Fo-Shou-San Ameliorates Chronic Cerebral Hypoperfusion-Induced Cognitive Impairment in Mice by Regulating NRF2/HO-1 Pathway Against Ferroptosis

Background : Fo-Shou-San (FSS) is a traditional Chinese medicine (TCM) decoction that can effectively treat vascular dementia (VD). In the face of unclear pharmacological mechanisms, we set out to validate that FSS treats chronic cerebral hypoperfusion (CCH)-induced cognitive impairment in mice. Methods : CCH animal model caused by permanent right unilateral common carotid arteries occlusion (rUCCAO) was established to verify that FSS could treat subcortical ischemic vascular dementia (SIVD). We performed novel object recognition test and Morris water maze test, observed morphological changes via HE and Nissl staining, and detected hippocampus apoptosis by TUNEL staining and oxidative stress by biochemical assays. Ferroptosis-related markers and NRF2/HO-1 signaling-related expressions were examined via qPCR and immunofluorescence staining. Results : We found that FSS ameliorated cognitive disorders, and lessened oxidative stress by decreasing MDA and GSH-PX while increasing the reduced glutathione (GSH)/oxidized glutathione disulfide (GSSG) ratio, which are associated with ferroptosis. Additionally, FSS reduced expression of SLC7A11 , GPX4 , ROX and 4HNE , as vital markers of ferroptosis. Further, FSS regulated NRF2/HO-1 signaling by downregulating NRF2 and HO-1 . Conclusions : Our study suggests that FSS may ameliorate chronic cerebral hypoperfusion-induced cognitive deficits through regulation of the NRF2/HO-1 pathway against ferroptosis. Taken together, our study highlights the neuroprotective efficacy of FSS.


Introduction
Vascular dementia (VD), the second most prevalent form of dementia, leads to cognitive impairment among elderly people [1]. Clinically, chronic cerebral hypoperfusion frequently causes subcortical ischemic vascular dementia (SIVD), a subtype of vascular dementia [2]. Pathological changes of SIVD include white matter injury, microglial activation, oligodendrocyte apoptosis and progressive demyelination [3][4][5]. The permanent right unilateral common carotid arteries occlusion (rUCCAO) mice model was performed as a chronic cerebral hypoperfusion (CCH) animal model mimicking pathophysiological characteristics of SIVD patients [6]. At present, there are no effective treatments for SIVD [7]. Thus, it is imperative for novel drugs to restrain the pathogenetic progress of SIVD.
Primary features of Traditional Chinese medicine (TCM) are synergistic and complementary effects and multi-target action [8]. Fo-Shou-San (FSS) is a traditional Chinese herbal decoction first published in Puji Benshi Fang in the Song Dynasty (AD 1132) by Xu Shuwei, containing Angelica sinensis (AS, Danggui, DG) and Ligusticum wallichii (LW, Chuanxiong, CX) in a weight ratio of 3:2. Initially, FSS was intended for obstetrical and gynecological diseases, such as dystocia and intrauterine fetal death [9]. According to TCM theory and previous research, FSS enhances immune, circulatory, and hematopoietic systems, whereby DG dispels blood stasis and CX nourishes blood circulation [10]. In recent years, FSS has been widely applied to various diseases, especially cerebral vascular impairment [11][12][13]. Accordingly, we set out to determine whether FSS could treat vascular dementia in mice, underlying its specific mechanism.
There is a rising focus on the impact of ferroptosis on neurological diseases [14]. Studies have revealed that inhibitors of ferroptosis lessen neuronal deterioration and improve neurological impairment induced by intracerebral hemorrhage and ischemic stroke [15][16][17]. Ferroptosis depends on the accumulation of iron ions, leading to nonapoptotic oxidation and lipid super oxidation including reactive oxygen species accumulation and polyunsaturated fatty acid consumption [18]. Three key components represent the main molecular mechanism of ferroptosis: reduced glutathione (GSH) utilized by glutathione peroxidase 4 (GPX4), cystine uptake by glutamate transporter solute carrier family 7 membrane 11 (SLC7A11), and existence of iron ions and lipid reactive oxygen species (ROS) [19].
Research has demonstrated a close relationship between vascular cognitive impairment (VCI) and ferroptosis [20]. In CCH-induced cognitive damage, iron deposition leads to oxidative stress wherein the most severe neuronal damage occurs in the CA1 region where iron most accumulates [21,22]. Cognitive impairment is associated with abnormal iron accumulation in SIVD patients' cortical zones [23]. Moreover, researchers found that enlarged nuclear factor erythroid 2-related factor 2 (NRF2) expression ameliorates the cognitive impairment caused by CCH [24].
Recently, it has been found that NRF2 performs a nuclear transcription factor to regulate cellular antioxidant response. Once activated, NRF2 binds with antioxidant response elements (ARE) to initiate a series of genes against oxidative stress [25,26]. There are two primary types of antioxidant substances related to NRF2 in the nervous system: GSH-related regulatory enzymes, and Heme oxidation system, in particular, heme oxygenase-1 (HO-1) [27,28]. Research has demonstrated that activation of NRF2 and its downstream target genes in vascular dementia models clarifies neuroprotective influence of NRF2 on ischemic injury and oxidative stress [29][30][31][32][33].
Herein, we set out to explore a potential therapeutic for SVID. First, we applied FSS treatment in rUCCAOinduced cognitive impairment mice. We then investigated the underlying pharmacological mechanisms whereby FSS regulates the NRF2/HO-1 pathway against ferroptosis.

Preparation of FSS
FSS comprises Angelica sinensis (Lot No. G01201124-06) and Ligusticum wallichii (Lot No. B21051401-01), which were obtained from the Guangzhou Medicine Company in accordance with the standard in Pharmacopoeia of People's Republic of China. The herbs were separated into small fragments in a weight ratio of 3:2, soaked in 10 volumes of purified water for 2 h, and boiled in the volatile oil extraction device for 6h. Residues were immersed in eight volumes of distilled water overnight. The extract was boiled three times for one hour. The filtrate was then concentrated by rotary evaporator. Lastly, the crude water-extracted drugs combined with the volatile oil were stored at -20°C [9,[34][35][36].

Animals
Male C57BL/6 mice, 12 weeks old and 25-28 g, were acquired from Experimental Animal Center of Guangzhou University of Chinese Medicine (License: SCXK2018-0034). All procedures under principles for animal experiments were approved by Guangzhou University of Chinese Medicine Animal Ethics Committee (approval ID: 20210615006). Moreover, nimodipine (N837938, Mack-lin, China), a dihydropyridine Ca 2+ antagonist, has been used for neuropharmacological treatment of vascular cognitive impairment [37,38], so that we used nimodipine as positive control.
The rUCCAO mice were placed under general anesthesia through a facemask with 5% isoflurane for anesthesia induction and 1.5% isoflurane for anesthesia maintenance. Exposed by a middle cervical cut, the right common carotid artery was double ligated with 6-0 silk distal and proximal to heart. The same process without carotid ligation was operated on the sham group. Body temperature was sustained until mice were woken from anesthesia [39][40][41][42][43].
The mice were separated into six groups (n = 12/group): sham group (same procedure as rUCCAO without carotid ligation and physiological saline administration), VD model group (subjected to rUCCAO and administrated with physiological saline), nimodipine group (subjected to rUCCAO and administrated with 15 mg/kg nimodipine), FSS low dosage (FSSL) group (subjected to rUCCAO and administrated with 0.5 g/kg FSS), FSS middle dosage (FSSM) group (subjected to rUCCAO and administrated with 1 g/kg FSS), and FSS high dosage (FSSH) group (subjected to rUCCAO and administrated with 2 g/kg FSS). Intragastrical administration followed rUCCAO once daily from the time of surgery (day 0) for 32 days. Following behavioral tests performed from day 24 to day 32, mice were sacrificed to store brain samples for further analysis (Fig. 1A).

Novel Object Recognition Test
The experiment site was kept bright and quiet. On day 24, mice were placed into a white experimental box (40 cm × 40 cm × 40 cm) with their back opposite to two of the same, equidistant objects. After five minutes of exploration, mice were removed. On day 25, one of the familiar (F) objects was substituted with a new object (N), and mice were placed into the box again for 5 minutes. The experimental box along with the objects were disinfected with 75% ethanol solution between each test. Time spent exploring familiar (F) object and new (N) object was recorded to calculate discrimination index percentage as ((N-F)/(N+F)) *100% [44].

Morris Water Maze Test
Procedures were performed as described in a previous study [45]. In a five-day learning period (days 27-31), mice were randomly trained to discover a hidden platform at the center of one of the four equal quadrants of a pool in order to calculate the average escape latency. On the sixth day of testing (day 32), swimming parameters were recorded of mice exploring the pool without the platform to assess cognitive memory.

Brain Tissue Sample Preparation
On day 33 following rUCCAO surgery, cervical dislocation was performed followed by intracardiac injection of phosphate buffer solution (PBS). Brain tissue was collected and divided for liquid nitrogen storage at -80°C for biochemical analysis, as well as for morphological and immunohistochemistry analysis as IHC, IF and TUNEL fixed in 4% paraformaldehyde.

Immunohistochemistry
The brains of four mice from each group were embedded in paraffin, cut 5 µm width, dewaxed with xylene, and rehydrated by an alcohol series. After dipping in Nissl Staining Solution (Beyotime, Shanghai, China) at room temperature for 10 min, the slides were dehydrated through a succession of ethanol, permeabilized by xylene and mounted via neutral resin. Captured by a microscope (Nikon 80i, Tokyo, Japan), quantitative analysis per visual field of the hippocampus CA1 region was performed by Image J software, which identifies Nissl Bodies staining blue-purple as surviving neurons to calculate percentage of Nissl-positive neurons to total number of neurons [46]. For HE staining, slices were dipped in 5 min-hematoxylin solution and in 1 min-eosin solution to image and measure using the same methods for Nissl staining.

Immunofluorescence
The frozen sections of brain (30 µm) were permeabilized in 0.1% Triton X-100 and saturated with primary antibodies of anti-SLC7A11 antibody, anti-ROX antibody, anti-4HNE antibody, anti-NRF2 antibody, anti-HO-1 antibody, at 4°C for 48 h. Next, the sections were saturated with goat secondary antibodies for 2 h; the nucleus was visualized with DAPI (1.0 µg/mL) for 10 min at dark. Using a fluorescent microscope (DMI8, Leica, Germany), images were quantified in intensity of immunofluorescence staining by Image J analysis of mean gray value [47,48].

TUNEL Staining
TUNEL detection employed the one-step TUNEL kit (Beyotime, Shanghai, China); brain slices were saturated in 4% paraformaldehyde and incubated in immunostaining strong permeable solution at room temperature. Then, slices were saturated with TUNEL reaction mixture, PBSwashed three times and stained with DAPI (1.0 µg/mL) in the dark. Images were seized and measured by the same method as immunofluorescence staining.

Quantitative Real-Time Polymerase Chain Reaction
RNA was isolated by RNAiso Plus (Takara, Beijing, China) from brain tissue. PrimeScript™RT Master Mix (Takara, Beijing, China) compounded first-strand cDNA under the corresponding protocol. The cDNA was enlarged by TB Green® Premix Ex Taq™ II (Takara, Beijing, China) in a 20 µL reaction mix in specified equipment (ABI-7500, Applied Biosystems, United States) using the following primers, see Table 1.

Statistical Analysis
Data were analyzed by GraphPad Prism 8.0 software (GraphPad Inc., La Jolla, CA, USA), and results shown as mean ± SEM with One way analysis of variance (ANOVA) for multigroup comparison according to experimental requirements. A p-value < 0.05 indicated statistical significance.

FSS Improved Cognitive Functions after Chronic Cerebral Hypoperfusion
Some reports have found that chronic cerebral hypoperfusion triggers cognitive damage [53,54]. To assess whether FSS could improve cognitive functions, we conducted a series of behavioral tests. FSS made a considerable improvement in discrimination skills for the novel object recognition test (Fig. 1B). In the probe test of Morris water maze test, results indicated that the number of platform crossings (Fig. 1C) and time spent in target quadrant (Fig. 1D) increased in mice treated with FSS and nimodipine compared to model group. As for the learning period, administration of nimodipine and FSS in mice raised their spatial learning ability to locate a find hidden platform as shown by less average escape latency versus the rUCCAO mice (Fig. 1E). This is a clear indication that FSS could improve cognitive functions after chronic cerebral hypoperfusion.

FSS prevented Hippocampal Neuron Loss through Chronic Cerebral Hypoperfusion
We applied HE and Nissl staining to examine neuron death in the CA1 area of hippocampus damaged by rUC-CAO. In HE staining, FSS and nimodipine treatment attenuated the morphological changes of VD mice, while neurons in the sham group had round or oval nuclei ( Fig. 2A and C). Similarly in Nissl staining, FSS and nimodipine treatment prohibited neuron loss ( Fig. 2B and D). Together, these data indicate that FSS could prevent hippocampal neuron loss through chronic cerebral hypoperfusion.

FSS Attenuated Apoptosis and Oxidative Stress after Chronic Cerebral Hypoperfusion
TUNEL staining showed that neurons in the CA1 area of hippocampus displayed apoptosis ameliorated by FSS and nimodipine (Fig. 3A,B). Moreover, as for the MDA, SOD and GSH-PX content in serum (Fig. 3C,D and E), MDA and GSH-PX were decreased by FSS and nimodipine versus the model group, while FSS showed no impact on SOD. Compared with model group, total glutathione, GSH and GSSG levels in corpus striatum decreased in the FSS and nimodipine treatment group. Accordingly, FSS and nimodipine could improve GSH/GSSG ratio in contrast with the model group (Fig. 3F). All data indicate that FSS could attenuate apoptosis and oxidative stress after chronic cerebral hypoperfusion.

FSS Suppressed Ferroptosis after Chronic Cerebral Hypoperfusion
To explore FSS against rUCCAO-induced ferroptosis, we revealed transcriptional mRNA measured by RT-qPCR and at post-transcriptional level by immunofluorescence staining. The FSS and nimodipine treatment group presented lower expression of GPX4 and SLC7A11, and different level of PTGS2 and LPCAT3 compared to the VD model group in cerebral cortex (Fig. 4A) and hippocampus (Fig. 4B). Further, FSS lessened expression level of SLC7A11, ROX and 4HNE in the CA1 area of hippocampus (Fig. 5A,B and C), indicating that FSS could decrease ferroptosis through System Xc−. These results demonstrate that FSS could suppress ferroptosis after chronic cerebral hypoperfusion.

FSS Regulated NRF2/HO-1 Pathway after Chronic Cerebral Hypoperfusion
NRF2/HO-1 signaling can lessen ferroptosis by reducing oxidative stress [55]. As shown in Fig. 6A and B, NRF2 and HO-1 mRNA were decreased by FSS and nimodipine in cerebral cortex and hippocampus versus the model group. The immunofluorescence staining in the CA1 area of hippocampus similarly indicate that FSS and nimodipine decreased NRF2 and HO-1 expressions compared with the model group ( Fig. 6C and D). This suggests that FSS could regulate NRF2/HO-1 pathway after chronic cerebral hypoperfusion.

Discussion
The decline of cerebral blood flow perfusion contributes to cognitive malfunction [56]. As an approach of permanent occlusion of major brain vessel in mice, UC-CAO leads to a drop by 35-55% in the ipsilateral cortical perfusion to induce CCH [57]. The animal model rUC-CAO was established to imitate the disease conditions of SIVD in order to investigate the impact of CCH on the VD pathophysiological mechanism and identify new treatment options [58].
In terms of syndrome differentiation and treatment of TCM, FSS has been extensively applied for cerebrovascular diseases to promote and nourish blood flow [59,60]. Herein, we verified that FSS could improve rUCCAOinduced cognitive deficiency regulated by NRF2/HO-1 pathway against ferroptosis.
We found that FSS treatment alleviated cognitive behavior disorders as well as hippocampal neuron damage after chronic cerebral hypoperfusion. FSS lessened impaired learning and memory abilities with higher discrimination index, more platform crossings, increased time spent in target quadrant and quicker escape latencies in the FSS treatment group compared with the VD model. Moreover, as evidenced by HE and Nissl staining, FSS improved neuronal morphological impairment in the VD group whereby hippocampal neurons disappeared. Chronic cerebral ischemia affects the hippocampus site connected with learning and memory skill [61]. Hence, the results of behavioral and immunohistochemical tests imply that FSS helps offset cognitive deficits.
Recently, apoptosis, oxidative stress, cholinergic nerve malfunction and energy metabolism syndromes have been associated with the intricate pathogenesis of VD [62,63]. According to our study, FSS treatment moderated the apoptosis, as shown via TUNEL staining. Similarly, the antioxidant effects of FSS have been validated by MDA and GSH-PX content in serum, and by GSH/GSSG ratio in corpus striatum after chronic cerebral hypoperfusion. MDA was observed as the final product of peroxidation reac- tion between free radicals and lipids, while antioxidant enzymes such as SOD and GSH-PX can maintain cellular redox homeostasis [64][65][66]. Cystine synthesizes and affects GSH, which disturbs the balance between lipid peroxidation and antioxidant stress [67]. Thus, one of the neuroprotective features of FSS is an antioxidant system that lessens lipid peroxidation injury.
Ferroptosis is a regulated cell death characterized by iron-mediated oxidative damage [18]. Canonical ferroptosis deactivates the main membrane mechanisms against peroxidation damage [68]. System Xc− plays an essential role in regulating lipid peroxidation. SLC7A11 is a cystine antiporter that sustains redox balance of GSH synthesis [69]. GPX4 is a selenoenzyme that converts antioxidant GSH to oxidized GSSG, thereby protecting cells from ferroptosis by restraining cytotoxic lipid peroxidation such as esterified lipid hydroperoxide [70]. Inhibition of GPX4 or system Xc− induces ferroptosis by dropping GSH while accumulating ROS and lipid peroxidation [71]. Brains that are replete with phospholipids are more vulnerable to lipid peroxidation damage [72]. Notably, ferroptosis has be found as a vital factor of VCI [73]. Herein, we found that SLC7A11 and GPX4 mRNA were increased in VD models in the cerebral cortex and hippocampus, and were reduced by FSS. Compared to the VD model group, the FSS treatment group showed different mRNA levels of PTGS2 and LP-CAT3, the main ferroptosis targets to deactivate GPX4 [74]. Further, our quantitative analysis of immunofluorescence staining revealed that FSS decreased the products of ROS as SLC7A11, ROX and 4HNE in the hippocampus CA1 area of mice, suggesting that FSS could decrease ferroptosis by enhancing System Xc− activity [75,76]. These data indicate that FSS could have protective effects against ferroptosis of the SLC7A11/GPX4 axis dysfunction after chronic cerebral hypoperfusion.
NRF2, a transcription factor, plays a role as a critical endogenous regulator of ferroptosis involved in genes of antioxidant and iron metabolism [77,78]. NRF2 stays inactive with Kelch-like ECH-associated protein 1 (Keap1) in cytosol [79]. Activated NRF2 is isolated from Keap1 and transferred to the nucleus, combining with enhancer sequences named ARE that target genes including HO-1 [80,81]. HO-1 has a protective effect against oxidative injury, apoptosis and inflammation, while abnormal HO-1 levels are related to the pathogenesis of neurodegeneration [82,83]. Studies have also verified that activation of NRF2 signaling shows neuroprotective potential to treat neurodegenerative diseases by inhibiting ferroptosis [28,84,85]. Moreover, NRF2 dysfunction causes oxidative damage in CCH-induced rats [86]. Thus, inactivated NRF2/HO-1 axis could be a promising treatment target in preventing ferroptosis. In our study, NRF2 and HO-1 mRNA in cerebral cortex and hippocampus, and NRF2 and HO-1 fluorescence intensity in the hippocampus CA1 region, indicate that FSS decreased these expressions compared with the model group. Overall, we propose that FSS could inhibit CCH-caused ferroptosis by regulating NRF2/HO-1 pathway.

Conclusions
Our study reveals that FSS likely ameliorates chronic cerebral hypoperfusion-induced cognitive impairment in vivo, and is potentially neuroprotective against ferroptosis by regulating the NRF2/HO-1 pathway. Additional studies are warranted to identify the fine pharmacological mechanisms of FSS.

Availability of Data and Materials
The data used and analyzed in the study are available from the corresponding author upon reasonable request.

Author Contributions
JYW and QW designed the study. JYW, JHS, YFX and GCC implemented the experiments. WZ and QW provided reagents and materials. CY and LND analyzed the data. JYW, JHS, YFX and GCC edited the manuscript. CY, LND, WZ and QW revised and supervised the paper.

Ethics Approval and Consent to Participate
All procedures under principles for animal experiments were approved by Guangzhou University of Chinese Medicine Animal Ethics Committee (approval ID: 20210615006).