Paeoniflorin improves cognitive dysfunction, restores glutamate receptors, attenuates gliosis and maintains synaptic plasticity in cadmium-intoxicated mice

2.1 Animal

ICR mice (8 weeks, male, 40 in total) were provided by Beijing Vital River Laboratory Animal Technology Co., ltd., under a license No. SYXK-2017–0005. Mice were randomly divided into 4 groups (n = 10), control, CdCl₂ (5 mg/kg), CdCl₂ (5 mg/kg) + PF (50 mg/kg), PF (50 mg/kg). The CdCl₂ was injected intraperitoneally every other day for 28 days. PF was administered by gavage once a day for 28 days, while saline was used in the control group. Animal experiments were carried out in accordance with the National Institutes of Health Laboratory Animal Care and Use Guidelines (NIH Publication No. 80–23) and approved by Animal Care and Use Committee of Minzu University of China.

2.2 Behavioral analysis

The Morris water maze (MWM) test was performed to determine the cognitive ability of mice and consisted of two phases of continuous experiments lasting 7 days (Zhong et al., 2019, Li et al., 2022). (1) Acquired training (first 6 days), to assess the learning ability of mice. (2) Probe trial (day 7 of the test) that used to assess the spatial memory ability of mice. In brief, animals were placed in a circular pool (110 cm in diameter and 50 cm in height) divided into four quadrants, with one quadrant placed with a platform (10 cm in diameter) submerged 1 cm of the water surface. A video-image system was placed above the center of the pool, which recorded the swimming path of the animals, which was then analyzed by software (Taimeng WMT-100). In acquired training (behavioral days 1–6), mice were allowed to find the platform freely within 1 min and the latency time was recorded. In probe trail (behavioral day 7), the platform was move away. The moved distance, residence time, and the crossing times of the animals in the target quadrant in the area where the original platform was placed, was recorded within 1 min.

The Y-maze spontaneous alternate experiment was used to test the animals' willingness to explore the new environments (Ren et al., 2021). The maze consisted of three equal length Y-shaped arms (30 cm long, 6 cm wide and 15 cm high) with an angle of 120° between them. Animals were allowed to explore freely from the top of one arm of the maze. Entry was considered when all four limbs of the animal were inside the arms. Correct spontaneous alternation occurred when the entering arm was different from the first two. The order in which the animal entered each arm over a 10-min period was recorded. Finally, the total alternation time and the number of spontaneous alternations were calculated.

A forced swimming test (FST) was performed to assess the depression-like phenotype of the animals (Yan et al., 2010). The experiment was performed in a transparent plastic cylindrical water tank (diameter 10 cm, height 38 cm). Water at 25 ± 1 °C was poured into the tank (height 25 cm), the mice were gently placed above the water surface and then released for swimming for a 6-min swim. The first minute was set as an adaption and was not recorded. The next 5 min were recorded by software Taimeng FST-100. The immobility time of mice within these 5 min was calculated.

2.3 H&E staining and immunohistochemistry (IHC)

Mice were anesthetized with isoflurane, dissected, thorax opened, and the left ventricle was perfused with saline until the liver tissue turned white and the heart stopped beating. Hematoxylin-eosin (H&E) staining was performed for histological observation. In brief, sections were dewaxed with xylene and ethanol, stained with H&E staining solution (#G1003, Servicebio) and observed under a microscope (NIKON DS-U3). Prefrontal cortex, hippocampus and other brain tissues were embedded in paraffin, sectioned, antigen repaired, and then IHC stained. Astrocytes and microglia were detected by GFAP (1:100, # 168251-AP) and IBA1 (1:100, #10904–1-AP) specific antibodies purchased from Proteintech Group or Bioss. DAB (antigen positive expression) and hematoxylin (nuclear staining) signals were collected under a microscope (NIKON DS-U3) and processed with Image Pro Plus 6.0 software (Media Cybernetics, Inc.).

2.4 Immunofluorescence

Brain sections were fixed in 4 % paraformaldehyde and dehydrated in 20 % and 30 % sucrose solutions. The sections were then washed 3 times with 1xPBS and sealed with 0.4 % Triton X-100 and 5 % goat serum (#SL038, Solarbio) for 2 h. After sealing and rinsed in 1xPBS for 3 times, sections were incubated with the primary antibody was incubated at 4^oC overnight. The primary antibody is shown in Supplementary Table S1. After washed away the free primary antibody in 1xPBS for 3 times, the secondary antibody was incubated. The nuclei were stained with Hoechst33342. Images were taken with a TCS SP8 confocal microscope (Leica Microsystems, Germany).

2.5 Western blotting

The hippocampus and prefrontal cortex were dissected and proteins were extracted. Equal amount of proteins were loaded and separated by 10 % SDS-PAGE and then transferred to nitrocellulose membrane, sealed with 5 % skim milk for 1 h, washed 3 times with TBST, followed by incubation with primary antibody at 4 °C overnight. The antibodies used are shown in Supplementary Table S1. After incubation with the secondary antibody, images were developed by the Odyssey CLX infrared fluorescence scanning imaging system. The relative protein levels were normalized to β-actin based on the optical density of the blotted bands and quantified by Image J software.

2.6 RT-PCR

Total RNA was purified from brain tissue using TRIzol reagent (#15596018, Ambion) as previously described. First strand cDNA was reverse transcribed from 2.0 μg RNA using EasyQuick RT MasterMix Kit (#CW 2019 M, Cowin Bio.) and then diluted in DEPC water at 1:10. Quantitative PCR were performed in a 10 μL reaction on a Light Cycler 96 Real-Time PCR System (Roche) using 2x RealStar Green Fast Mixture (#A301-05, GenStar) containing primers (Supplementary Table S2) for IL-1β, IL-6, TNF-α, IFN-γ, and IL-10, respectively. Three replicates for each template were performed and five samples were examined.

2.7 Statistical analysis

All data were expressed as mean ± s.e.m. One-way or two-way Analyses of Variance (ANOVA) were performed to examine the significance of differences between groups and are indicated separately in each figure. Otherwise, Mann Whitney U test was performed for nonparametric analysis. p < 0.05 was considered significant.

3.1 PF attenuates Cd-induced systemic toxicity in ICR mice

To determine the effectiveness of PF (Fig. 1A) in reducing the toxicity of Cd, ICR mice were exposed to intraperitoneally administered Cd for 28 days (Fig. 1B) or while being given PF treatment. During the experiment, mice were weighted on days 1, 14 and 28. As shown in Fig. 1C, mice in the control or PF group gained normal body weight over the same period, indicating that PF was not toxic to mice (Fig. 1C, PF group). However, mice treated with Cd lost significantly more weight compared to the control group, especially on day 28 (p < 0.01), while mice treated with Cd + PF loss less weight (p < 0.05). In addition, increased proteins were detected in the urine collected on day 28. The results showed that Cd significantly increased urinary protein in mice, indicating that Cd toxicity disrupts the integrity of the kidney (Fig. 1D). Consistently, kidney atrophy was observed at the end of the experiment after dissection of the organs, as the kidney index (organ weight/body mass) decreased (Fig. 1E, kidney, p < 0.001). In addition, hepatomegaly and encephalomegaly was found as the liver and brain indices increased (Fig. 1E, liver and brain, p < 0.001). Again, PF treatment protected organs from Cd-induced toxicity and recovered them to normal. Furthermore, we performed H&E staining on sections of liver, kidney, small intestine and testis and found that the toxicity of Cd induced severe damage but was alleviated by PF treatment (Fig. 1F). Normally, hepatocytes were arranged radially around the central vein without dilatation and congestion, inflammatory cell infiltration and other pathological changes. However, in Cd-intoxicated mice, the hepatocytes were irregularly arranged with blurred hepatic cord boundaries and inflammatory cell infiltration around the central vein (Fig. 1F, liver in Cd group). In contrast, in Cd + PF treated mice, hepatocytes were neatly aligned and inflammation was alleviated, indicating PF-mediated protection of the liver from Cd-induced toxicity. Similarly, renal cells were tightly arranged with well-defined tubular lumen and glomerular structure. After Cd toxicity, renal cells were loosely arranged with marked swelling, narrowed tubular lumen and irregular glomeruli (Fig. 1F, kidney with Cd). Again, the above abnormalities were ameliorated by PF. In addition, Cd-induced toxicity to the small intestine and testes was attenuated by PF treatment. Notably, Cd toxicity was clearly observed as manifested by atrophy of spermatogonial epithelial cells and a dramatic reduction in spermatocyte numbers (Fig. 1F, testis with Cd). In other words, Cd was able to effectively penetrate the testicular barrier in the blood and induce toxicity. Fortunately, PF is also effective in mitigating the toxicity of Cd in the testis (Cd + PF). Thus, PF is a potential natural product to attenuate the toxicity of Cd in ICR mice.

Close

Fig. 1

Cd induced systemic injury in ICR mice but can be attenuated by PF treatment. (A) Chemical structure of Paeoniflorin (PF). (B) Timing of animal experiments. i.p. qod., intraperitoneally injection, every other day. i.g. qd., injection by gavage, once daily. (C) Body weight of mice. PF prevented the weight loss of Cd-treated mice. (D) PF reduced the abnormal increase in urinary protein caused by Cd toxicity. (E) Organ indices of liver, kidney and brain. Cd-induced toxicity caused hepatic edema, cerebral edema, and renal atrophy according to organ/body mass index. (F) H & E staining was used to observe the toxicity of Cd in different organs and the protection PF against Cd-induced toxicity. Arrows indicate morphological changes in Cd-treated tissues. In this figure, C, two-way ANOVA, post hoc Tukey’s test was performed. In D, a nonparametric analysis was performed by the Mann-Whitney U test. In E, a one-way ANOVA, post hoc Dunnett’s test was performed. n.s, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Export to PPT

3.2 PF alleviates the AD-like and depression-like phenotypes induced by Cd toxicity in ICR mice

Behavioral tests were performed on ICR mice after modeling and drug treatment. A Y-maze test was performed to determine the willingness of mice to explore new environments. We found that Cd-intoxication reduced the willingness of mice to explore new environments, but could be restored by PF treatment (Fig. 2A). In addition, Cd-intoxication induced a depression-like phenotype in mice, as determined by the forced swimming test (FST) (Fig. 2B). In the test, Cd-induced immobility was significantly prolonged in the FST, while PF treatment reduced the duration of immobility, indicating an improvement in Cd toxicity (p < 0.001).

Close

Fig. 2

Cd-intoxication induced AD-like and depressive-like behaviors in ICR mice, but can be restored by PF treatment. (A) Spontaneous alterations and number of entries in the Y-maze test in mice. (B) Mice in the Cd group exhibited longer immobility time in the FS test, indicating a depressive-like phenotype; which was ameliorated by PF treatment. (C) Escape latency (with platform) in ICR mice in the acquired training test. (D-E) Representative swim paths and thermal-infrared trajectories for MWM tests obtained on behavioral days 6 and day 7, respectively. The location of the underwater platform is marked with a red circle to represent the target area. (F) Probe trail of MWM tests (day 7) showing distance traveled, residence time and number of crossings in the target quadrant where the platform was originally placed. In this figure, in A-B and F, a one-way ANOVA, post hoc Dunnett’s test was performed. In C left panel, two-way RM ANOVA, post hoc Tukey’s test was performed. n.s, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001 (Cd vs Control); ##, p < 0.01, ###, p < 0.001 (Cd + PF vs Cd).

Export to PPT

In MWM experiments, we carried out acquired training (last 6 days) and probe trail successively. During acquired training, mice were allowed to find the platform submerged in water. During training, the mice acquired a memory for the platform position, so the time to find the platform (escape latency) gradually decreased with training (Fig. 2C, control). However, Cd-intoxication impaired the mice’s memory, so it took longer to find the platform compared to the control (Fig. 2C, Cd). In contrast, PF treatment maintained the learning and memory abilities of the mice, similar to the controls, indicating the attenuation of Cd-toxicity (Fig. 2C, Cd + PF). Recording of swimming occupancy plot also showed that the route to find the platform was much more complicated in Cd-intoxicated mice compared to the other groups (Fig. 2D). Moreover, mice did not differ significantly in the MWM test, indicating that Cd or PF did not affect the physiological movements of mice (Fig. 2C, right column). Consistently, after completion of acquired training, probe trail was performed in which the platform was removed and mice were allowed to swim in the maze for 1 min and the swimming path was recorded (Fig. 2E). We found that in the control, PF and Cd + PF groups, mice spent most of their time in the quadrant where the platform was initially placed, with the exception of the Cd-treated group. Statistical analysis of swimming paths also confirmed that Cd-treated mice spent less time in the target quadrant area, traversed it less frequently and moved a greater distance than the control and even the Cd + PF group (Fig. 2F). In conclusion, Cd intoxication caused cognitive dysfunction in mice, which could be restored by PF treatment (Fig. 2).

3.3 PF reduces Cd toxicity in mice by recovering the activity of glutamate receptors

The glutamate ionotropic receptor NMDA regulates learning and long term potentiation in the brain. To study whether Cd toxicity affects NMDA receptor activity and to examine the protective role of PF, the activity of NMDA receptor 2A and 2B was detected by immunofluorescence and Western blotting using antibodies specific for their phosphorylation (Fig. 3, Fig. S1). As shown in Fig. 3A, p-NMDA receptors (2B) were clearly expressed on the surface of cell membrane and in the cytoplasm of hippocampal and prefrontal cortex neurons. However, p-NMDAR2B was reduced after Cd treatment compared with controls (Fig. 3A, p-NMDAR). After PF treatment, the levels of p-NMDAR2B were recovered in the Cd + PF group. Western blotting also confirmed the restoration of protein levels of these receptors (Fig. 3B and C).

Close

Fig. 3

PF recovered the activity of glutamate receptors in response to Cd-induced toxicity. (A) Representative immunofluorescence images of phospho-NMDAR2B and phospho-GluR2 in brain sections of ICR mouse. Images of hippocampal area DG and prefrontal cortex (PFC) are shown in Supplementary Materials Fig. S1. (B) Protein levels in the hippocampus were detected by Western blot analysis. (C) Quantification of relative protein levels as indicated by blot intensity in (B) by Image J software. One-way ANOVA, post hoc Tukey’s test was carried out. n.s, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Export to PPT

Other important types of GluRs are AMPA receptors, of which GluR1 and GluR2 are the two main subtypes, both of which are regulated by protein phosphorylation (Steinberg et al., 2006, Lee et al., 2010). In the experiment, phospho-GluR2 (p-GluR2) decreased after Cd treatment compared to the control (Fig. 3A, p-GluR2), and also confirmed by Western blotting in the Cd-treated samples (Fig. 3B and C). In contrast, p-GluR1 was up-regulated after Cd treatment. These results imply that Cd-intoxication induced the destruction of GluRs in the CNS. However, PF treatment restored the Cd-induced changes in GluR1 and GluR2, attenuating Cd toxicity (Fig. 3, Cd + PF).

3.4 PF attenuates Cd-induced gliosis in ICR mice

The status of microglia and astrocytes was detected, as they are the major regulators of neuroinflammation in the CNS. Immunofluorescence (IF) staining of brain slices from the hippocampus and prefrontal cortex were performed using antibodies specific to GFAP and IBA1 (Fig. 4). The signal and number of microglia (IBA1) and astrocytes (GFAP) in hippocampal DG, CA1 and cortex (PFC) were significantly increased after Cd treatment (Fig. 4A and B, Fig. S2). Under Cd-induced toxicity, gliosis was activated, as evidenced by the enlargement of cell bodies and lengthening of protrusions. In contrast, the number of microglia and astrocytes decreased with PF treatment. Consistently, protein levels examined by Western blotting confirmed opposite changes of GFAP and IBA1 in response to Cd and PF as revealed by IF (Fig. 4C). Furthermore, we performed IHC staining of brain slices. We found that Cd induced toxicity increased the number of GFAP and IBA1-positive cells in the DG, CA1 and CA3 regions of the hippocampus and in the prefrontal cortex, whereas PF effectively reduced the increase in these cells (Fig. 5, Fig. S3). Consistently, the results of RT-PCR to determine the expression of inflammatory factors such as IL-1β, IL-6, TNF-α, as well as IFN-γ and anti-inflammatory factor IL-10 also confirmed the anti-inflammation effect of PF (Fig. 5D). Thus, PF effectively alleviated Cd-induced gliosis.

Close

Fig. 4

Representative immunofluorescence images and blots of GFAP and IBA1. (A) Cd-induced toxicity in the brain was manifested by an increase in GFAP- and IBA1-positive cells, indicating activation of microglia and astrocytes. However, PF treatment attenuated the Cd-induced neurotoxicity and restored cellular homeostasis. Regional images of hippocampal DG and prefrontal cortex (PFC) are shown in Supplementary Material Fig. S2. (B) Cell counts were calculated from the images. (C) Western blots analysis of GFAP and IBA1 protein levels. In B-C, one-way ANOVA, post hoc Dunnett’s test were performed. n.s, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Export to PPT

Close

Fig. 5

Immunohistochemical staining of GFAP and IBA1 in the prefrontal cortex (PFC) and hippocampus. (A-B) Representative IHC images show that Cd-induced toxicity activated the inflammatory response in astrocytes and microglia, as indicated by the increased levels of GFAP and IBA1. However, PF restored Cd-induced inflammation. Images of GFAP and IBA1 in hippocampal CA1 are showed in Supplementary Materials Fig. S3. (C) Statistical analysis of the number of GFAP- and IBA1-positive cells in the hippocampus and prefrontal cortex was performed. (D) The mRNA expression levels of inflammatory factors in brain tissue were determined by RT-PCR. In C and D, a one-way ANOVA, post hoc Dunnett’s test was performed. n.s, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Export to PPT

3.5 PF alleviates Cd-induced gliosis by reducing HIF-1α levels and protecting neurons

HIF-1α is involved in cellular adaptation to oxidative stress and regulates inflammation (Ashok et al., 2017). Therefore, we performed co-immunostaining of HIF-1α with GFAP, as shown in Fig. 6. HIF-1α was mainly expressed in the cytoplasm. After Cd intoxication, the level of HIF-1α was significantly increased, along with a plenty of activated astrocytes, as shown by GFAP-labeling (Fig. 6A, Fig. S4). Compared to controls, astrocytes were activated and their number increased after Cd-treatment, whereas PF attenuated the toxicity of Cd and reduced the increment of HIF-1α and the activation of astrocytes. We also confirmed that neuronal responses to Cd and PF treatment were opposite to the changes in astrocytes (Fig. 6B, Fig. S4). In detail, Cd toxicity disrupted the activity of neurons and even caused substantial loss of neurons, especially in hippocampus (Fig. 6B, Fig. S4), by staining neurons with specific antibodies to the protein marker MAP2, an important cytoskeletal protein that is closely related to neuroplasticity and plays an important role in dendritic and branchial extension (Lingwood et al., 2008). However, PF treatment prevented Cd-induced loss of neurons and maintained CNS homeostasis. Western blotting of protein levels of HIF-1α and MAP2 also showed results consistent with immunofluorescence staining (Fig. 6C).

Close

Fig. 6

PF inhibited Cd-induced inflammation and protected neurons by attenuating the increase of HIF-1α. (A) Immunofluorescence staining of GFAP and HIF-1α in the hippocampus. Cd induced an increase in HIF-1α levels and the number of GFAP-labeled astrocytes. PF mitigated the increase in HIF-1α expression and astrocyte activation, thereby reducing inflammation. (B) Immunofluorescence staining for MAP2 in PFC and hippocampus. Images of hippocampal CA4 (A) and CA1 (B) are showed in Supplementary Materials Fig. S4. (C) Western blotting for protein levels of GFAP and HIF-1α as well as the neuronal synaptic proteins synaptophysin and PSD95 in the hippocampus. The intensity of the blotted bands was quantified using Image J software. One-way ANOVA, post hoc Dunnett’s test was performed. n.s, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Export to PPT

In addition, the reduction of postsynaptic densifying membrane 95 (PSD95) and synaptic vesicle protein (synaptophysin) in brain tissues of AD patients is closely correlated with the degree of cognitive decline (Qin et al., 2004, Tajiri et al., 2013). We found that PSD95 and synaptophysin were significantly reduced following Cd treatment (Fig. 6C), implying impairment of neuronal synaptophysin by Cd toxicity. However, consistent with the maintenance of neurons by PF treatment, protein levels of PSD95 and synaptophysin were restored. In other words, PF attenuated Cd-induced neuronal damage and protected neurons, especially by restoring synaptic plasticity.

At the mechanistic level, we also explored molecular changes within microglia. CD68 (also named macrogliadin in mice) is a lysosomal membrane protein that is highly expressed by macrophages and activated microglia, but less so by resting microglia, as it is essential for lysosomal activity of microglia to clear damaged cellular components or molecules. It is therefore positively correlated with pathological and cognitive function in AD (Minett et al., 2016). Using immunofluorescence, we found co-localization of microglia (stained by IBA1) with CD68 (Fig. 7), especially in Cd-intoxicated brain samples, indicating microglia activation and upregulation of CD68. It can be seen that this is a stress response of the mouse brain to Cd toxicity, and is very similar to the pathological features of AD. This result is also consistent with the behavioral manifestations of the mice described previously. That is, the cognitive impairment induced by Cd toxicity corresponds to a series of changes in glial cells and neurons. However, PF treatment again effectively attenuated the toxicity of Cd and alleviated the increase in CD68, indicating a reduction in lysosomal activity in microglia. In other words, PF protected neurons from Cd-induced toxicity and maintained CNS homeostasis (Fig. 8).

Close

Fig. 7

Immunofluorescence staining of IBA1, CD68 in prefrontal cortex (PFC) and hippocampus. (A) Cd toxicity activated microglia, as indicated by increased levels of CD68 and number of IBA1-positive cell. PF attenuated Cd toxicity, microglia activation and the increase of CD68.

Export to PPT

Close

Fig. 8

Schematic graph showing PF-mediated attenuation of Cd toxicity and restoration of homeostasis in the brain.

Export to PPT