Branched-chain amino acids ameliorate CD4⁺ T-cell-associated gut immune inflammation in Parkinson’s disease

Abstract

Previous studies have shown that alterations in the gut microbiota and its derived metabolites, branched-chain amino acids (BCAAs), are correlated with T-cell-associated immune imbalance and Parkinson’s disease (PD). However, the associations among BCAAs, gastrointestinal dysfunction and T-cell-related gut inflammation remain unclear. This study showed that the constipation symptoms in the PD mice persisted after chronic MPTP treatment. An imbalance in the CD4⁺ T-cell subtypes was observed in the colonic lamina propria (cLP), mesenteric lymph nodes (mLNs), and spleen. Metagenomic and metabolomic analyses showed that microbial dysbiosis promoted BCAA degradation rather than biosynthesis, and reduced BCAA levels were confirmed in the serum. BCAA supplementation alleviated constipation symptoms and increased Th1 and Th17 cell infiltration in the cLP, mLNs and spleen were significantly attenuated after BCAA treatment. This study highlights the therapeutic value of BCAAs in mitigating gut immune inflammation-associated constipation symptoms in PD.

Introduction

Parkinson’s disease (PD) is the most common neurodegenerative movement disorder and is characterized by the loss of dopaminergic neurons (DNs) and the deposition of pathologic alpha-synuclein (α-Syn) in the substantia nigra (SN). In addition to motor symptoms, PD patients suffer from a series of nonmotor symptoms, such as autonomic dysregulation, sleep disorders, and cognitive impairment¹. Gastrointestinal (GI) dysfunctions, including sialorrhea, dysphagia, gastroparesis, and constipation^2,3, are among the most prominent nonmotor symptoms in PD patients. It has been reported that constipation, the most prevalent nonmotor symptom that manifests over a decade before the diagnosis of PD⁴, affects up to 80% of patients with PD⁵. In addition, constipation is associated with a greater risk of developing PD⁴ and is positively correlated with PD severity and duration⁶. However, the mechanisms of constipation in the context of PD remain elusive.

In recent years, an increasing number of studies have suggested that GI dysfunction is closely related to changes in the composition of immune cells, such as CD4⁺ T cells, in the gut. CD4⁺ T cells can be divided into different subtypes, including Th1, Th2, Th9, Th17, Th22, Tfh, and regulatory T (Treg) cells⁷. Mounting evidence has revealed the imbalance in Th1, Th17 and Treg cell proportions in the periphery and central nervous system of PD animals and/or patients^8–11, and CD4⁺ T cells can participate in the pathogenesis of PD through direct effects on DNs^12,13 or indirect regulation of immune inflammation¹⁴. Although the presence and effects of T cells have been explored in the brains of PD models, the mechanisms of T-cell-associated gut inflammation are still unclear.

To date, various PD mouse models, including neurotoxic models such as 6-OHDA¹⁵, MPTP¹⁶, rotenone¹⁷, and human α-Syn A53T transgenic mice¹⁸, have demonstrated GI dysfunction. Moreover, it has been suggested that constipation symptoms often coexist with dysbiosis of the intestinal microbiota¹⁹ in PD patients. In fact, gut dysbiosis in PD patients²⁰ and mouse models²¹ have been identified in numerous studies, and the changes in microbiota-derived metabolite levels may vary with disturbances in the gut microbiome²². Recently, gut-derived branched-chain amino acids (BCAAs) have attracted broad attention in the context of neurodegenerative diseases²³. BCAAs (leucine, isoleucine and valine) are essential amino acids that can only be obtained from the diet or synthesized by the gut microbiota²⁴, which plays an important role in regulating intestinal immunity and maintaining intestinal barrier function^25,26. As previously reported, BCAAs exert neuroprotective and anti-inflammatory effects on the brain and markedly alleviate constipation symptoms in rotenone-induced PD mice²⁷. Additionally, BCAAs can also serve as important regulators of Th1, Th17 and Treg cell differentiation and function in vitro and in vivo²⁸. On the basis of this evidence, we speculate that BCAAs are likely to attenuate GI dysfunction through regulating T-cell-related gut immune imbalance in PD.

In this study, we used MPTP-induced chronic PD mice to investigate GI dysfunction and gut immune imbalance. The gut dysbiosis and altered microbiota-derived BCAA profiles in chronic PD mice were assessed by metagenomic and metabolomic analyses. In addition, to further determine the role of BCAAs in GI dysfunction, a BCAA-rich diet was applied to evaluate the effect on GI immune imbalance to identify a promising strategy to treat GI dysfunction in PD patients.

Results

Time-dependent motor impairment and constipation symptoms were observed in the chronic PD mouse model

To evaluate GI dysfunction in PD, we established a chronic PD mouse model and collected feces on the day before model induction and 3, 7 and 14 days after model induction (Fig. 1A). Moreover, we evaluated motor impairment in the chronic PD model mice. In the OFT, the total distance traveled by MPTP-treated mice did not differ from that traveled by control mice, whereas a shorter distance traveled in the center was observed in MPTP-treated mice, with the most obvious effect occurring on the 7th day after model induction (Fig. 1B, C). A decreased duration in the rotarod test and extended time to descend in the pole test were observed (Fig. 1D, E). Moreover, IF staining and WB also revealed decreased TH expression in the SN of MPTP-treated mice, especially on the 7th day (Fig. 1F–H). These data suggested the successful establishment of the model and indicated that the most obvious difference in motor symptoms appeared on the 7th day after model induction.

Fig. 1. Chronic PD mice presented time-dependent motor impairment and constipation symptoms.

A Experimental design and timepoints of MPTP-induced chronic PD mice. B, C Total distance traveled or distance traveled in the central area in the OFT. D Latency time in the rotarod test. E Time to descend in the pole test. F The expression level of TH in the SN of different groups of PD mice was assessed by WB. G, H DNs in the substantia nigra pars compacta (SNpc) were identified by IF staining and quantified by ImageJ. I–K After 1 h of fecal collection, the fecal pellets were quantified (I), and the water content (J) and wet weight (K) were determined for the different groups. L The colon length of PD mice was measured and compared among different timepoints. *P < 0.05; **P < 0.01; ***P < 0.001. n = 5 in each group. Scale bar = 200 μm.

After fecal collection, we found that the fecal pellets, water content and wet weight of the feces were significantly lower in the MPTP-treated group than in the control group on the 7th day after model induction (Fig. 1I–K). In addition, the colon length was shorter in MPTP-treated mice than in control mice (Fig. 1L). In line with the motor impairment results, these results suggested that MPTP-treated mice developed significant constipation symptoms and colonic pathology, with the MPTP (7 d) group experiencing the most obvious intestinal alterations.

Compared with WT mice, the higher expression level of α-Syn and decreased level of TH in the SN of A53T mice revealed the successful establishment of the transgenic PD mouse model (Fig. S1A). Meanwhile, we discovered that the A53T mice also exhibited lower fecal water content (Fig. S1B) than the WT mice, implying that the constipation symptoms were present in the transgenic PD mouse model, which was consistent with previous reports^18,29.

Gut immune imbalance and mucosal barrier impairment occurred in the chronic mouse model of PD

The Th17/Treg axis balance is pivotal for maintaining intestinal homeostasis and plays a key role in the pathogenesis of GI dysfunction³⁰. Therefore, we aimed to assess the percentages of Th1, Th17, and Treg cells in the spleen, mLNs, and cLP of chronic PD mice. The results revealed that Th1 and Th17 cell proportions were consistently increased in the gut and spleen, whereas Treg cells presented no evidence of change, accompanied by a proinflammatory bias of the Th1/Treg axis and Th17/Treg axis towards Th1 and Th17 cells, respectively (Fig. 2A–L).

Fig. 2. Chronic PD mice exhibited Th17/Treg-associated gut immune dysregulation and inflammation in a time-dependent manner.

A–D The proportions of Th1 (A), Th17 (B), and Treg (C) cells in CD4⁺ T cells were determined by flow cytometry in the spleen of PD mice, and the Th1/Treg and Th17/Treg ratios were calculated (D). E–H The proportions of Th1 (E), Th17 (F), and Treg (G) cells in the mLNs of PD mice and changes in the Th1/Treg and Th17/Treg axes (H). I–L The proportions of Th1 (I), Th17 (J), and Treg (K) cells in the cLP of PD mice and changes in the Th1/Treg and Th17/Treg ratios (L). M The expression levels of ZO-1 and Occludin in the different groups were evaluated by WB. N, O The protein levels of proinflammatory cytokines (IL-1β, IL-6, and TNF-α) were assessed by WB (N), and the relative mRNA levels of these cytokines were determined by qPCR (O). *P < 0.05; **P < 0.01; ***P < 0.001. n = 5 in each group.

PD mice are characterized by intestinal barrier disruption and gut inflammation²¹. We further explored intestinal barrier integrity and gut inflammation in MPTP-induced chronic PD mice. The expression of the colonic tight junction proteins ZO-1 and Occludin decreased in MPTP-treated mice, especially at 7 days after model induction (Fig. 2M). WB and qPCR analyses also revealed that the expression of IL-1β, IL-6, and TNF-α was elevated in the colons of MPTP-treated mice (Fig. 2N, O). Similarly, we also validated the decreased expression of ZO-1 and Occludin and increased levels of IL-1β, IL-6 and TNF-α in the colon of the transgenic A53T PD mice, as shown in Fig. S1C, D. These findings suggested that the chronic PD mice were characterized by a compromised gut barrier and increased colonic inflammation.

Gut dysbiosis and BCAA dysregulation in the chronic PD mouse model

Gut microbiome alterations have been extensively reported in PD animals³¹ and patients³². We also collected feces from PBS- and MPTP-treated mice on day 7 after model induction and subjected them to metagenomic sequencing analysis to determine microbial differences between the two groups. Microbial genes were decreased in MPTP-treated mice compared with control mice (Fig. 3A). Next, the decreased Chao1 and Shannon indices of the α-diversity analysis demonstrated that microbial community richness was also compromised after MPTP treatment (Fig. 3B, C). Moreover, PCoA based on Bray-Curtis distances revealed that the microbial species distributions of the two groups were significantly distinct from each other, as their community structural similarity was low (Anosim R value = 0.6907, P value = 0.0033, Fig. 3D). The α-diversity and β-diversity analyses indicated that the MPTP group was characterized by a lower microbial richness and was different from the PBS group in microbial composition.

Fig. 3. Dysbiosis in chronic PD mice was associated with BCAA dysregulation.

A Venn diagram showing the number of microbial genes identified in the PBS and MPTP groups. B, C The Chao1 index and Shannon index revealed altered α-diversity of the microbiota in the MPTP group compared with the PBS group. D PCoA and Anosim, which are based on Bray-Curtis distances of the microbiota, revealed significant compositional differences at the species level between the PBS group and the MPTP group. E Distribution of the top 10 microbial classes in the PBS and MPTP groups on the basis of their relative abundance. F Heatmap of the relative abundance of the 30 most abundant species in each sample. G LEfSe analysis revealed differential taxonomies in the PBS and MPTP groups (LDA scores >4). H KEGG database annotation of microbial genes revealed differentially abundant pathways between the PBS and MPTP groups via LEfSe analysis (LDA scores >2). I Comparison of the relative abundance of BCAA degradation and biosynthesis pathways in the PBS and MPTP groups. k kingdom, p phylum, c class, o order, f family, s species. **P < 0.01; ***P < 0.001. n = 6 in each group.

To further reveal microbial compositional differences, we identified the top ten classes in the two groups and found that the relative abundances of the Bacteroidia and Verrucomicrobiae classes were significantly increased in the MPTP group, whereas the abundances of the Deferribacteres, Clostridia, and Deltaproteobacteria classes were increased in the PBS group (Fig. 3E and Fig. S2). The heatmap demonstrated the top 30 microbial species on the basis of their relative abundance in each sample, and we discovered that species of Akkermansia muciniphila, Muribaculaceae, Barnesiella, Duncaniella, and Bacteroidales were relatively more abundant in the MPTP group; whereas the PBS group was enriched in Lachnospiraceae, Mucispirillum schaedleri, Oscillibacter, Acetatifactor muris, and Clostridium species (Fig. 3F). Notably, in the MPTP group, the Bacteroidota and Verrucomicrobia phyla were enriched, as mentioned above; however, the Firmicutes and Deferribacteres phyla were enriched in the PBS group (Fig. 3F). Specifically, the Lachnospiraceae family, Eubacteriales order, Clostridia class, Firmicutes phylum and Bacteria kingdom were the taxa with the highest rankings in the PBS group, whereas the Muribaculaceae family, Bacteroidales order, Bacteroidia class and Bacteroidota phylum were the taxa with the highest rankings in the MPTP group according to the LDA scores (all with scores greater than 4) in the LEfSe analysis (Fig. 3G).

The functional annotation of the KEGG database indicated that the microbiota in the MPTP group were associated with metabolism-associated pathways, including amino acid metabolism, and involved neurodegenerative disease-related unigenes (Fig. 3H). More importantly, relative abundance analysis of BCAA-associated pathways suggested that the MPTP group was higher in BCAA degradation but lower in BCAA biosynthesis pathways (Fig. 3I). This interesting discovery suggested that dysbiosis might contribute to BCAA imbalance and that alterations in BCAA levels could be involved in the pathogenesis of gut abnormalities in PD mice.

Combined untargeted metabolomics of feces and targeted metabolomics of serum revealed perturbations in BCAA levels in chronic PD mice

To determine the relative changes in BCAAs in feces, untargeted metabolomics was performed with the LC-MS platform. The OPLS-DA score plot revealed that the metabolic profiles of the PBS group and MPTP group were distinctively separated (Fig. 4A). The OPLS-DA model was validated in 200 permutation tests, with R² X value = 0.311, R² Y value = 0.999, Q² value = 0.856, and P value <0.005 (Fig. S3A).

Fig. 4. Combined metabolomics revealed perturbations in BCAA levels in chronic PD mice.

A The OPLS-DA score plot demonstrated differences in the fecal metabolomic profiles of the PBS and MPTP groups. B Raw intensities of three differentially abundant BCAA-associated oligopeptides in the PBS and MPTP groups. C Fold changes in the top ten upregulated (orange) and downregulated (green) fecal metabolites in the MPTP group compared with the PBS group. D The relative contents of fecal amino acids and their metabolites in the PBS and MPTP groups were determined by Z-score standardization and were presented in the heatmap. E Heatmap showing the correlation of BCAA-related oligopeptides with microbial species. F PCA of targeted amino acids and their metabolites in the serum of the PBS and MPTP groups. G Fold changes in the top ten upregulated and downregulated amino acids and their metabolites in the serum of the MPTP group. H Comparison of the serum levels of leucine, isoleucine and valine in the PBS and MPTP groups. *P < 0.05; **P < 0.01. n = 6 in each group.

Among the differentially abundant metabolites between the two groups, BCAA-associated oligopeptides were detected, and some of them, including Val-Val-Val, Ile-Ile-Val and Leu-Leu, were reported to be elevated in the feces of the MPTP group (Fig. 4B). To better elucidate the overall changes in differential fecal metabolites, we listed the top ten dysregulated metabolites in the MPTP group. As shown in Fig. 4C, the BCAA-related oligopeptide His-Lys-Leu-Val-Val had the second-highest increase in abundance in the MPTP group compared with the PBS group. In addition, according to the VIP scores in the OPLS-DA model, we acquired the top metabolites between groups (VIP value >1, P value <0.05) and found that the BCAA-related oligopeptides Phe-Ile-Asp-Leu-Asn and His-Lys-Leu-Val-Val were the top metabolites with increased abundance in the MPTP group (Fig. S3B). The heatmap of differentially abundant metabolites also highlighted a significant increase in amino acids and their metabolites in the MPTP group compared with those in the PBS group (Fig. 4D). These results suggested that fecal BCAA levels were likely elevated after MPTP treatment.

Next, Spearman correlation analysis was performed between differentially abundant metabolites and differential microbial taxa to determine their relationships, and a heatmap of BCAA-related oligopeptides and associated microbial species with significant differences was generated (Fig. 4E). Consistent with our findings from the metagenomic analysis, the increased abundances of the four BCAA-related oligopeptides were positively correlated with the increased abundances of Muribaculaceae, Bacteroidales, Barnesiella, and Duncaniella species (the Bacteroidota phylum), and negatively correlated with the decreased abundances of Lachnospiraceae, Clostridium, Oscillibacter and Mucispirillum schaedleri species (the Firmicutes and Deferribacteres phyla) in the MPTP group (Fig. 4E).

Interestingly, contradictory results regarding BCAA changes in blood have been reported in PD patients^33,34 in recent years. To further determine the exact change in BCAAs in the MPTP group, we applied targeted metabolomic analysis to quantify 94 amino acids and their metabolites in the serum by the UPLC-MS/MS platform. First, PCA suggested that there was good separation of the metabolic distribution between the PBS and MPTP groups (Fig. 4F). Second, the top ten increased and decreased amino acids and their metabolites in serum were ranked on the basis of their fold changes in the MPTP group vs the PBS group, with isoleucine ranking as one of the most decreased metabolites (Fig. 4G). Next, we paid special attention to BCAA levels in the serum, and found that all three individual BCAAs were significantly decreased in the serum after MPTP treatment (Fig. 4H). This discovery corroborated the aforementioned results of microbial differential KEGG pathways of BCAA degradation and biosynthesis, implying a deficiency of BCAAs in the systemic circulation of chronic PD mice.

To verify the alteration of BCAAs in PD mice, we also performed targeted metabolomic sequencing analysis of amino acids in the feces and serum of WT and A53T mice. The results showed that, as we expected, the three individual BCAAs seemed to be increased in the feces of A53T mice, although no statistical significance (Fig. S4A). Meanwhile, the three BCAAs were all significantly decreased in the serum of A53T mice (Fig. S4B), which was consistent with our findings of the MPTP-induced PD mouse model.

Supplementation with BCAAs alleviated motor disability and gut dysfunction in chronic PD mice

To identify the role of BCAAs in PD, the PBS-treated mice were fed the Normal BCAAs diet as the control group (PBS), and the MPTP-induced PD mice were fed the Normal BCAAs diet (MPTP) or High BCAAs diet (MPTP + High BCAAs) during the same period. We first identified that the dietary supplementation of BCAAs increased BCAA levels in the colon and serum of the PBS-injected mice (Fig. S5A, B), but it had no significant effect on the brain or the gut of these mice (Fig. S5C–E). Next, we evaluated the influence of BCAAs on the MPTP-induced PD mice. As shown in Fig. S5A, B, the High BCAAs diet successfully increased the levels of BCAAs in the colon and serum of MPTP mice. According to our aforementioned results, behavioral tests and fecal collection were conducted on day 7 after model induction, while sacrifice was completed on day 3 and day 7 to determine the T-cell percentages in the gut and spleen (Fig. 5A). In the OFT, the MPTP + High BCAAs group presented a significant increase in the distance traveled by the MPTP group, although no difference was detected in terms of distance traveled in the center (Fig. 5B, C). In addition, the High BCAAs diet prolonged the retention time in the rotarod test and increased the speed of pole climbing in the MPTP group, indicating great improvement in motor ability after BCAA supplementation (Fig. 5D, E). WB analysis revealed that the expression of TH in the SN was significantly increased in the MPTP + High BCAAs group (Fig. 5F), and IF staining of the brain also revealed a greater number of TH⁺ neurons in the SN (Fig. 5G) in the MPTP + High BCAAs group than in the MPTP group. These results suggested that motor function impairment and neurodegeneration were partially reversed in MPTP-treated mice fed the High BCAAs diet.

Fig. 5. Supplementation with BCAAs alleviated motor disability and gut dysfunction in a chronic PD mouse model.

A Experimental design of the BCAA diet intervention in chronic PD model mice and the timepoints of the behavioral test, fecal collection and sacrifice. B, C Motor trajectories of the mice in the OFT, total distance traveled, and the distance traveled in the central area were monitored. D Latency time in the rotarod test. E Time to descend in the pole test. F The expression level of TH in the mouse SN was examined by WB. G IF staining of DNs in the SNpc of the mice, and the number of TH⁺ neurons was compared. H–J Fecal pellets (H), water content (I), and wet weight (J) were determined after 1 h of fecal collection. K The colon length of the three groups of mice was measured and compared. *P < 0.05; **P < 0.01; ***P < 0.001; ns not significant. n = 5 in each group. Scale bar = 200 μm.

On the other hand, the fecal output results revealed that although wet weight remained unchanged among the three groups, the High BCAAs diet significantly increased the number of fecal pellets at 1 h and the fecal water contents of PD mice (Fig. 5H–J). The colon lengths of the MPTP-treated mice increased after they were fed the High BCAAs diet (Fig. 5K). These findings indicated that MPTP-treated mice benefited from high BCAA consumption in terms of alleviating constipation symptoms and gut pathology.

BCAA treatment ameliorated the gut Th17/Treg imbalance and intestinal inflammation in the chronic PD mouse model

Flow cytometry of splenocytes revealed that the High BCAAs diet significantly reduced the proportions of proinflammatory Th1 and Th17 cells in CD4⁺ T cells and ameliorated the splenic Th1/Treg and Th17/Treg axis bias of the MPTP group on day 7 after model induction (Fig. 6A–D). Moreover, on day 3 after model induction, the MPTP + High BCAAs group presented lower Th1 and Th17 cell percentages in mLNs than the MPTP group did, accompanied by recovery of biased Th1/Treg and Th17/Treg axes, although the Treg cell ratio did not change (Fig. 6E–H). Similarly, we also found that the cLP of the MPTP + High BCAAs group had less infiltration of proinflammatory Th1 and Th17 cells than the MPTP group did. Notably, the proportion of anti-inflammatory Treg cells was greatly reduced after MPTP treatment and seemed not to be influenced by the High BCAAs diet. However, the biased Th1/Treg and Th17/Treg axes were ameliorated after BCAA treatment (Fig. 6I–L) on day 3 after model induction. The combined results demonstrated that BCAA supplementation could alleviate the bias of the CD4⁺ T-cell-related immune axis in gut-associated tissues and the spleen.

Fig. 6. BCAA treatment ameliorated the gut Th17/Treg imbalance and intestinal inflammation in a chronic PD mouse model.

A–D The proportions of Th1 (A), Th17 (B), and Treg (C) cells in CD4⁺ T cells in the spleens of the PBS, MPTP and MPTP + high BCAAs groups were determined by flow cytometry. The altered Th1/Treg and Th17/Treg axes in the spleen (D). E–H Percentages of Th1 (E), Th17 (F), and Treg (G) cells in CD4⁺ T cells in the mLNs of the three groups of mice. Changes in the Th1/Treg and Th17/Treg ratios in the mLNs (H). I–L Changes in the proportions of Th1 (I), Th17 (J), and Treg (K) cells in CD4⁺ T cells in the cLP of the three groups of mice. Changes in the Th1/Treg and Th17/Treg axes (L). M The expression levels of ZO-1 and Occludin were assessed by WB. N, O The expression levels of colonic proinflammatory cytokines (IL-1β, IL-6, and TNF-α) were examined by WB (N), and their relative mRNA levels were determined by qPCR (O) in the three groups. *P < 0.05; **P < 0.01; ***P < 0.001; ns not significant. n = 5 in each group.

Next, with respect to gut barrier integrity, the expression of the gut barrier proteins ZO-1 and Occludin was significantly increased after the MPTP mice were fed the High BCAAs diet (Fig. 6M). Moreover, the protein and mRNA levels of the proinflammatory cytokines IL-1β, IL-6, and TNF-α were also lower in the colonic tissue of the MPTP + High BCAAs group than in that of the MPTP group (Fig. 6N, O), indicating reduced colonic inflammation levels after BCAA supplementation. These results demonstrated that the High BCAAs diet could ameliorate the immune imbalance of the intestinal and systemic Th1/Treg and Th17/Treg axes and mitigate colonic inflammation in MPTP-induced PD mice.

Discussion

The gut microbiome, metabolites, and T-cell-associated immune imbalance may be involved in the pathogenesis of PD; however, the mechanism is still unclear. In this study, we found that the gut immune imbalance in chronic PD model mice was associated with an imbalance in CD4⁺ T-cell subtypes. Additionally, dietary supplementation with BCAAs effectively alleviated neurodegeneration, constipation symptoms and the gut immune imbalance in PD mice. For the first time, we provide new evidence that BCAAs regulate gut dysfunction and immune imbalance in PD.

Alterations in CD4⁺ T-cell subsets are involved in the pathogenesis of PD, among which the imbalance in the Th17/Treg axis plays an important role^8,11,14. We observed a significant increase in the proportions of Th1 and Th17 cells in the spleen, mLNs and cLP of chronic PD model mice, and the Th17/Treg balance shifted towards Th17 cells after model induction, which was in accordance with the results of our previous study⁸. Similarly, elevated proportions of splenic Th1 and Th17 cells were also reported 7 days after acute MPTP administration, but no changes in the proportion of Treg cells were detected³⁵, which was consistent with our findings. Interestingly, we noticed that the increased infiltration of Th17 cells and the imbalanced Th17/Treg axis in the gut were more prominent than those of Th1 cells were, which suggested that Th17 cells might be a dominant change in the CD4⁺ T-cell-related imbalance in the gut immune system in PD mice. A previous study³⁶ revealed that PD patients with constipation (especially those with slow-transit constipation) had increased frequencies of Th17 cells in the periphery. Moreover, experimental autoimmune encephalomyelitis (EAE) mice with constipation presented increased percentages of Th17 cells, reduced percentages of Treg cells and increased Th17/Treg ratios in the spleen and inguinal lymph nodes³⁷. Another example is that autism spectrum disorder (ASD) children with GI symptoms have higher frequencies of Th17 populations in the periphery than do ASD children without GI symptoms³⁸. These results suggested the involvement of the Th17/Treg axis imbalance in the neurological disease-associated constipation symptoms.

Accompanied by gut immune dysregulation, increased local inflammation levels and damage to the intestinal barrier were induced in colonic tissue by chronic MPTP treatment. Numerous studies have reported that colonic inflammation, gut leakage and elevated LPS or LBP levels are associated with disruption of the gut barrier in PD mice^21,39–41. Chronic gut inflammation leads to elevated intestinal permeability, allowing excessive passthrough of harmful microorganisms, toxic substances and enteric metabolites, which induce a systematic inflammatory cascade, destroy the integrity of the blood-brain barrier (BBB), and eventually aggravate neurodegeneration as has been previously reviewed in detail^2,42,43.

In recent years, an increasing number of studies have explored microbial community alterations in PD. The decreased richness and distinctive profiles of the gut microbiota in MPTP-induced PD mice have been repeatedly documented^40,44. However, the results on differential taxa identified by fecal metagenomics or 16S rRNA sequencing in reports only partially agree with each other and discrepancies remain^20,45. The Bacteroidetes (also known as Bacteroidota) and Firmicutes phyla are prevalent in the gut microbiota of both PD patients and healthy controls⁴⁶. An increase in Bacteroides, and a decrease in Firmicutes and the Firmicutes/Bacteroidetes (F/B) ratio have been found in PD patients compared with healthy subjects⁴⁷. In PD animals, decreased Firmicutes and increased Bacteroidetes abundances were also reported in the feces of MPTP-induced PD model mice^40,44, which was consistent with our results. The F/B ratio is considered an indicator of the gut health status, with an abnormal F/B value associated with a greater risk of disease⁴⁸. Therefore, an imbalanced F/B ratio serves as a potential biomarker of dysbiosis in both PD patients and animals. Additionally, one report demonstrated a decrease in the Lachnospiraceae family from the Firmicutes phylum and its key members in the stool of PD patients, which was consistent with our current findings⁴⁹. Among the decreased taxa in MPTP-induced PD mice, the abundance of anti-inflammatory Lachnospiraceae seemed to decrease in the GI tract of PD patients as the disease progressed, indicating a continuous decrease in beneficial bacteria with the deterioration of PD^50,51. Regarding the increased Akkermansia muciniphila in PD mice⁵², which was also verified in our study, evidence has suggested its multifaceted role in PD pathogenesis, including causing increased gut inflammation and triggering compensatory anti-inflammatory effects in the host^53,54. Recently, its protective role in alleviating neurodegeneration and intestinal permeability in MPTP-induced PD mice has also been reported⁵⁵. Therefore, the increase in Akkermansia muciniphila might be simply due to dysbiosis or serve as a compensatory change after the disruption of gut homeostasis in PD.

BCAAs are among the most important microbiota-derived metabolites that can be obtained from animal-based food and the biosynthesis of the gut microbiota⁵⁶. In the present study, we identified the increased degradation and decreased synthesis of BCAAs in MPTP-treated PD mice in comparison to PBS-treated mice, indicating decreased production of BCAAs in PD mice. Moreover, we also detected significant correlations between the abundances of oligopeptides containing BCAAs and those of differential microbial species, which revealed that the abundances of BCAA-related metabolites were positively related to that of Bacteroidota and negatively associated with that of Firmicutes. This relationship suggested that the increase in fecal BCAA-related oligopeptides might be associated with the perturbation of specific microbiota in the gut of PD mice, which provides directions for further investigation of certain microbial taxa. However, there is currently no unified conclusion on the changes in the content of BCAAs in feces. A previous study³⁴ indicated that, compared with early-stage PD patients, advanced PD patients had a reduced fecal abundance of genes related to BCAA biosynthesis and a lower number of predicted genes associated with BCAA biosynthesis. Although this observation partially coincided with our discovery in terms of reduced fecal gene numbers of BCAA biosynthesis, the lower concentration of fecal BCAAs in advanced PD patients, as reported, was not identified in our study. However, elevated levels of fecal isoleucine and leucine have also been detected in PD patients compared with control subjects⁴⁹. In PD animals, reduced microbial BCAA biosynthesis and fecal BCAA levels were demonstrated in the stool of rotenone-induced PD mice, the latter of which was the opposite of the results of the current study¹⁷. A few factors may result in contradictory results for fecal BCAAs, such as a limited sample size, dynamic changes in the biosynthesis and degradation of amino acids, and variations in enrollment criteria, the detection methodologies and adjusting factors⁵⁷. Although the changes of BCAAs in the serum and feces of the MPTP model have been validated in the A53T mice in our study, more clinical samples and studies are still needed for verification.

BCAAs have been reported to ameliorate neurodegeneration, regulate immunity and mitigate gut dysfunction as mentioned previously^25–27,58. As expected, BCAA supplementation greatly alleviated motor impairment, neurodegeneration and GI dysfunction in MPTP-treated PD mice. More importantly, the Th1/Treg and Th17/Treg imbalances were also attenuated after high BCAA intake, along with decreased colonic inflammation and reduced gut barrier damage. Previous studies have revealed the role of BCAAs in regulating the proliferation, differentiation and function of T cells, especially Th1, Th17, and Treg cells²⁸. This involved activation of mammalian target of rapamycin complex 1 (mTORC1) by BCAAs in T cells, and mTORC1 was significant in integrating environmental cues, including amino acids, sensing cellular energy state and regulating cell biology⁵⁹. It was reported that deficits in Slc7a5 and Slc1a5 would lead to a reduction of BCAAs uptake, thus inhibiting T cells’ differentiation into Th1 and Th17 in an mTORC1-dependent manner^60,61. However, leucine supplementation also exhibited adverse effects, as it inhibited the mTORC1 pathway and downregulated Th1 and Th17 cells⁶². Additionally, leucine and isoleucine had been proposed as important regulators in Treg cells’ functions and proliferation^63,64, although mTORC1’s role seemed to be bidirectional. It was reported that mTORC1 inhibition would lead to Treg differentiation other than Th1 or Th17 cells⁶⁵; however, mTORC1 deficits would also cause weakened immune-suppressing functions of Tregs⁶⁶. Therefore, different BCAA molecules may exert their influence on T cells in a complicated and interactive manner. In the present study, the influence of BCAAs on immune balance in vivo might be a result of the combined effects of leucine, isoleucine and valine on PD mice. As a previous study demonstrated, BCAAs mainly exert protective effects on PD pathogenesis and might serve as potential therapeutic strategies for PD treatment in the future²⁷. Overall, this is the first study to identify the role of BCAAs in regulating gut immune inflammation-related constipation symptoms in PD model mice.

There are also several limitations in our study. First, although BCAA supplementation alleviated Th1-, Th17- and Treg-associated peripheral immune imbalances in chronic PD mice, the specific mechanisms involved remain unknown. Second, the beneficial effects of BCAAs on PD still need to be verified in clinical studies. Given that PD patients usually suffer from constipation in the prodromal phase of the disease, developing a nutritional supplement for mitigating GI dysfunction is promising. We will continue to explore the above limitations in our future work.

Collectively, our study revealed time-dependent patterns of neurodegeneration, GI dysfunction and peripheral immune imbalance in an MPTP-induced chronic PD model. Gut dysbiosis-related BCAAs were dysregulated in feces, whereas a significant reduction in BCAA levels was observed in the serum of PD mice. BCAA supplementation attenuated neurodegeneration, mitigated GI symptoms, restored peripheral T-cell immune balance and reduced gut inflammation in PD mice. This study highlights the role of gut microbiota-derived metabolites in PD pathogenesis and proposes promising therapeutic strategies for mitigating gut dysfunction in PD.

Methods

Animals

Male C57BL/6 J mice (5–8 weeks of age, 20–25 g) and A53T mice (12 months old) were purchased from GemPharmatech Co., Ltd, Guangdong, China. All the mice were housed in a specific pathogen-free animal facility of Tongji Hospital under a 12/12 h light/dark cycle and had ad libitum access to food and water. All the animal experiments were approved by the Animal Welfare Ethics Committee of Tongji Hospital (IACUC Issue No. TJH-202311034) and were conducted in compliance with the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 86-23, 1985). Sample sizes per group were determined to ensure reproducibility and minimize animal use in compliance with institutional animal care guidelines. Five to six male mice were randomly allocated per group, stratified by age to maintain homogeneity. All experiments were double-blinded to reduce bias. Mice were anesthetized with isoflurane for blood collection, and for tissue collection, they were euthanized by cervical dislocation after anesthesia.

In vivo PD mouse model and BCAA intervention

A chronic PD mouse model was established by the intraperitoneal injection of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP, 25 mg/kg, Sigma, USA), which was dissolved in phosphate-buffered saline (PBS, 5 mg/mL). The C57BL/6J mice received MPTP injection twice a week for 5 consecutive weeks⁶⁷. Behavioral tests and fecal sample collection were performed at different timepoints (the day before MPTP model induction, and 3, 7, and 14 days after model induction) followed by animal sacrifice (Fig. 1A). After the mice were sacrificed, the SN of mice was obtained; the spleen, mesenteric lymph nodes (mLNs), and colonic lamina propria (cLP) were collected to examine alterations in CD4⁺ T-cell subsets.

In addition, the C57BL/6J mice were administered the same volume of PBS intraperitoneally as the control group. The feces and serum of the PBS- and MPTP-treated mice were obtained 7 days after model induction for sequencing. The transgenic A53T mouse model is a well-known PD animal model which replicates the key features of Lewy body pathology in PD patients. The 12-month-old male wild-type (WT) mice and human α-Syn A53T transgenic mouse model of Parkinson’s disease were used for verification^68,69.

The specific diet fodder was customized by Shulaibao (Wuhan) Biotechnology Co., Ltd., according to the 1BCAA and 2BCAA formulas as previously described⁷⁰ (Table S1). The mice received PBS or MPTP administration, and were fed the Normal or High BCAAs diet for 5 weeks during the same period. Behavioral tests, fecal collection and tissue collection were performed 3 or 7 days after model induction, depending on the timepoints of the most obvious changes.

Animal behavioral test

Before the behavioral tests, the mice were preadapted to the test environment for 2 h. All the mice received behavioral training before the formal start of behavioral testing.

Open field test (OFT)

The mouse was placed in the center of the field and monitored for 5 min, during which the entire movement process of the mouse was recorded by the ANY-maze Video Tracking System (v7.16, Stoelting, USA). The movement trajectory was plotted, and the total travel distance and distance in the center were automatically calculated.

Rotarod test

The mice were placed on the rotarod apparatus (IITC Life Science, USA), which was programmed to uniformly accelerate from 4 to 40 rpm within 5 min to test their motor coordination ability. The latency to fall was automatically recorded by the apparatus when the mouse fell off the rotarod, and those that lasted longer than 5 min were recorded as 300 s. All the mice were tested three times at 30-min intervals, and the average time was considered the fall latency in the rotarod test.

Pole test

A vertical wooden pole (50 cm in length and 1 cm in diameter) equipped with a cork on the top was utilized. The time required for the mice to reach the bottom from the top of the pole was recorded as the descending time. Each mouse was tested three times at 30-min intervals, and the average duration was recorded as the time to descend.

Fecal collection

The mice were placed in individual clean cages without food, water or padding at 9:00 in the morning. After 1 h, the fecal pellets from each mouse were immediately collected, counted and sealed in separate sterile 1.5 mL tubes. Next, the fecal pellets were dried at 65 °C overnight and weighed again the next day. The water content was then calculated as (wet weight-dry weight)/wet weight × 100%.

Flow cytometry

Single-cell suspensions of the spleen and mLNs were prepared. Single cells from the cLP were obtained as previously described⁷¹. For surface marker staining, the cells were incubated in PBS containing 1% BSA and a FITC-conjugated rat anti-mouse CD4 antibody (1:200, 100405, Biolegend, USA) for 30 min on ice. The staining of the intracellular markers was performed following the instructions of a Transcription Factor Buffer Set Kit (BD Biosciences, USA), and the antibodies used were as follows: Brilliant Voilet 421 rat anti-mouse IL-17A (1:200, 506926, BioLegend, USA) or PE rat anti-mouse IL-17A (1:200, 561020, BD Biosciences, USA) for Th17 cell staining; PE rat anti-mouse IFN-γ (1:200, 554412, BD Biosciences, USA) or PE-Cy 7 rat anti-mouse IFN-γ (1:200, 561040, BD Biosciences, USA) for Th1 cell staining; and Alexa Fluor 647 rat anti-mouse Foxp3 (1:200, 563486, BD Biosciences, USA) for Treg cell staining. Flow cytometry data were then obtained from Cytoexpert software (Beckman Coulter, USA), and further analysis was performed with FlowJo software (FlowJo LLC, USA). The gating strategies for Th1, Th17 and Treg cells in the spleen, mLNs and cLP of the mice are shown in Fig. S6.

Immunofluorescence (IF) staining analysis

For IF staining, frozen SN slices were incubated at room temperature for 10 min and fixed in 4% paraformaldehyde (PFA). After washing, the tissue area was blocked with a quick block solution (Beyotime, China) and incubated with the primary antibody (anti-tyrosine hydroxylase (TH; rabbit, 1:400; ab137869, Abcam, UK)) at 4 °C overnight. The next day, the tissue was incubated with the secondary antibody (YSFluor 488 goat anti-rabbit lgG (H + L) (1:400; 33106ES60, Yeasen Biotechnology, China)) in the dark at room temperature for 1 h. After washing, the slides were then mounted with Antifade Mounting Medium with DAPI (Beyotime, China) and covered with cover slips. Images were acquired by confocal microscopy (FV1200; Olympus, Japan), and further image analysis was completed by ImageJ software (NIH, USA).

Protein extraction and western blotting (WB)

Tissues (colon and SN) were lysed in RIPA lysis buffer supplemented with protease inhibitor cocktail, phosphatase inhibitor and PMSF (all from Servicebio, China). The concentrations of the samples were quantified with a bicinchoninic acid kit (Beyotime, China). The same amount of protein was loaded on SDS-PAGE gels (Yeasen Biotechnology, China), and the protein was transferred onto nitrocellulose (NC) membranes after electrophoresis. Next, the NC membranes were blocked in 5% skim milk for 1 h at room temperature and then incubated with primary antibodies at 4 °C overnight. The primary antibodies used for WB were as follows: anti-TH (rabbit, 1:1000; ab137869, Abcam, UK), anti-zonula occludens-1 (ZO-1, rabbit, 1:200; 61-7300, Thermo Fisher Scientific, USA), anti-Occludin (rabbit, 1:5000; 27260-1-AP, Proteintech, USA), anti-β-actin (mouse, 1:10,000; 66009-1-Ig, Proteintech, USA), anti-IL-1β (rabbit, 1:1000; 12703, Cell Signaling Technology, USA), anti-IL-1β (rabbit, 1:1000; A1112, ABclonal, China), anti-IL-6 (rabbit, 1:1000; A0286, ABclonal, China) and anti-TNF-α (rabbit, 1:1000; A11534, ABclonal, China). Next, the membranes were incubated with secondary antibodies for 1 h at room temperature. The secondary antibodies used were as follows: HRP-conjugated goat anti-rabbit IgG (1:5000, SA00001-2, Proteintech, USA) and HRP-conjugated goat anti-mouse IgG (1:5000, SA00001-1, Proteintech, USA). The protein bands were visualized with ECL Plus reagents (Servicebio, China) and visualized with a BLT GelView 6000 Pro imaging system (Guangzhou Biolight Biotechnology, China). The relative intensities were measured with ImageJ software (NIH, USA).

Quantitative real-time polymerase chain reaction (qPCR) analysis

Total RNA was extracted from the colonic tissue with RNAiso Plus reagent (Takara, Japan). Reverse transcription of RNA into cDNA was completed with Hifair Ⅲ 1st Strand cDNA Synthesis SuperMix for qPCR (gDNA digester plus) (11141ES60, Yeasen Biotechnology, China), and qPCR amplification was performed with Hieff qPCR SYBR Green Master Mix (No Rox) (11201ES50, Yeasen Biotechnology, China) with a Bio-Rad CFX96 qPCR system (Bio-Rad Laboratories, USA) or a Roche LightCycler 480 Instrument (Roche, Switzerland) according to the manufacturer’s instructions. The sequences of primers used for qPCR are listed in Table S2. Relative mRNA levels were determined by the 2^−ΔΔCT method.

Fecal metagenomic sequencing analysis

Cluster generation was performed with the cBot cluster generation system, and the library preparations were sequenced by the Illumina NovaSeq platform. Gene prediction was performed with MetaGeneMark (v3.38). DIAMOND software was used to compare unigene sequences with those of bacteria, fungi, archaea and viruses from the NCBI NR database. The LCA algorithm of the MEGAN software was used to determine the species annotation messages of sequences.

α-Diversity analysis was performed using the Chao1 and Shannon indices. β-Diversity analysis (principal coordinates analysis (PCoA) and analysis of similarities (Anosim)) was performed and reported using Bray-Curtis distances to reflect the microbial structural similarity of samples in different groups. The most abundant microbial classes of the two groups are shown in a bar plot on the basis of their relative abundance using ggplot2 (v3.3.6). A heatmap of the most abundant species in each sample was also generated to compare the differences in their relative abundances by ComplexHeatmap (v2.12.0). Linear discriminant analysis (LDA) was performed, and the results of the LDA effect size (LEfSe) analysis of the differential microbial taxa are shown (over four LDA scores). For functional annotation, unigenes were compared with the Kyoto Encyclopedia of Genes and Genomes (KEGG) database through the DIAMOND software, and the relative abundance of different functional levels was calculated after filtering the comparison results.

Fecal untargeted metabolomic analysis

The parameters of the mass spectrometry conditions are shown in Table S3. The raw data obtained from liquid chromatography-mass spectrometry (LC-MS) methods were converted to mzXML format by ProteoWizard software. The XCMS program was used for peak extraction, alignment and retention time correction. Metabolite identification was performed by matching the acquired data against multiple databases, including the laboratory’s self-built database, integrated public database, AI database and metDNA.

Orthogonal partial least squares discriminant analysis (OPLS-DA) was performed with the MetaboAnalystR package (v1.0.1), and a score plot was created to demonstrate the differences between groups. Next, the variable importance in projection (VIP) based on the OPLS-DA model (VIP score >1) and P value <0.05 (Student’s t-test) were combined as a cut-off to filter differentially abundant metabolites. A heatmap was created by the ComplexHeatmap package (v2.9.4) after Z-score standardization. For correlation analysis between fecal differential microbial taxa and metabolites, Spearman correlation analysis and significance tests were performed by functions cor and corPvalueStudent from R software, respectively, and a heatmap was created by the ComplexHeatmap R package (v2.9.4).

Targeted metabolomic analysis of amino acids and related metabolites in feces and serum

Venous blood was drawn from the orbital plexuses of MPTP- and PBS-treated mice. An LC-ESI-MS/MS system (UPLC, ExionLC AD; MS/MS, QTRAP 6500+) was utilized for sample extract analysis. The AB 6500 + QTRAP LC-MS/MS System, equipped with an ESI turbo ion-spray interface, was operated in positive and negative ion modes and controlled by Analyst software (v1.6, AB Sciex). The ESI source operation parameters were as follows: ion source, turbo spray; source temperature, 550 °C; ion spray voltage (IS), 5500 V (positive), −4500 V (negative); curtain gas (CUR), 35.0 psi; declustering potential (DP) and collision energy (CE) for individual MRM transitions were optimized. A particular set of MRM transitions was observed in each period on the basis of the amino acids eluted within the duration.

Unsupervised principal component analysis (PCA) was conducted with the statistical function prcomp in R software after unit variance scaling of the data. Significantly changed metabolites between groups were determined by absolute Log₂FC. The top-ranked differential amino acids and their related metabolites are shown according to their FC values.

Feces and serum of the WT and A53T mice were also collected for verification of BCAA changes by targeted metabolomic analysis of amino acids and related metabolites.

Examination of total BCAA concentrations in the colon and serum

The total BCAA concentrations in the colon and serum were examined using the Branched-Chain Amino Acid Assay Kit with WST-8 (Beyotime, China). For sample preparation, 10 mg of colon was lysed in 100 μL BeyoLysis^TM Buffer A for Metabolic Assay. After centrifugation, the supernatant was used for the detection of colonic BCAA concentrations. The blood samples were placed at room temperature for 2 h, and the supernatant was collected after centrifugation. The dilution factors of the colon and serum samples were determined in the preliminary experiment to ensure that the BCAA concentrations fell within the standard curve range. After adding the working solution to the samples, the absorbance at 450 nm was detected. The BCAA concentrations were then identified based on the leucine standard curve, as indicated in the manufacturer’s instructions.

Statistical analysis

The data were presented as the means ± standard errors of the means (SEMs). Unpaired Student’s t-tests were used for comparisons between two groups, and one-way ANOVA was applied for comparisons between more than two groups with post hoc tests using Tukey’s multiple comparisons. A P value <0.05 was considered statistically significant. *P value <0.05, **P value <0.01, ***P value <0.001. All the statistical analyses were performed with GraphPad Prism (v9.5.0).

Author contributions

K.A. and J.L. conceived and designed the whole study; K.A. conducted experiments, completed data analysis, and drafted the manuscript; J.L. supervised the study, reviewed the manuscript and provided intellectual content; Z.X., Z.M., and J.L. provided financial support; D.W., Y.Q., H.Y., and H.L. assisted with the experiments.

Data availability

All raw data generated or analyzed during this study are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41531-026-01375-y.