency of rearing and grooming were recorded as described previously[27].

Y-maze 
Alternation behavior was assessed 13 days after saline, extract or AgNPs administration for 8 minutes duration. A digital camera was used for recording the sequence of entries into the arms. The olfactory cues in the Y maze apparatus removed by alcohol swab before each test.

Novel object recognition (NOR) 
The test based on the animal’s desire to explore the new objects. The rats were allowed to navigate in the chamber for 10 minutes while two similar objects at a distance from each other placed in the chamber. After an hour, the new object was substitute with one of the old objects. The exploration time in the chamber was three minutes. The exploring time for new and old objects were recorded by a camera. The box and objects were cleaned in order to abolish the olfactory cues. Recognition index scores were calculated as the percentage of exploration time of novel object divided by total exploration time of new plus familiar object [28]. 

Elevated plus maze (EPM)
The EPM device had two closed and two open arms (50×10 cm) located in the opposite direction. The height was 50 cm above the floor and rats allowed to search freely for 5 min. The recorded films were used for calculation of the open arms time (OAT %), and the entrance to open arms (OAE %) in percentage as explained before[29].

Passive avoidance memory
The shuttle box apparatus was used for the measurement of passive avoidance memory. As mentioned before [29], the delay time of entry into the dark side of the box was recorded as step-through latency (STL). The time that animals spent in the dark chamber was reported as TDC.

Hippocampal MDA and SOD measurements
At the end of experiments, after animal decapitation, the hippocampus tissues were extracted and homogenized. The supernatants of homogenates kept at -70°C. The total protein concentrations were evaluated via the Bradford technique [30]. Lipid peroxidation level and SOD activity were determined based on the kit instructions. 

Lipophilic fluorescent products (LFPs) assay
The extraction of hippocampus LFPs were based on the Goldstein and McDonagh description [31], and Wilhelm and Herget’s modification [31]. The cold PBS (1:10 w/v) was used for hippocampus homogenization. The one ml of hippocampus homogenates was mixed with four ml of chloroform and methanol (2:1, v/v). After shaking and adding the distilled water and centrifuge (200×g, 20 min), the lower phase was used for fluorescent assay [32, 33]. To identify fluorescence maxima, all samples were analyzed and the fluorescence was measured in two synchronous fluorescence spectra, 25, and 50 nm difference between excitation and emission according to the previous our study [25].

Statistical analysis
The statistical analysis software was GraphPad Prism (version 7.04). After the normality evaluation by the Shapiro-Wilk test, the locomotion data, behavioral tasks, MDA level, and SOD activity were compared using one-way analysis of variance (ANOVA). The Tukey’s test was considered as the post hoc. The two-way analysis of variance and repeated measure were performed for the fluorescence intensity comparisons. Findings were reported as the mean ± SEM. The exact p-value was provided for all variables.

RESULTS
3.1. Physicochemical properties of green-synthetized AgNPs
Spectroscopy of UV-visible: The AgNPs excitation absorption was around 320-370 nm and absorption peak intensity was 1.53 a.u (Fig.1A).
TEM and Zeta potential: The average diameter size of 91.68 nm was detected by TEM images (Fig.1B). There were different shapes involving spherical, triangular, and hexagonal. In addition, zeta potential was -17.4 mV (Fig.1C). The zeta potential was comparable to other studies about the AgNPs green synthesis [33, 34]. The Coulombic repulsion forces due to the negative surface potentials of AgNPs prevent aggregation. 
X-ray diffraction analysis (XRD) and Fourier-transform infrared spectroscopy (FTIR): The XRD technique was used to assess crystal structure and AgNPs phase purity. The AgNPs diffraction peaks has been shown in Fig.1D. The AgNPs demonstrated 2 theta degrees of 38.22°, 44.27°, 63.96°, and 77.21° (111), (200), (220), and (311) indices respectively (JCPDS card no: 65-2871). In addition, crystal sizes 72 nm. The FTIR analysis showed several functional groups, including amine, carboxylic acid, ketone, aldehyde, phenol, and sulfoxide. There were prominent peaks, including 3397.85, 1723.53, 1612.53, 1386.27, 1070.29 cm-1 for N-H, C=O, and C=O stretching, as well as C-H bending, and S=O.

AgNPs Effects on behavioral indices 
The locomotion (F(6,63)=1.69, p=0.13), rearing numbers (F(6, 63)=1.30, p=0.26), and grooming behavior (F(6, 63)=1.01, p=0.42) were alike in different groups (Table 1). The weight of animals was recorded two times, at the beginning and research end. The body weight gain among groups before (F(6, 63)=0.85, p=0.53) and after (F(6, 63)=1.77, p=0.11, Table 1) experiments were similar. 

Working memory in the Y-maze 
We observed no significant difference regarding the alternative behavior among intact, sham and extract groups (F(6, 63)=6.44, p=0.98, Fig.2A). The same statistical results were found for 0.1 ppm (64.8 ± 3.6, vs. 67.3±2.87, p=0.99), and one ppm (60 ± 2.53, p=0.34) of AgNPs compared with the sham group. The percent of alternative behavior significantly decreased in groups that received 10 (55.6 ± 1.36, p=0.017) and 100 (53.2 ± 2.13, p=0.002) ppm AgNPs. The number of total arm entries were similar among groups (F(6,63)=0.64, p=0.69, Fig.2B). 

Recognition memory in the NOR test
Recognition index was significantly altered (F(6, 63)=31.42, p=0.000, Fig.2C). We found no significant difference among intact (79.08±3.7), sham (68.4±2.007), and extract (74 ±3.46) groups. The same condition was detected between 0.1 ppm (69.6 ± 3.73, p=0.99) and the sham group. However, the one (54.2 ±2.21, p=0.014), 10 (47.3±1.71, p<0.0001) and 100 (34.7 ±2.45, p<0.0001) ppm of AgNPs caused significant reduction in recognition index when matched with the sham group data.
    
Anxiety–like behavior in the EPM test
The OAT% (F (6, 63)=15.0, p=0.000) and OAE% (F (6, 63)=6.14, p=0.000) were significantly changed. The alterations of OAT% (22.3 ±0.97 vs. 18.1 ± 0.7, p=0.16, Fig.3A) and OAE% (43.4±1.88 vs. 43.8 ± 4.57, p=0.99, Fig.3B) were similar in the intact and sham groups. The same result was detected between the intact and extract groups in OAT% (18.7 ±1.02, p=0.32) and OAE% (35.0 ± 1.39, p=0.26). AgNPs administration, 0.1 (18.1±1.65, p=0.99), and 1ppm (15.6±0.98, p=0.75) had no significant effects on OAT% versus the sham group. However, the 0.1 (32.6±2.47, p=0.048) and 1ppm (31.6±1.37, p=0.02) caused a significant reduction in OAE%. The 10 and 100 ppm significantly decreased OAT% (12±1.41, p=0.008; 8.8±1.18, p=0.001) and OAE% (31.6±1.89, p=0.02; 26.5±3.11, p=0.0001) as compared to the sham group. The changes in the number of total arm entries among groups (F (6,63)=1.11, p=0.36, Fig.3C) were not significant. These results imply that, the doses of 10 and 100 ppm were anxiogenic. 
 
Passive avoidance memory test
We observed the significant changes in the STL (F(6, 63)=90.0, p=0.000), and TDC (F(6, 63)=73.24, p=0.000). The differences between the intact and sham groups in the STL (151±7.84 vs.139±8.75, p=0.79, Fig.3D) and TDC (16.2±1.1 vs.16.5±1.2, p=0.99, Fig.3E) were nonsignificant. The STL (135±5.52, p=0.47) and TDC (23.1±4.19, p=0.98) were similar in the intact and extract groups. The AgNPs (0.1 ppm) had no significant effects on STL (137±7.81, p=0.99) and TDC (16.1±1.17, p=0.99) versus the sham group. While, the one (56.4±2.38, p<0.0001), 10 (22.3±2.82), and 100 (23.6±4.48, p<0.0001) ppm caused a significant reduction in the STL and TDC (69.1±13.9; 129.4±7.19; 137.7±6.49) versus the sham group.

Hippocampal MDA levels and SOD activity
We observed a significant alteration in MDA level among groups (F(6,63)=147.2, p=0.000, µmol/mg protein, Fig.4A). The intact and sham groups displayed no significant difference (3.01±0.31 vs. 3.22±0.52, p=0.99). The extract showed no significant changes versus the sham groups (4.86±0.71, p=0.71). We detected no significant difference between the AgNPs (0.1 ppm) and the sham group (3.38±0.59, p=0.99). The AgNPs of 1(8.97±0.23), 10 (13.4±0.89) and 100 (23.3±0.74) ppm caused a significant rise in MDA level when checked with the sham group. 
Our finding revealed the significant change in SOD activity (F(6,63)=65.27, P=0.000, U/mg protein, Fig.4B). The result of comparison between the intact and sham groups were 152±1.34 vs. 152±0.97, p=0.99. The same results were observed about the SOD activity in the extract (149±0.99, p=0.33) and 0.1 ppm AgNPs group (148±1.69, p=0.56) compared with the sham group. The SOD activity was decreased in one (132±2.48), 10 (132±2.3) and 100 (118±1.28) ppm AgNPs groups. Finally, the 0.1 ppm caused no oxidative induction. 

AgNPs Effects on fluorescence spectra of LFPs in the hippocampus 
A significant difference was detected between the fluorescence intensity (F(20, 120)=537.9, p< 0.001), and different groups (F(6, 36) = 44.62, p<0.001) and a significant interaction (F(120, 720) = 4.97, p< 0.001) was detected between them. In synchronous spectra analysis of the hippocampus, we found no significant difference between the intact, sham and extract groups in both synchrony of 25 and 50 nm. In SYN25, eight significant specific fluorescence maxima 270/295, 290/315, 320/ 345, 330/355, 360/385, 370/395, 410/435, and 420/445nm were detected in the 1, 10, and 100ppm AgNPs groups as matched with the sham data (Fig 5A). In SYN50, the most pronounced difference was observed at 270/320, 290/340, 300/350, 310/360 and 320/370 nm (Fig 5B).

DISCUSSION
Our finding revealed the ICV-administration of green-synthesized AgNPs (10, and 100 ppm) caused memory impairment and induced anxiety-like behavior. In addition, seven significant fluorescence maxima were distinguished in the AgNPs groups (1, 10, and 100ppm). However, the 0.1 ppm of AgNPs had no effects on MDA, SOD activity and LFPs. Exposure to the low doses of AgNPs, possibly through the hormesis effect, resisted the neurotoxicity. We focused on the dorsal hippocampus for the reason that, it is crucial for spatial learning[35] , retention of object recognition [36], and rat’s passive avoidance memory [37].
The antioxidant and anti-inflammatory effects of AgNPs have been reported [8] and their usage has been dramatically enhanced. The green synthesis mediates biocompatible AgNPs for in vivo administration [38]. The plant extracts contribute to the stabilization of silver ions [39] and occupy the surface of nanoparticles [40], augmenting the potentiality of the nanoparticles [16]. Toxicity of chemical synthesized AgNPs to neurons, astrocytes and glial cells has been reported [41-43]. The neurotoxicity of various sized (25, 40, and 80 nm) silver nanoparticles have also been studied [44]. The larger AgNPs in comparison with the smaller particles, has less effects on rat brain microvessels to produce an inflammatory response. The ROS production and NMDA receptor activation have been reported as important mechanisms of AgNPs neurotoxicity in cultured neurons [45]. 
Nevertheless, the size, dose, and exposure time of AgNPs are important issues in biological responses. In the current study, the effects of AgNPs were dose-dependent. We found that 0.1 ppm caused no oxidative damage and cognitive impairment which may be due to the hormesis effect. We observed that high doses of AgNPs (10, and 100 ppm) induced working memory damage. The NOR task that depends on the integrity of the temporal regions [46], and is helpful for the cognitive deficit detection [47], was also impaired. In addition, the EPM results showed that high doses of AgNPs were anxiogenic. 
Our data revealed the passive avoidance memory impairment in animals treated by high doses of AgNPs as indicated by a decrease in the STL and an increase in TDC. The high doses of AgNPs caused lipid peroxidation demonstrated by a rise in the hippocampal MDA level and was accompanied by a decrease in the activity of SOD. The nanoparticles neurotoxicity may be related to oxidative stress [48]. After AgNPs internalization, the silver ions release, and their reaction via several targets initiate the cytotoxicity[12] and induce oxidative stress.
We can interpret that cognitive impairment in the high doses of AgNPs groups was associated with an oxidative imbalance in the hippocampus tissue. High MDA level is a non-specific or common component in various diseases. We further evaluated the alteration of redox homeostasis by the measurement of fluorescence intensity of LFPs. The fluorescence intensity peaks were detected in specific excitation/emissions. The pattern of alterations was not the same in different doses implying the involvement of different LFPs composition. Monitoring the patterns of changes in the fluorescent spectra of LFPs may help to find the most appropriate biomarker to evaluate nanoparticle toxicity. 
In addition, the AgNPs antioxidant effects have been reported in other studies. For instance, Gonzalez-Carter, et al. showed their antioxidant effects against lipopolysaccharide-reactive oxygen species production[8]. Metallothioneins are antioxidants and interact with amyloidogenic peptides, and their up-regulation has been proposed as a therapeutic approach [49]. Interestingly, astrocytes exposure to AgNPs caused metallothioneins up-regulation [50]. Therefore, AgNPs may prevent oxidative damage by a compensatory mechanism such as metallothioneins’ contribution. More studies are required to confirm these aspects.  

CONCLUSION
The ICV-AgNPs (10, and 100 ppm) administration caused memory impairment and altered the hippocampal redox status, verified by a rise in MDA level and decreased SOD activity. We further confirmed the alteration of redox status by detecting an elevated fluorescence intensity of LFPs in hippocampus. However, the low dose of AgNPs (0.1 ppm), possibly through the hormesis effect, had no destructive effects on behavioral indices and redox status. Understanding the biological effects of AgNPs can be used to make future safe classes of AgNPs in the food and drug industries. 

ACKNOWLEDGMENTS
The Electrophysiology Research Center, Neuroscience Institute, Tehran University of Medical Sciences (Grant numbers: 96-03-139-36130) have supported this work, and we would like to thank them for their financial support of this research project. 

DECLARATION OF COMPETING INTEREST
The authors declare that they have no conflict of interest.

ABBREVIATIONS 
FTIR: Spectroscopy of Fourier-transform infrared, Green-synthesized, silver nanoparticles: AgNPs, ICV: Intracerebroventricular, LFPs: Lipophilic fluorescent products, MDA: malondialdehyde, SOD: superoxide dismutase, TEM: transmission electron microscopy