This study was performed on male C57BL/6 mice aged 14–16 weeks and weighing 25–28 g, the mice were supplied by the Experimental Animal Center of Fujian Medical University, and the study was approved by the IACUC of FJMU (ethics protocol#number was FJMUIACUC 2020-0015). The animals were housed under standard vivarium conditions (12-h dark/light cycle, 22 ± 1°C ambient temperature) with an ad libitum access to food and water. All efforts were made to minimize animal suffering and reduce the number of animals used.

Animals were randomly assigned to following groups (n = 10 per group):

1. Control group: No intervention;

Rosmarinic acid (Yuanye technology, Shanghai, China, Cat# B20862) was administered once a day by i.p. injection of 40, 60, and 80 mg/kg RA, while NS was administered by i.p. injection in the normal saline group (Wei et al., 2021).

After 30 min of RA injection, the animals were intraperitoneally injected with MPTP (Sigma-Aldrich, MO, United States, Cat# M0896) four times at a dose of 20 mg/kg at 2-h intervals. The same number of NS was injected into the NS group. The schedule of animal experiments is shown in Figure 1A.

Exponentially growing human neuroblastoma cell line SHSY5Y (Chinese Academy of Sciences Cell Bank, Shanghai, China, Cat#SCSP-5014) were maintained in a 1:1 mixture of Ham’s F12 and Dulbecco Modified Eagle Medium (DMEM) (Sigma-Hyclone, MO, United States, Cat# SH30023.01) supplemented with 10% heat-inactivated FBS (Sigma- Gibco, MO, United States, Cat# 16140089), 100 IU/ml of Penicillin and 100 ug/ml of streptomycin (Sigma- Gibco, MO, United States, Cat#15140148), in a humidified atmosphere of 5% of CO₂ in the air at 37°C (Thermo Fisher Scientific, MA, United States). Cells were sub-cultured once they reached 80–90% confluence. Rosmarinic acid (20, 40, 60, 80, 100, and 200 μM) and MPP⁺ (400 μM) (Sigma-Aldrich, MO, United States, Cat# M7068) were dissolved in normal saline then added into the culture medium. The cells were seeded into 25 cm² culture flasks for 24 h, then divided into 9 groups of: control, PBS, MPP⁺, MPP⁺ + 20 μM RA, MPP⁺ + 40 μM RA, MPP⁺ + 60 μM RA, MPP⁺ + 80 μM RA, MPP⁺ + 100 μM RA and MPP⁺ + 200 μM RA.

The mice’s motor performance was assessed using the following experiments: pole test, open-field test (OFT), and rotarod test.

The pole test is a useful method to measure the bradykinesia and movement balance in a mouse PD model. The pole test was performed on the 1st day after the last MPTP injection. The animals were placed facing upward near the top of a wooden pole (diameter 8 mm, height 50 cm with a rough surface). The time needed for the mice to climb down the pole and place four feet on the floor (defined as locomotor activity time, T-LA) was recorded. The mice were pre-trained three times before MPTP injections. Every mouse was tested three times, and the average of three trials was calculated for statistical analyses (Park et al., 2013).

The ambulatory behavior and exploratory ability were assessed in an open field test conducted on the 2nd day after MPTP injection. The animals were individually placed in the center of an acrylic apparatus (40 cm*40 cm*40 cm). The tracks, hotspots, and average speed, were monitored using Smart 3.0 software (Panlab, Barcelona, Spain). The apparatus was cleaned with 75% ethanol solution and dried between trials to avoid the presence of olfactory cues (Li et al., 2016).

Sensorimotor coordination was assessed with a rotarod test [YLS-4C Rota Rod with automatic timers and falling sensors (Yiyan Scientific, Shangdong, China)]. Animals were evaluated on the rotarod day after MPTP injection and pretrained three times before PD induction. All mice were pretrained for three consecutive days with two trials per day (5 rpm, 1 min → 10 rpm, 2 min → 15 rpm, 3 min) with 15-min intervals between each trial. In the test session, the test rotation speed was increased to 40 rpm, and the length of time taken until the mouse fell from the rod was recorded. Every mouse was tested three times, and the average of three trials was calculated for statistical analyses. The rotarod test was conducted by examiners blinded to the treatment.

On the third day after MPTP injection, the animals were sacrificed by anesthetic overdose, and the brains were immediately removed with different methods for the following experiments.

The mice were sacrificed, as mentioned above. The brains were removed, and the SN was dissected and processed with 3% glutaraldehyde (Tianjin Hengxing Chemical Reagent Co., Ltd., Tianjin, China), 1.5% paraformaldehyde (Tianjin Hengxing Chemical Reagent Co., Ltd., Tianjin, China), and 0.1 M PBS (Sigma- Gibco, MO, United States, Cat# 70011069) at 4°C per 24 h each. This procedure was followed by post-fixations in 1% osmium tetroxide and 1.5% potassium ferrocyanide at 4°C for 1.5 h. After washing with PBS, the samples were dehydrated in a graded series of ethanol (75%, 95%, 100% v/v) and embedded in an Epon-Araldite solution (Ted Pella, Redding, CA, United States) at 60°C for 72 h. Then, the sections were cut at 100 nm thickness using an ultramicrotome (Leica, Wetzlar, Germany) and imaged under an EM208 transmission electron microscope (Philips, Eindhoven, the Netherlands).

The mice were sacrificed and intracardially perfused with 30 ml of 0.1 M PBS and then fixed with 4% PFA. The brains were removed and post-fixed in 4% PFA at 4°C overnight and then immersed in a solution containing 30% sucrose (BioSharp, Hefei, China, Cat# BS127A) 0.1 M PBS for cryoprotection. The tissues were cryo-sectioned at 15 μm thickness using a cryo-microtome (Leica, Wetzlar, Germany) and then placed on coated slides.

The slides were pretreated with 3% hydrogen peroxide for 15 min to remove endogenous peroxidase activity. The slides were incubated overnight with the rabbit anti TH antibody (1:1,000 dilutions) (Abcam, Cambridge, UK, Cat# ab112) in 5% NGS (Beijing Zhongshan Jinqiao Biotechnology Co., Ltd., Beijing, China, Cat# ZLI-9021) +PBST, rinsed with 0.1 M PBS three times, 5 min each, and then incubated with biotinylated goat anti-rabbit IgG (Beyotime Biotechnology, Shanghai, China, Cat#A0277) at room temperature for 2 h. The second antibody was removed, and the slides were rinsed with 0.1 M PBS and incubated with the avidin-biotin-peroxidase complex reagent at room temperature for 1 h. The samples were colored in 3,3-diaminobenzidine (DAB) solution (Thermo Fisher Scientific, MA, United States, Cat# 34002) for 3–5 min using distilled water to terminate the response. The number of DA (TH positive) neurons was only counted from the sections in the region containing the medial terminal nucleus (MTN) because this region has been previously shown to express the highest level of virus-mediated gene expression after intrastriatal infection. We also used the MTN as a landmark to evaluate consistent levels of SNc (Huang et al., 2017). One section per animal and three animals per group were analyzed. We also did normalization to divide the number of DA neurons by the area (number of DA neurons/area). All the stained sections were viewed and photographed using a light scope to measure the density of TH-positive cells in the SN, and the images were scanned using ImageJ software (ImageJ, NIH, United States) to complete the quantitative analysis of DA neurons.

Following the manufacturer’s instruction, the MMP change was determined using a JC-1 Assay Kit (Beyotime Biotechnology, Shanghai, China, Cat# C2003S). In brief, primary cortical neurons (5 × 10⁵ cells/well) cultured in 6-well plates were harvested. After washing, neurons were resuspended in a 500 μL growth medium and incubated with JC-1 working solution (2 μM final concentration) in the dark at 37°C for 20 min. For positive control, the mitochondrial uncoupler, CCCP (10 μM final concentration), was added simultaneously with JC-1. CCCP is a positive control group in JC-1, which is a powerful uncoupling agent for mitochondrial oxidative phosphorylation, which promotes the permeability of mitochondrial intima to H⁺, resulting in mitochondrial apoptosis. The loss of membrane potential on both sides of the intima induces apoptosis. Samples were then analyzed using a confocal microscope (LSM 750, Zeiss, Gottingen, Germany). Data were processed with the ImageJ software, and the results were represented as the relative ratio of green to red fluorescence intensity.

The intracellular ROS levels were measured using a Reactive Oxygen Species Assay Kit (Beyotime Biotechnology, Shanghai, China, Cat# S0033S) 2′, 7′-dichlorofluorescein-diacetate (DCFH-DA), which is easily oxidized to fluorescent dichlorofluorescein (DCF) by intracellular ROS, is its principal component, and therefore, the ROS levels were quantified. Briefly, the cells were seeded in 96-well plates as described above and exposed to MPP₊ and various concentrations of RA for different time intervals. The cells were incubated with DCFH-DA for 20 min at 37°C and then observed using a confocal microscope (LSM 750, Zeiss, Gottingen, Germany). Data were processed with the ImageJ software, and the results were represented as the relative ratio of green to red fluorescence intensity.

Cell viability was evaluated by an enhanced cell counting kit-8 (CCK-8) assay following the manufacturer’s instruction (Dojindo, Kumamoto, Japan, Cat# CK04-11). In short, SH-SY5Y were seeded in a 96-well plate at a concentration of ×1 104 (200 μL/well) for 7 days. After treatment, add 10 μL of CCK-8 reagent to each well and incubate at 37°C for 2 h. Use a microplate reader (SpectraMax®i3x, Molecular Devices, United States) to measure the sample’s optical density (OD) value at 450 nm.

Three days after MPTP injection, the animals were sacrificed by anesthetic overdose and intracardially perfused with 30 ml of ice-cold 0.1 M PBS. Each SN was dissected and homogenized on ice in RIPA lysis buffer (Beyotime Biotechnology, Shanghai, China, Cat# P0013) containing phenylmethylsulfonyl fluoride (PMSF) (Beyotime Biotechnology, Shanghai, China, Cat# ST506) for 30 min. Tissue lysates were obtained by centrifugation (Thermo Fisher Scientific, MA, United States) at 12,000 rpm for 5 min at 4°C.

The protein concentration was quantified using bovine serum albumin as the standard by a BCA protein assay (Beyotime Biotechnology, Shanghai, China, Cat# P0012). After tissue lysate of cells was mixed with loading buffer and boiled for 5 min, 30 μg of denatured protein was loaded per lane, resolved by 10% SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel electrophoresis) (Beyotime Biotechnology, Shanghai, China, Cat# P0012A) for 90 min at 80 V, and then transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, MA, United States, Cat# ISEQ00010). The membranes were blocked with 5% BSA in tris-buffered saline and incubated with primary antibodies (CLPP, HSPA9, HSPE1, LONP1, and SIRT4 for cell specimen; CLPP, HSPA9, HSPE1, LONP1, and SIRT4 for tissue specimen) [HSPE1 (Abcam, Cambridge, UK, Cat# ab181606), HSPA9 (Abcam, Cambridge, UK, Cat# ab129201), CLPP (Abcam, Cambridge, UK, Cat# ab126102), LONP1 (Abcam, Cambridge, UK, Cat# ab103809) and SIRT4 (Proteintech, Wuhan, China, Cat# 66543-1-Ig)] at 4°C overnight, followed by incubation for 90 min at room temperature with secondary antibody conjugated to HRP [HRP-labeled Goat Anti-Rabbit IgG (Beyotime Biotechnology, Shanghai, China, Cat# A0208), HRP-labeled Goat Anti-Mouse IgG (Beyotime Biotechnology, Shanghai, China, Cat# A0216)]. The bound antibodies were then visualized by enhanced chemiluminescence (ECL). Gels were recorded under ultraviolet light with a GelDoc XR system and analyzed with the ImageJ software.

In this study, the statistical analysis used was one-way ANOVA followed by post hoc comparisons using Tukey’s test or Dunnett’s T3 test for comparisons among multiple groups, Tukey’s test was used when variances were assumed equal, and Dunnett’s T3 test was used when variances were assumed to be unequal across groups, and the data were analyzed with SPSS 24.0 (IBM, Armonk, NY, United States). All data are presented as the mean ± SD. p < 0.05 was considered statistically significant.

The effective behaviors were examined on day 1 post-MPTP (Figure 1B). The animals were then subjected to various motor function tests to detect behavioral dysfunction, including the pole test, rotarod test, and OFT.

In the poletest, the MPTP-injected mice showed a significant prolongation of climbing time compared with the control and PBS (control vs. MPTP, p < 0.05; NS vs. MPTP, p < 0.01; Figure 1B), but this prolongation was significantly improved by RA pre-administration (MPTP vs. MPTP+ 80 mg/kg RA, p < 0.05; Figure 1B).

When subjected to the rotarod test, the MPTP-injected mice showed a significantly decreased latency-to fall (control vs. MPTP, p < 0.01; NS vs. MPTP, p < 0.001; Figure 1C), but the pre-injection with RA effectively alleviated this motor abnormalities (MPTP vs. MPTP+ 60 mg/kg RA, p < 0.01; MPTP vs. MPTP+ 80 mg/kg RA, p < 0.001; Figure 1C).

There was a significant decrease in the movement distance of MPTP-injected mice in the OFT (control vs. MPTP, p < 0.01; NS vs. MPTP, p < 0.001; Figure 1D), but this deficit was improved with the pre-injection with RA (MPTP vs. 60 mg/kg RA, p < 0.05; MPTP vs. MPTP+ 80 mg/kg RA, p < 0.05; Figure 1D). Additionally, a significant decrease was observed in the average movement speed of MPTP-injected mice (control vs. MPTP, p < 0.01; NS vs. MPTP, p < 0.001; Figure 1E), but this deficit was improved with the pre-injection with RA (MPTP vs. MPTP+ 60 mg/kg RA, p < 0.05; MPTP vs. MPTP+ 80 mg/kg RA, p < 0.01; Figure 1E). Representative examples of movement paths are shown in Figure 1F.

The MPTP-injection showed a significant effect on the dopaminergic neurons in SN (control vs. MPTP, p < 0.001; NS vs. MPTP, p < 0.001; Figure 2). Compared with the MPTP injection alone, pre-injection with RA significantly decreased the loss of TH⁺ cells in the SN. Interestingly, no difference was observed between different doses of RA (p > 0.05; Figure 2).

The changes in mitochondrial ultrastructure in MPTP-mice were investigated by transmission electron microscopy 30.5 h after RA administration to evaluate the therapeutic effect of pre-injection of RA on the ultrastructural abnormalities of brain mitochondria after MPTP administration.

The mitochondria were categorized into the following types (Del Dotto et al., 2017): more than four cristae (type I), two or three cristae (type II), not more than one crista (type III) (Figures 3A,B). The control group showed 84% type I mitochondria, and the NS group showed 92%. However, the mitochondria suffered extensive damage in the MPTP injury group: about 63% of class III mitochondria, 28% of class II mitochondria, and 9% of class I mitochondria, indicating a dramatic reduction of the mitochondria cristae; meanwhile, the length of mitochondria was significantly shortened. However, RA pre-injection significantly alleviated abnormal mitochondrial ultrastructure, showing 70% of class I mitochondria, 23% of class II mitochondria, and 7% of class III mitochondria in the MPTP+ 60 mg/kg RA group. The mitochondrial length was significantly improved in the MPTP+ 60 mg/kg RA and MPTP+ 80 mg/kg RA group (Figure 3C). These data indicate that RA pre-injection can effectively prevent the ultrastructure of mitochondria from being compromised by MPTP administration.

To investigate how RA protects against dopaminergic neurodegeneration, SH-SY5Y cells were used to identify the related molecular mechanism. As shown in Figure 4A, MPP⁺ resulted a significantly decrease of cell viability (MPP⁺ vs. control, p < 0.001; MPP⁺ vs. PBS, p < 0.001; Figure 4A) and pre-injection with RA significantly reduced the MPP⁺-induced reduction in cell viability in a dose-dependent manner. (MPP⁺ vs. MPP⁺+ 20 μM RA, p < 0.01; MPTP vs. MPTP+ 40 μM RA, p < 0.01; MPP⁺ vs. MPP⁺+ 80 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 100 μM RA, p < 0.001; Figure 4A). To evaluate oxidative stress, we measured the activity of ROS. MPP+ increased the green/red fluorescence intensity in the SH-SY5Y cells (MPP⁺ vs. control, p < 0.05; Figure 4C) and these changes were significantly reversed by RA in a dose-dependent manner (MPP⁺ vs. MPP⁺+ 60 μM RA, p < 0.05; MPP⁺ vs. MPP⁺+ 80 μM RA, p < 0.05; MPP⁺ vs. MPP⁺+ 200 μM RA, p < 0.05; Figure 4B). Therefore, we further evaluated mitochondrial function after MPP+ and RA administration. JC-1 was then utilized to detect mitochondrial membrane potential (MMP). As expected, the green/red fluorescence intensity increased in the MPP₊ group when stained with JC-1 (MPP⁺ vs. control, p < 0.001; MPP⁺ vs. PBS, p < 0.001; Figure 5B), and this ratio was markedly decreased by preincubation with RA (MPP⁺ vs. MPP⁺+ 20 μM RA, p < 0.01; MPP⁺ vs. MPP⁺+ 40 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 60 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 80 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 100 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 200 μM RA, p < 0.001; Figure 4C).

To explore whether the treatment with RA is related with the mtUPR, Western blotting was used to detect the expression of molecular chaperones and proteases in the mtUPR proteins, including CLPP, LONP1, HSAP9, and HSPE1. 1) An obvious increase was observed in the expression of CLPP in the MPP⁺ group compared with the control and PBS (control vs. MPP⁺, p < 0.001; PBS vs. MPP⁺, p < 0.001; Figure 5A), which was significantly improved with the pre-administration of 40, 60, 80, 100 or 200 μM RA (MPP⁺ vs. MPP⁺+ 40 μM RA, p < 0.01; MPP⁺ vs. MPP⁺+ 60 μM RA, p < 0.05; MPP⁺ vs. MPP⁺+ 80 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 100 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 200 μM RA, p < 0.001; Figure 5A). 2) Compared with the control and PBS, MPP⁺ led to an increase in the expression of LONP1 (control vs. MPP⁺, p < 0.001; PBS vs. MPP⁺, p < 0.001; Figure 5B), while the pre-injection of 40, 60, 80 or 100 μM RA reduced this increase significantly (MPP⁺ vs. MPP⁺+ 40 μM RA, p < 0.01; MPP⁺ vs. MPP⁺+ 60 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 80 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 100 μM RA, p < 0.001; Figure 5B). 3) The expression of HSPA9 in the MPP⁺ group obviously increased compared with the control and PBS (control vs. MPP⁺, p < 0.001; PBS vs. MPP⁺, p < 0.001; Figure 5C), while the pre-administration of 40, 60, 80, 100 or 200 μM RA inhibited this increase significantly (MPP⁺ vs. MPP⁺+ 40 μM RA, p < 0.05; MPP⁺ vs. MPP⁺+ 60 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 80 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 100 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 200 μM RA, p < 0.001; Figure 5C). 4) An obvious increase was observed in the expression of HSPE1 in the MPP⁺ group compared with the control and PBS (control vs. MPP⁺, p < 0.001; PBS vs. MPP⁺, p < 0.01; Figure 5D), which was significantly inhibited with the pre-administration of 20, 40, 60, 80, 100 or 200 μM RA (MPP⁺ vs. MPP⁺+ 20 μM RA, p < 0.05; MPP⁺ vs. MPP⁺+ 40 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 60 μM RA, p < 0.01; MPP⁺ vs. MPP⁺+ 80 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 100 μM RA, p < 0.001; MPP⁺ vs. MPP⁺+ 200 μM RA, p < 0.01; Figure 5D).

For further verification, we explored the molecular mechanism of RA alleviating MPTP-induced SN mitochondrial dysfunction. 1) A significant increase was observed in the expression of CLPP in the SN after injecting with MPTP (control vs. MPTP, p < 0.001; NS vs. MPTP, p < 0.001; Figure 6A), while pretreatment with 40 or 80 mg/kg RA significantly reduced this increase (MPTP vs. MPTP+ 40 mg/kg RA, p < 0.05; MPTP vs. MPTP+ 80 mg/kg RA, p < 0.001; Figure 6A). 2) A significant increase was observed in the expression of LONP1 in the SN after injecting with MPTP (control vs. MPTP, p < 0.01; NS vs. MPTP, p < 0.05; Figure 6B), which was significantly inhibited with the pretreatment with 60 or 80 mg/kg RA (MPTP vs. MPTP+ 60 mg/kg RA, p < 0.01; MPTP vs. MPTP+ 80 mg/kg RA, p < 0.001; Figure 6B). 3) An increased protein expression of HSAP9 in the SN was observed after injecting with MPTP (control vs. MPTP, p < 0.01; NS vs. MPTP, p < 0.05; Figure 6C), and this increase was reversed by the pre-administration of 80 mg/kg RA (MPTP vs. MPTP+ 80 mg/kg RA, p < 0.01; Figure 6C). 4) Compared with the control and PBS, MPTP led to an increase in the expression of HSPE1 in the SN (control vs. MPTP, p < 0.01; NS vs. MPTP, p < 0.01; Figure 6D), while pretreatment with 80 mg/kg RA significantly reduced this increase (MPTP vs. MPTP+ 80 mg/kg RA, p < 0.01; Figure 6D).

For exploring whether RA alleviated MPTP/MPP⁺-induced mitochondrial damage via sirtuin signaling. The SH-SY5Y cells were administered with MPP₊ and various concentrations of RA. Increased protein expression of SIRT4 was observed in the MPP⁺ group compared with the control (control vs. MPP⁺, p < 0.05; Figure 5E), and this increase was reversed by the pre-administration of 80, 100 or 200 μM RA (MPP⁺ vs. MPP⁺+ 80 μM RA, p < 0.05; MPP⁺ vs. MPP⁺+ 100 μM RA, p < 0.05; MPP⁺ vs. MPP⁺+ 200 μM RA, p < 0.05; Figure 5E). For further verification, a significant increase was observed in the expression of SIRT 4 in the SN after injecting with MPTP (control vs. MPTP, p < 0.01; NS vs. MPTP, p < 0.05; Figure 6E), while pretreatment with 80 mg/kg RA significantly reduced this increase (MPTP vs. MPTP+ 80 mg/kg RA, p < 0.01; Figure 6E).