Tuina massage may protect the brain in cerebral palsy rats by improving gut bacteria and gut lining

Male Sprague–Dawley rats (7 days old, weighing 20–30 g) were obtained from Shanghai Jihui Laboratory Animal Co., Ltd. (Shanghai, China) and housed at the Laboratory Animal Center of the Shanghai Public Health Clinical Center. The pups were nursed by their dams and maintained under standard laboratory conditions: temperature 22–27°C, relative humidity 50%–70%, and a 12‐h light/dark cycle, with ad libitum access to standard chow and water. All experimental procedures were performed in accordance with the ethical guidelines approved by the Experimental Animal Ethics Committee of the Shanghai Public Health Clinical Center (approval No. 2022‐A004‐01).

On postnatal day 7 (P7), pups were anesthetized with 5% isoflurane. The left common carotid artery was carefully exposed and ligated. After confirming hemostasis, the skin incision was sutured and disinfected. Following recovery from anesthesia, the pups were placed in a hypoxic chamber at 37°C containing 92% nitrogen (N₂) and 8% oxygen (O₂) for 3 h. After hypoxic exposure, the pups were returned to their dams for regular feeding.

The experimental procedure is illustrated in Figure 1A. The study spanned from P1 to P40. On P6, pups were randomly assigned to three groups: control, model, and Tuina. On P7, pups in the control group underwent a sham operation, whereas those in the model and Tuina groups underwent induction of the CP model. Behavioral assessments included the negative geotaxis test, conducted from P6 to P12, and daily body weight measurements, from P7 to P40. The balance beam test was performed from P36 to P39. On P40, the modified neurological severity score (mNSS) was assessed, and tissue samples were collected for further analysis. Tuina intervention was administered to the Tuina group starting on P8 for 32 consecutive days. Each session lasted 15 min, with a massage frequency of 120 strokes per minute. The intervention followed a specific sequence: (1) spinal manipulation, (2) limb manipulation, and (3) abdominal rubbing. The specific procedures were performed in accordance with previously established protocols (Qiao et al. 2023).

Fecal samples were collected aseptically, immediately flash‐frozen in liquid nitrogen, and stored at −80°C until further analysis. Gut microbiota profiling was performed using 16S rRNA gene sequencing. Total DNA was extracted from fecal samples, and the V3–V4 region of the 16S rRNA gene was amplified using specific primers. The PCR products were purified, quantified, and sequenced on the Illumina MiSeq platform using paired‐end reads. Sequencing data were processed and analyzed using QIIME2 software, including quality filtering, operational taxonomic unit clustering, and diversity analysis. Taxonomic annotation was performed against the Greengenes database.

The negative geotaxis test was performed on rats in all three groups on P6‒P12. Each rat was placed in a head‐down position on an inclined plane set at a 45° angle, and the time required for the rat to reorient from the head‐down to the head‐up position was recorded.

The balance beam test was conducted to evaluate motor coordination and balance in young rats. A wooden beam measuring 1 cm in width and 50 cm in length, elevated 30 cm above the surface, was used. Each rat was placed at one end of the beam and encouraged to traverse to the opposite end. The time required to cross the beam and the number of slips (defined as any instance of a paw slipping off the beam) were recorded. Prior to the formal test, rats underwent a 3‐day training period (P36–P38) to acclimate to the apparatus. The formal test was conducted on the fourth day (P39), during which each rat completed three trials at 30‐min intervals. The average crossing time and the mean number of slips were calculated for each rat as indices of motor coordination and balance.

On P40, an mNSS assessment was performed to evaluate neurological function. The mNSS criteria are summarized in Table, with higher scores indicating more severe neurological deficits. S1

A 100‐µL plasma sample was collected from each experimental group. Triethylammonium bicarbonate (TEAB, 4.5 µL, 1 M) and 8 M urea buffer solutions were prepared, and alkylation was performed by adding iodoacetamide (5 µL, 375 mM) to each sample. The samples were then washed with acetone (660 µL) and incubated overnight at 37°C. Subsequently, samples from each group were labeled using the iTRAQ kit. The labeled samples were mixed in equal volumes, desalted, and lyophilized. The dried samples were reconstituted in a pH 10 solution containing 2% acetonitrile and centrifuged. High‐performance liquid chromatography separation was performed using a C18 column maintained at 45°C. After vacuum drying, the eluates were dissolved in buffer and fractionated using the Dionex NCS3500 system, followed by analysis via nano‐electrospray ionization tandem mass spectrometry (nano‐ESI–MS/MS).

Total protein was extracted from the samples, and the concentration was determined and normalized accordingly. Equal amounts of protein were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and subsequently transferred onto polyvinylidene difluoride membranes. The membranes were blocked at room temperature for 1 h in TBST buffer containing 5% nonfat milk. Following blocking, the membranes were incubated overnight at 4°C with specific primary antibodies: Nrf2 (1:2000, Abcam, ab31163), BDNF (1:2000, Abcam, ab108319), GPX4 (1:2000, Abcam, ab125066), ZO‐1 (1:5000, Bioss, ab‐4023R), occludin (1:2000, Bioss, bs‐1001R), and β‐actin (1:2000, CST, 4970S). After three washes with TBST, the membranes were incubated with HRP‐conjugated secondary antibodies (anti‐rabbit, 1:5000, CST, 7074S) at room temperature for 2 h. Protein bands were visualized using an enhanced chemiluminescence kit (Merck Millipore) and imaged with a ChemiDoc MP system (Bio‐Rad). Densitometric analysis was performed using ImageJ software.

Paraffin‐embedded sections were dried at 60°C for 2 h and sequentially treated with xylene, absolute ethanol, graded ethanol, and distilled water. Antigen retrieval was performed by boiling the sections in diluted 50x EDTA solution, allowing them to cool to room temperature, and subsequently washing them with PBS. The sections were blocked with 10% serum for 10 min at room temperature, followed by overnight incubation at 4°C with primary antibodies diluted at ratios of 1:100–1:500 (Nrf2, 1:200, Abcam, ab31163; BDNF, 1:200, Abcam, ab108319; GPX4, 1:200, Abcam, ab125066; ZO‐1, 1:200, Bioss, ab‐34023R; occludin, 1:500, Bioss, bs‐10011R; IL‐6, 1:200, Abcam, ab9324; TNF‐α, 1:200, Abcam, ab6671). The following day, the sections were washed with PBS and incubated with fluorescent secondary antibodies for 45 min. After additional PBS washes, the sections were counterstained with DAPI for nuclear visualization and imaged using a confocal microscope.

After drying the paraffin sections at 60°C for 2 h, they were sequentially treated with xylene, absolute ethanol, graded ethanol, and distilled water. For hematoxylin and eosin (H&E) staining, the sections were immersed in hematoxylin for 15 min, differentiated in acid alcohol, blued in weak ammonia water, and stained with 0.5% aqueous eosin for 10 min. Subsequently, the sections were dehydrated through a graded ethanol series, cleared in xylene, and mounted with neutral resin. For Luxol fast blue (LFB) staining, the sections were incubated in LFB solution for 20 h, rinsed with 95% ethanol to remove excess dye, differentiated with lithium carbonate, counterstained with eosin, dehydrated through graded ethanol, cleared in xylene, and mounted with neutral resin. Finally, all stained sections were examined and photographed under a microscope for analysis.

Serum samples were analyzed using ELISA kits purchased from Shanghai Enzyme‐linked Biotechnology Co., Ltd. The following kits were used in this study: Rat Lipopolysaccharide (LPS) ELISA Kit (Catalog No. ml003197V), Rat Diamine Oxidase (DAO) ELISA Detection Kit (Catalog No. ml003420V), Rat D‐Lactic Acid (D‐LA) ELISA Kit (Catalog No. ml273710V), Rat Superoxide Dismutase (SOD) ELISA Kit (Catalog No. ml077379V), Rat Malondialdehyde (MDA) ELISA Kit (Catalog No. ml077384V), and Rat Glutathione Peroxidase (GSH‐Px) ELISA Kit (Catalog No. ml097316V).

Statistical analyses were performed using SPSS version 25.0 (Chicago, IL, USA). Data are presented as mean ± standard deviation (SD). The normality of data distribution was confirmed using the Shapiro‒Wilk test, and the homogeneity of variances was verified using Levene's test. Group differences were analyzed using one‐way analysis of variance (ANOVA), followed by post hoc tests. Specifically, when variances were homogeneous, the Least Significant Difference test was applied; otherwise, Dunnett's T3 test was used. A significance level of α = 0.05 was used, and p < 0.05 was considered statistically significant. Graphs were generated using GraphPad Prism 9.5.

To evaluate the effects of Tuina on the growth and development of CP rats, body weight was monitored across all groups. Although body weight gradually increased in each group, rats in the model group exhibited a significantly slower rate of weight gain than those in the control and Tuina groups, with significant differences observed from P9 to P39 (Figure 1B). Brain injury was then assessed in CP rats. Gross examination revealed that the control group had an intact and symmetrical bilateral cerebral cortex, whereas the model group exhibited marked damage, including cortical atrophy and liquefaction. The Tuina group showed only mild atrophy without obvious defects (Figure 1G). H&E staining (Figure 1H) showed that cortical neurons in the control group were numerous, were densely arranged, and exhibited normal nuclei with no obvious pathological changes. In contrast, the injured cortex in the model group contained many swollen, atrophic, and ruptured neurons, with irregular cell sizes, increased staining intensity, and abundant cellular debris. Neuronal density was reduced, and the cells were disorganized, with decreased cytoplasmic volume. Following Tuina intervention, brain tissue injury was markedly attenuated, with fewer necrotic neurons and a more organized, compact neuronal arrangement. Similarly, LFB staining revealed that the Tuina group had less demyelination and vacuolation than the model group (Figure 1I). Collectively, these findings indicate that Tuina promotes the growth and development of CP rats and significantly reduces brain tissue injury.

To evaluate the therapeutic effects of Tuina on motor dysfunction, CP rats were subjected to a series of behavioral tests. In the negative geotaxis test (Figure 1C), no significant differences were observed between groups at baseline (P6). After model induction (P7–P10), the model group exhibited significantly longer response times than the control group. Notably, the consistently shorter response times in the Tuina group than in the model group, which did not reach statistical significance at P8 and P9, became statistically significant by P10. Further assessment with the balance beam test revealed significantly shorter crossing times and fewer foot slips in the Tuina group than in the model group (Figure 1D,E). Consistent with these findings, the Tuina group also demonstrated significantly better neurological function scores (Figure 1F). Collectively, these results indicate that Tuina significantly improves motor function in CP rats.

To determine whether Tuina ameliorates gut microbiota dysbiosis in CP rats, we performed 16S rRNA gene sequencing on fecal samples from all groups. β‐Diversity analysis based on principal coordinates analysis (PCoA) revealed distinct clustering of the gut microbiota between the control, model, and Tuina groups (Figure 2A). Furthermore, STAMP and LEfSe analyses identified significant differences in microbial community composition between groups (Figure 2D–G). Collectively, these results indicate that CP induction caused marked microbial dysbiosis, which was substantially ameliorated by Tuina intervention. At the phylum level, the model group showed a characteristic decrease in Firmicutes and an increase in Bacteroidetes, resulting in an altered Firmicutes/Bacteroidetes ratio. Notably, Tuina intervention effectively reversed this dysbiosis (Figure 2B). Analysis of the top 20 genera showed that Lactobacillus was most abundant in the Tuina group, exceeding its levels in both the control and model groups (Figure 2C). STAMP and LEfSe analyses further corroborated these findings, demonstrating that Lactobacillus—including L. gasseri, L. johnsonii, and unclassified Lactobacillus taxa—was significantly enriched in the Tuina group. In contrast, genera, such as Rothia and QAMM01, were specifically enriched in the model group (Figure 2E–G).

Given the established role of gut microbiota‐derived SCFAs in regulating central nervous system function, we hypothesized that the observed microbial shifts may influence SCFA levels. Serum analyses supported this hypothesis: the model group showed significantly reduced concentrations of hexanoic, isobutyric, butyric, and valeric acids. Tuina intervention partially restored hexanoic, isobutyric, and valeric acid levels, although it had only a limited effect on butyric acid (Figure 2H–K). In summary, these findings suggest that Tuina restores gut microbiota homeostasis in CP rats, markedly increases the relative abundance of Lactobacillus, and promotes the production of SCFAs.

We further evaluated the effects of Tuina on intestinal health in CP rats. HE staining revealed pronounced inflammatory lesions in the model group, characterized by disruption of the mucosal architecture, reduced numbers of glands, and goblet cell atrophy. These pathological alterations were markedly ameliorated following Tuina treatment (Figure 3E). In addition, IF staining showed that the fluorescence intensities of the proinflammatory cytokines IL‐6 and TNF‐α in intestinal tissues were significantly lower in the Tuina group than in the model group (Figure 3A,B), a finding further confirmed by quantitative analysis (Figure 3C,D). Collectively, these findings indicate that Tuina effectively attenuates intestinal pathological injury and inflammatory responses in CP rats.

We further evaluated the effects of Tuina on intestinal barrier integrity in CP rats. Western blotting and IF analyses showed that Tuina significantly upregulated the expression of the tight junction proteins ZO‐1 and occludin (Figure 4). In parallel, Tuina treatment significantly reduced plasma levels of the gut permeability markers LPS, DAO, and D‐LA (Figure 5A–C), and attenuated systemic oxidative stress by decreasing MDA levels and increasing the activities of SOD and GSH‐Px (Figure 5D–F). These results indicate that Tuina improves intestinal function in CP rats by strengthening barrier integrity, reducing permeability, and mitigating oxidative stress.

Given that SCFAs are key mediators of gut–brain axis signaling with important effects on the nervous system, we examined cortical molecules associated with SCFA activity, including the proinflammatory cytokines IL‐6 and TNF‐α, neurotrophic factor BDNF, and antioxidant proteins Nrf2 and its downstream effector GPX4. The cortical fluorescence intensities of IL‐6 and TNF‐α were significantly higher in the model group than in the Tuina group (Figure 6). In addition, cortical BDNF expression was markedly reduced in the model group but significantly increased in the Tuina group (Figure 7A,B,F). Western blot analysis further demonstrated that Tuina markedly increased cortical expression of Nrf2 and GPX4 in CP rats (Figure 7A,D,E). Consistent with these results, IF staining confirmed a pronounced upregulation of Nrf2 in the Tuina group (Figure 7C). These findings suggest that SCFAs may mediate the attenuation of neuroinflammation, the regulation of neurotrophic factor expression, and the improvement of oxidative stress in the brain.