Curcumin Regulates Gut Microbiota and Exerts a Neuroprotective Effect in the MPTP Model of Parkinson's Disease

Objectives The experiment aimed to explore the effects of curcumin on motor impairment, dopamine neurons, and gut microbiota in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) mice model. Methods Mice were randomly assigned to six groups: normal control group, solvent control group, MPTP group, curcumin-low-dose group (40 mg/kg), curcumin-medium-dose group (80 mg/kg), and curcumin-high-dose group (160 mg/kg). After 14 days, each group of mice was subjected to the pole text, the hanging test, and the open-field test. Tyrosine hydroxylase (TH) immunohistochemistry was used to observe the survival of nigrostriatal dopamine neurons. Moreover, ultrastructural changes were observed with a transmission electron microscope in mice striatal tissue cells. Then, 16S rRNA was used to assess changes in the gut microbiota. Results (1) Each dose of curcumin reduced pole climbing time and increased suspension score and total distance moved dose-dependently. (2) All curcumin groups improved cell wrinkling and vacuolar degeneration, increased the number of TH positives, improved cell survival, and the higher the dose of curcumin, the better the effect. (3) There were differences in microbiota composition and a relative abundance among the groups. The relative abundance of Patescibacteria, Proteobacteria, and Verrucomicrobia was higher in the MPTP group. The relative abundance of Patescibacteria, Enterobacteriaceae, Enterococcaceae all decreased in all curcumin groups. In addition, the Kyoto Encyclopedia of Genes and Genomes pathways showed a reduction in the superpathway of N-acetylneuraminate degradation after medium- and high-dose curcumin administration. Conclusions Curcumin regulates gut microbiota and exerts a neuroprotective effect in the MPTP mice model. This preliminary study demonstrates the therapeutic potential of curcumin for Parkinson's disease, providing clues for microbially targeted therapies for Parkinson's disease.


Introduction
Parkinson's disease (PD) is one of the most common neurodegenerative diseases with a high incidence and disability rate, which seriously afects the quality of life of patients [1]. Lewy's body composed of α-synuclein (α-Syn) abnormally gathered in the pars compacta of substantia nigra is an important pathological feature of PD [2]. More and more evidence show that PD is not only a nervous system disease but also a digestive system disease [3].
In the past few decades, human microbiota has become an area of most interest, with many studies focusing on its impact on diseases [4]. Gastrointestinal dysfunction has also been consistently associated with PD and occurs earlier than movement disorders. Gastrointestinal dysfunction can be considered an early biomarker of PD [5]. Moreover, gastrointestinal dysfunction preceding dyskinesia in PD patients may be associated with dysregulation of the "microbegut-brain axis." Lai et al. found changes in the abundance of Erysipelotrichales, Proteobacteria, Prevotellaceae, and Clostridiales in intestinal microorganisms by constructing a PD mice model [6]. In PD patients, the level of Escherichia-Shigella, Proteus, Streptococcus, and Enterococcus were signifcantly higher than that in healthy individuals [7].
Dysbiosis of the gut microbiota and increased intestinal permeability overstimulate the immune system, leading to a neuroinfammatory response [8]. Ten, the α-Syn of the enteric nervous system is misfolded and aggregated and passed through the vagus nerve to the dorsal nucleus [9,10].
One study transplanted gut microbes from PD patients into a healthy mouse model and found that it exacerbated dyskinesia in the mice [3], and antibiotic treatment reversed this efect [11]. In addition, patients with PD constipation who received fecal microbiota transplants saw improvements in their intestinal microbiota imbalance, postural instability, and gait abnormalities [12], suggesting that intestinal microorganisms can be used as targets to improve the condition by changing the composition of intestinal microorganisms.
Curcumin, a polyphenol extracted from the rhizomes of Curcuma longa, is converted into biologically active metabolites in the intestine by microbial digestion [13]. Nowadays, it has been shown to have multiple efects such as anti-infammatory [14], antioxidant [15], anticancer [16], and mitochondrial protection [17]. Te bidirectional interplay between curcumin and gut microbes has been demonstrated, as curcumin is not only metabolized by the enzymes of the gut microbes to produce active metabolites but it also strengthens the intestinal barrier and changes the composition of the gut microbes [18].
Curcumin has shown an extremely high therapeutic potential in PD. In animal models, dyskinesia can be improved after the use of curcumin [19]. Many studies have also confrmed that curcumin increased the survival of tyrosine hydroxylase (TH) striatal fbers and nigrostriatal dense part neurons in rat PD model induced by 6hydroxydopamine (6-OHDA) [20], exhibiting increased dopamine (DA) levels [21]. Concomitantly, curcumin can exert anti-infammatory efects by inhibiting microgliamediated neuroinfammation and reducing interleukin-2, chemokines, and cyclooxygenase-2 [22]. Curcumin has also been shown to increase the electrical activity of hippocampal neurons in rotenone-induced PD model rats [23] and regulate the activity of mitochondrial enzyme complex [24]. In addition, a recent clinical study found that curcumin improved motor and nonmotor symptoms and reduced the aggregation of phosphorylated α-Syn in skin biopsies in PD patients [25].
However, it is not entirely clear whether curcumin can improve clinical performance by modulating the gut microbes in PD mice. Terefore, in this experiment, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) was used to establish PD mice model. We used three diferent doses of curcumin for the intervention and 16S rRNA gene sequencing to analyze the microbiota diversity in a mice model to see if curcumin could regulate the gut microbes and thus play a therapeutic role.

Drug Preparation.
Formulations of MPTP: 100 mg MPTP powder dissolved in 1 mL saline, prepared into 100 mg/mL storage solution, stored in −20°C refrigerator, and diluted to 3 mg/mL before use.
Formulations of low-, medium-, and high-dose solutions of curcumin: 20 mg, 40 mg, and 80 mg of curcumin powder which were accurately weighed added to 50 μl of dimethyl sulfoxide. Ten, the mixture was vortexed and shaken until it was completely dissolved. After adding sesame oil for injection to 5 mL, the mixture was stored in a 4°C refrigerator protected from light. Te curcumin solution was mixed by vortex shaking before use and prepared into 4 mg/mL, 8 mg/ mL, and 16 mg/mL injections, respectively.

2.2.
Animals. 72 clean grade healthy male C57BL/6 mice (6-8 weeks old and 18-22 g) were used in the experiment. Tose mice were housed on a 12-hours light/12-hours dark cycle with food and water available ad libitum. Experimental animals, facilities, and feed were provided by the Animal Experiment Center of Zhejiang Chinese Medical University. All experimental protocols were approved by the Experimental Animal Management and Ethics Committee of Zhejiang Chinese Medical University (Ethics Approval No. 20180604-01).

Animal Model Preparation and Grouping.
Mice were randomly assigned to six groups: normal control (NC) group, solvent control (SC) group, MPTP group, curcuminlow-dose (CUR-L) group, curcumin-medium-dose (CUR-M) group, and curcumin-high-dose (CUR-H) group. Tere were 12 rats in each group.
Mice in the MPTP group received a daily intraperitoneal injection of MPTP (30 mg/kg) for 7 consecutive days to establish a subacute PD model. Te CUR-L, CUR-M, and CUR-H groups received a combination of both treatments in which curcumin (40 mg/kg, 80 mg/kg, 160 mg/kg) was given 1 hour after administration of MPTP intraperitoneally once a day for 7 consecutive days and every 24 hours after fnal MPTP for 7 consecutive days. Te NC group received a daily intraperitoneal injection of sterile saline (30 mg/kg) for 7 consecutive days. Te SC group received a combination of both interventions in which sesame oil containing 1% dimethyl sulfoxide was given 1 hour after administration of sterile saline intraperitoneally daily for 7 consecutive days and every 24 hours after fnal sterile saline for 7 consecutive days.

Animal Experiments.
Te following behavioral tests were performed in each group of mice after 14 days. For the pole test, the climbing bar was made with reference to the literature method [26]. Te mice were placed at the top of the pole (1 cm diameter and 55 cm height). Ten, we corded the time of climbing down the pole. Each animal was tested three times with a minimum of 30 minutes between tests, and the pole climbing time was the average of all three trials. For the hanging test, the mice were strung up by their forepaws on a horizontal wire [27]. Te mice were given three points for grasping the wire with both hind paws, two points for one hind paw, one point for not being able to grab the wire with either hind paw, and zero points for dropping directly. Te results were calculated as average of three Te experiments were conducted from 9 am to 12 pm, and the mice were thoroughly cleaned of feces and urine before the experiments to ensure a clean environment.

Taking Samples and Preparing
Tissues. Mice were placed in their respective sterilized cages after the last behavioral test and allowed to defecate freely. Each mice's feces were collected and kept at −80°C in sterile EP tubes after they were generated. Tree specimens were taken from each group. Afterward, mice were anesthetized with 0.3% pentobarbital sodium (0.025 mL/g). Te right atrial appendage was severed at the same time that a perfusion needle was placed into the left ventricle, and saline (4°C) was quickly infused until the outfow was clear. Four mice in each group were randomly selected to be perfused with 4% paraformaldehyde and then decapitated and the brain was removed. Ten, gradient alcohol dehydration and xylene transparent were performed. Te tissue blocks were then embedded in wax, serially sliced into 5 μm-thick coronal slices, and stored at 4°C until used for immunohistochemical staining. In each group, two mice were chosen at random, and a striatal tissue that was less than 1 mm * 1 mm * 1 mm was extracted. For transmission electron microscopy, the tissues were preserved at room temperature after being treated in 2.5% glutaraldehyde.
2.6. TH Immunohistochemistry Was Used to Observe the Survival of Nigrostriatal DA Neurons. Te prepared tissue sections of each group were subjected to antigen repair by the water bath method, incubated at room temperature, and closed with serum (4°C). Tissue sections were titrated with a mice anti-TH polyclonal antibody (1 : 500) and left overnight (4°C). Hematoxylin was used to stain the sections after diaminobenzidine treatment and secondary antibody incubation was done at room temperature. After that, the sections were fractionated in hydrochloric alcohol, returned to blue, dehydrated in gradient ethanol, and fnally sealed. Two felds of view were taken for each mice under a high magnifcation microscope, and eight felds of view were taken for each group to observe the quantity and morphological alterations in the mice's substantia nigra of THpositive neurons. Te Nano Zoomer digital pathology slide scanner is used to quickly scan slides and convert them into digital sections for data analysis. A barcode was added to the 5′ ends of the primers per sample and universal primers were sequenced. PCR amplifcation was performed in a total volume of 25 μL reaction mixture containing 25 ng of template DNA, 2.5 μL of each primer, 12.5 μL PCR Premix, and PCR-grade water to adjust the volume. Te prokaryotic 16S fragments were amplifed using PCR under the following conditions: initial denaturation at 98°C for 30 seconds; 32 cycles of denaturation at 98°C for 10 seconds, annealing at 54°C for 30 seconds, and extension at 72°C for 45 seconds; lastly, fnal extension was performed at 72°C for 10 minutes. With the help of electrophoresis on 2% agarose gels, the PCR products were verifed. Te PCR products were measured by Qubit and purifed using AMPure XT beads. Te amplicon pools were ready for sequencing, and the Library Quantifcation Kit for Illumina and Agilent 2100 Bioanalyzer, respectively, were used to measure the amount and size of the amplicon library. For the library sequencing, the NovaSeq PE250 platform was employed.

Bioinformatics Analysis and Statistical Analysis.
On a platform called Illumina NovaSeq, samples were sequenced. Based on each sample's distinctive barcode, pairedend readings were assigned to them and the barcode and primer sequence were removed. Using FLASH, paired-end readings were combined. According to the fqtrim (v0.94), quality fltering on the raw reads was carried out under precise fltering settings to produce the high-quality clean tags. Vsearch software (v2.3.4) was used to screen out chimeric sequences. Using DADA 2, we acquired the feature table and feature sequence after dereplication.
Te R package (v3.5.2) was used to create the fgures. Moreover, QIIME 2 was used to determine the alpha and beta diversity. Sequence alignment was performed using Blast, and each representative sequence's feature sequences were annotated using the SILVA database.
Data were analyzed using SPSS 20.0, and measurement data were expressed as mean ± standard deviation. To evaluate if the data had a normal distribution, the Shapiro-Wilk test was utilized. When comparing more than two groups, one-way ANOVA was used, followed by the LSD test if the variance was homogeneous and the Games-Howell test otherwise.

Curcumin Improves Motor Dysfunction in PD Mice.
To evaluate the ameliorative efect of curcumin on dyskinesia in PD mice, behavioral experiments were performed on mice. Te results of the pole text showed that there was no statistical diference between the NC and SC groups. In contrast to the NC group, the pole climbing time was prolonged in the MPTP, CUR-L, CUR-M, and CUR-H groups (P < 0.01). Moreover, the diferences between the MPTP and CUR-L groups were not signifcant (P > 0.05), but the pole climbing time of mice in both the CUR-M and CUR-H groups was shorter compared to the MPTP group (P < 0.05) (Figure 1(a)).
When compared to the NC group, the hanging test showed the elevated scores of the mice in the MPTP (P < 0.01), CUR-L (P < 0.01), CUR-M (P < 0.01), and CUR-H groups (P < 0.05). However, no statistically signifcant diference was seen between the MPTP and CUR-L groups (P > 0.05). When compared to the MPTP group, the scores of the mice in the CUR-M (P < 0.05) and CUR-H groups (P < 0.01) were all improved. Moreover, CUR-H and CUR-L groups difered from one another (P < 0.01) ( Figure 1 Te open feld test showed that the total distance of activity was reduced in both the MPTP group and the CUR-L, CUR-M, and CUR-H groups when compared to the NC group (P < 0.01). Similarly, no discernible diference existed between the MPTP and CUR-L groups (P > 0.05). In comparison to the MPTP group, mice in the CUR-M and CUR-H groups (P < 0.01) had longer total active distances. Additionally, compared to the CUR-L group, the total distance of activity was greater in the CUR-M (P < 0.05) and CUR-H (P < 0.01) groups (Figure 1(c)).
Tese results suggested that curcumin improved dyskinesia in PD mice dose-dependently.

Curcumin Reduces MPTP Neurotoxicity and Protects DA Neurons in PD Mice.
To assess the protective efect of curcumin on DA neurons in PD mice, TH immunohistochemically stained positive cells were used to represent the Similarly, the number of TH-positive neurons was higher in the CUR-M and CUR-H group compared to the MPTP group (P < 0.01), whereas there was no discernible diference between the MPTP and CUR-L groups (P > 0.05). In addition, the number of TH-positive neurons was increased in both the CUR-M and CUR-H groups compared to the CUR-L group (P < 0.05) (Figure 2(b)).
Ultrastructural changes were observed with a transmission electron microscope in mice striatal tissue cells. Te results showed that the striatal neurons in the NC group were oval in shape, with an intact envelope structure, no intracellular protein aggregation, no obvious cytoplasmic edema or vacuolation, an intact mitochondrial membrane structure, and clearly visible internal cristae. When comparing the SC and NC groups, there was no discernible diference in their electron microscopic performance. Te Evidence-Based Complementary and Alternative Medicine striatal tissue in the MPTP group showed structural destruction of neurons, intracytoplasmic edema, massive vacuolar degeneration, nuclear consolidation, chromatin aggregation, destruction of mitochondrial membrane structure, even swelling and rupture, blurred cristae structure, and no autophagosomes could be observed. Te CUR-L group also showed partial structural destruction of neurons, cytoplasmic edema, vacuolar degeneration, blurred mitochondrial cristae, and relatively intact mitochondrial membrane structure compared with the MPTP group. Intracellular edema and vacuolar degeneration were also seen in the CUR-M group, with a relatively intact envelope. When compared with the MPTP group, the mitochondrial cristae in the CUR-M group were clearer, and the mitochondrial membrane results were relatively intact. Autophagic lysosomes are visible in some tissues. Neuronal cell edema and vacuolar degeneration were improved in the CUR-H group compared with the MPTP group. Te rest of the CUR-H group behaved similarly to the CUR-M group and autophagic lysosomes were also seen ( Figure 3). Curcumin was shown to reduce the neurotoxicity of MPTP in PD mice and protect DA neurons dose-dependently. Te results showed that curcumin improved cell wrinkling and vacuolar degeneration, increased the number of TH positives, and improved cell survival.

Curcumin Regulates the Composition of Gut Microbiota in PD Mice.
To test our hypothesis that curcumin could improve the gut microbiota of PD mice, we further performed 16S rRNA gene sequencing. Using the QIIME 2 analysis process with better performance, we obtained the Feature data. Alpha diversity was used to assess diferences in colony richness and diversity between groups (Figures 4(a) and 4(b)). Te results suggest that the samples' microbiota was essentially covered by the sequencing depth and there was no discernible variation between the groups. Beta diversity was further assessed using PCA analysis (Figure 4(c)). Te results

Evidence-Based Complementary and Alternative Medicine
showed that the SC, NC, and MPTP groups were separated, indicating that there were signifcant diferences in the composition of the microbiota between the SC, NC, and MPTP groups of mice. Notably, the CUR-L, CUR-M, and CUR-H groups were separated from the MPTP group, indicating that curcumin has an efect on the gut microbiota of PD mice. To further investigate the characteristics of mice gut microbiota, we explored the relative abundance of mice gut microbiota of six groups. At the phylum level, compared with NC group, the relative abundance of Patescibacteria, Proteobacteria, and Verrucomicrobia in MPTP group increased, while the relative abundance of Bacteroidetes declined. Compared with the MPTP group, the relative abundance of Cyanobacteria in the CUR-L group increased, while the relative abundance of Patescibacteria and Verrucomicrobia decreased. In the CUR-M group, Actinobacteria, Acidobacteria, and Cyanobacteria's relative abundance increased, but Verrucomicrobia's and Patescibacteria's relative abundance declined. In the CUR-H group, relative abundance of Acidobacteria, Firmicutes, Actinobacteria, and Tenericutes increased while Bacteroidetes, Patescibacteria, and Verrucomicrobia dropped (Figure 4(d)).
A heatmap analysis was performed based on the relative abundance of each sample at the family level. Te abundance of Enterobacteriaceae and Enterococcaceae was negatively correlated with the dose of curcumin. More importantly, the relative abundance of Enterobacteriaceae and Enterococcaceae in the CUR-H group has approached that of the NC group (Figure 4(f )).  Evidence-Based Complementary and Alternative Medicine Te Venn diagram suggests that the MPTP, CUR-L, CUR-M, and CUR-H groups all have their own specifc gut microbiota (Figure 4(g)). Taken together, these data showed that PD mice have gut microbiota dysbiosis, and curcumin can regulate the distribution and composition of gut microbiota to improve gut microbiota dysbiosis in PD mice.

Analysis of Diferences in Gut Microbiota. Utilizing
Linear discriminant analysis Efect Size (LEfSe), biomarkers were looked for to identify the distinct bacterial groups within each group. When gut microbiota differences between the groups were analyzed using linear discriminant analysis, the fndings revealed a substantial diference in gut microbiota across the groups.
At the phylum level, the relative abundance of Patescibacteria and Verrucomicrobia in CUR-L is lower than that of the MPTP group, while the relative abundance of Cyanobacteria is higher than that of the MPTP group. Te CUR-M and MPTP groups did not difer from one another. Moreover, the relative abundance of Patescibacteria in CUR-H is lower than that of the MPTP group, while the relative abundance of Tenericutes and Actinobacteria is higher than that of the MPTP group.
At the family level, there was lower abundance of Akkermansiaceae, Saccharimonadaceae, Eubacteriaceaen and Atopobiaceae in the CUR-L group, while Anaeroplasmataceae, Gastranaerophilales-unclassifed, Eggerthellaceae, Rikenellaceae, and Prevotellaceae were more abundant. Te relative abundance of Eubacteriaceae in CUR-M was lower than that of the MPTP group, while the relative abundance of Burkholderiaceae, Prevotellaceae, Eggerthellaceae, Pseudomonadaceae, and Rikenellaceae was higher. Tere was lower abundance of Saccharimonadaceae, Eubacteriaceae, and Family-XIII in the CUR-L group, while Pseudomonadaceae and Eggerthellaceae were more abundant ( Figure 5(a)). In addition, the cladogram showed the differential distribution of gut microbiota between groups ( Figure 5(b)).

Prediction of Gene Function of Gut Microbiota. Te
Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways showed substantial group diferences by Phylogenetic Investigation of Communities by Reconstruction of Unobserved States 2 (PICRUSt2). Tere was a signifcant enrichment of 30 KEGG pathways between the NC and MPTP groups. Te administration of curcumin resulted in a number of functional changes in the gut microbiota of mice, such as superpathway of UDP-glucose-derivedO-antigen building blocks biosynthesis, inosine-5′-phosphate biosynthesis III, superpathway of N-acetylneuraminate degradation, cob (II) yrinate a, c-diamide biosynthesis I (early cobalt insertion), superpathway of glycolysis and Entner-Doudorof and GDP-mannose biosynthesis (Figures 6(a)-6(d)).

Discussion
As the number of people with PD increases, PD is a growing concern. PD is treated chiefy with drugs such as levodopa preparations, and herbal medicine can play a synergistic role. However, none of the existing therapies can stop the degenerative changes in PD, which forces us to look for new therapies for PD. Curcumin is the main active ingredient of traditional Chinese medicine turmeric, which can be converted into biologically active metabolites in the intestinal tract through microbial digestion after consumption. Tus, curcumin can exert a wide range of pharmacological efects through the "microbe-gut-brain axis" and can be used to treat many chronic diseases [28].
Te MPTP-induced PD mice model is more consistent with the course of human PD and partially refects human PD pathology. Terefore, we explored the therapeutic efects of three diferent concentrations of curcumin on PD mice by constructing MPTP model mice at 4 mg/mL, 8 mg/mL, and 16 mg/mL. We assessed the motor function of mice by the pole test, hanging test and open feld test. It is clear that curcumin improves motor impairment in MPTP mice dosedependently. In addition, curcumin also reversed the neurotoxicity of MPTP in PD mice. Inhibiting the function of the mitochondrial complex I under the control of monoamine oxidase, MPTP is a lipophilic neurotoxin that easily crosses the blood-brain barrier and causes degenerative necrosis of DA neurons [29]. Coincidentally, curcumin inhibits monoamine oxidase B and protects DA neurons [17]. Our data showed that diferent doses of curcumin improved DA neuron morphology and number, mitochondrial structure, and neuronal vacuolar degeneration, with higher doses being more efective.
To understand the therapeutic efects of diferent doses of curcumin on gut microbiota, we used 16S rRNA gene sequencing to analyze the diferences in the composition of gut microbiota in each group of mice. From the beta diversity analysis, it is clear that MPTP-induced changes in gut microbiota occurred in the mice model, and diferent doses of curcumin afected gut microbiota. At the phylum level, the relative abundance of Patescibacteria, Proteobacteria, and Verrucomicrobia was higher in the MPTP group. Following curcumin delivery, Patescibacteria and Tenericutes showed dose-dependent changes in relative abundance, with Patescibacteria decreasing and Tenericutes increasing in all groups. Te relative abundance of Cyanobacteria increased in the CUR-H group.
Te changes in Verrucomicrobia [30] and Proteobacteria [31] abundance in the MPTP group were consistent with the results of previous studies. Verrucomicrobia at low concentrations can aid in the intestinal barrier's metamorphosis. However, excessive quantities of Verrucomicrobia may increase intestinal leakiness and encourage toxin translocation to the enteric nervous system if the intestinal barrier has been compromised and there is infammation [30]. Additionally, Proteobacteria gives rise to elevated luminal lipopolysaccharide levels in PD, which may impair intestinal integrity [32]. However, there are no reports about the relationship between Patescibacteria and PD. Patescibacteria is an interesting but currently not widely understood anaerobic bacterium [33]. In the cerebral cortex and hippocampus of a rat model of chronic unpredictable mild stress (CUMS), Zhang et al. found that it was positively associated with 5hydroxy tryptamine (5-HT), brain-derived neurotrophic factor (BDNF), and glucagon-likepeptide-1 (GLP-1) [34]. Unfortunately, Patescibacteria has no relevant studies in animal models of PD. Several studies have confrmed that 5-HT [35], GLP-1 [36], and BDNF [37] all have therapeutic efects on PD model mice. However, in the MPTP group, the relative abundance of Patescibacteria in this study increased signifcantly and decreased with the use of curcumin. It needs to be considered that Patescibacteria is an autotrophic microorganism symbiotic with other bacteria [38] and its physiological and biochemical metabolic pathways are not yet clear. Further PD animal model studies are urgently needed to clarify the changes in this microbiota and its neurotransmitters.
Cyanobacteria are linked to the etiology of PD because they create β-N-methylamino-L-alanine, which depletes glutathione, a key antioxidant, by activating metabotropic glutamate receptors [39]. Tis neurotoxin is also incorrectly incorporated into human proteins to replace serine, leading to misfolding and aggregation of proteins [40]. Te relative abundance of Cyanobacteria increased only in the CUR-H group, which may remind us to pay attention to the dose when using curcumin.
At the family level, the relative abundance of Enterobacteriaceae, Enterococcaceae, Saccharimonadaceae, and Bifdobacteriaceae was higher in the MPTP group, and the relative abundance of Rikenellaceae, Prevotellaceae, and Burkholderiaceae was lower. After curcumin administration, the relative abundance of Eubacteriaceae, Rikenellaceae, Prevotellaceae, Anaeroplasmataceae, Eggerthellaceae, Ruminococcaceae, and Burkholderiaceae all increased and the relative abundance of Saccharimonadaceae, Enterococcaceae, and Eubacteriaceae all decreased. Interestingly, Enterobacteriaceae and Enterococcaceae has a dose-dependent. It should be noted that the relative abundance of Enterobacteriaceae and Enterococcaceae in the CUR-H group has approached that of the NC group.
Bifdobacteriaceae, which can behave as opportunistic pathogens and induce infection and excessive immunological stimulation in immune-compromised people [41], has been found in higher levels in PD patients in many studies [42]. Bifdobacteriaceae can also raise levels of 5-HT [43]; therefore, more research is required to understand its function.
Additionally, the impact of Enterobacteriaceae on PD has been the subject of numerous research. Citrobacter can induce mitochondrial antigen presentation (MitAP). Recently, E. coli has also been shown to cause PD-like motor symptoms in Pink1-/-mice and to produce mitochondriaspecifcauto-reactive CD8+ T cells via MitAP, causing autoimmunity in DA neurons [44]. Second, E. coli can disrupt the distribution of tight junction proteins and actin flaments in intestinal cells, thereby increasing intestinal permeability, disrupting intestinal barrier function [45,46], and inducing neuroinfammation. In addition, curli, a cell surface amyloid protein abundantly expressed in E. coli, promotes α-Syn aggregation and increases activation of α-Syn overexpressing mouse microglia [47]. In addition, clinical studies have shown that the relative abundance of Enterobacteriaceae is positively correlated with the severity of gait difculties [48]. Combined with our results on motor function and DA neurons, it can be hypothesized that curcumin may protect DA neurons and improve motor impairment in MPTP mice by improving the relative abundance of gut microbiota, such as Enterobacteriaceae. More importantly, high doses of curcumin achieve the best results.
We further used KEGG-based functional prediction and found that the function of the mice gut microbiota was quite altered after curcumin administration, such as superpathway of UDP-glucose-derivedO-antigen building blocks biosynthesis, inosine-5′-phosphate biosynthesis III, superpathway of N-acetylneuraminate degradation, cob (II) yrinate a, c-diamide biosynthesis I (early cobalt insertion), and superpathway of glycolysis and Entner-Doudorof and GDP-mannose biosynthesis. Among them, the superpathway of N-acetylneuraminate degradation is one of the most interesting pathways in this study. Our results showed that the superpathway of N-acetylneuraminate degradation was upregulated in the MPTP group and that this pathway was reduced after curcumin administration. N-acetylneuraminate is a nine-carbon sugar found in vertebrates and some bacteria. It has been shown that Nacetylneuraminate inhibits amyloid fbrillation and aggregation of α-Syn [49], while afecting phosphorylation [50]. Besides. N-acetylneuraminate is components of gangliosides, which can scavenge free radicals and have a signifcant promotion efect on nerve regeneration. Gangliosides ameliorate the neurotoxicity of MPTP, activate autophagy, and participate in α-Syn clearance, thereby reducing the accumulation of α-Syn in neuronal cells [51]. Coincidentally, the pathology of PD is characterized by the aberrant accumulation of α-Syn in the nigrostriatal dense part of Lewy's microsomes. Our results suggest that curcumin may downregulate Nacetylneuraminate degradation, reduce the accumulation of α-Syn, and exert antioxidant efects in the treatment of PD. Notably, this pathway was not significantly diferent in the CUR-L group, suggesting that there may be a dose-dependent efect of curcumin on this pathway.
Tere are some limitations of our current study. First, our fecal samples per group of mice were too small and need to be further expanded. Second, because of the low bioavailability of curcumin, it may be able to explore whether its therapeutic efect would continue to increase or remain at a certain level along with further increases in the curcumin dose. Furthermore, curcumin has been shown to inhibit the activation of the PI3K/ Akt/mTOR signaling pathway in PD cell models [52]. Tis foundation allows us to further investigate the mechanism of action of curcumin, which is our area of interest for future research. More in-depth studies are needed to verify our fndings.

Conclusions
Our study shows that curcumin can efectively regulate the changes of gut microbiota, improve the dysfunction of MPTP mice, reduce the neurotoxicity of MPTP in PD mice, and protect DA neurons, and there is some correlation with the dose. Tis preliminary study demonstrates the therapeutic potential of curcumin for PD, providing clues for microbially targeted therapies for PD.

Data Availability
Te data used to support the fndings of this study are included in the article.