Yellow cattle nose cartilage was procured from Xi’an Huimin Street (Shanxi, China). Chondroitin sulfate standard (CAS 9007-28-7, purity 95%) was obtained from Macklin Company (Shanghai, China). H₂O₂ and Vc were purchased from Chengdu Kelong Chemical Co., Ltd. (Chengdu, China). Glucan standards (1, 5, 12, 25, 50, 80, 150, and 270 kDa) were obtained from Sigma Aldrich (St. Louis, MO, United States). Lipopolysaccharide (LPS) was purchased from Sigma Aldrich (St. Louis, MO, United States); FeSO₄ and FeCl₃·6H₂O were purchased from Tianjin Damao Chemical Reagent Factory. 2,3,5-Chlorotriphenyltetrazolium (TTC) was obtained from Shanghai Lanji Technology Development Co., Ltd. Enzyme-linked immunosorbent assay (ELISA) kits for IL-1β and TNF-α were purchased from Shanghai Xinle Biotechnology Co., Ltd. (Shanghai, China). Malondialdehyde (MDA) and reduced glutathione (GSH) detection kits were obtained from Nanjing Jiancheng Biotechnology Research Institute. Acetic acid, propionic acid, butyric acid, isobutyric acid, valeric acid, and n-valeric acid standards were purchased from Aladdin Bio-Chem Technology Co., Ltd. (Shanghai, China). BIOZOL reagent was purchased from Hangzhou Bioer Technology Co., Ltd. (Zhejiang, China). Reverse Transcription Kit (UEIris RT mix with DNase) and 2× SYBR Green qPCR Master Mix were purchased from Suzhou Hengyu Biotechnology Co., Ltd. (Jiangsu, China). All other reagents used in this study were manufactured in China and met the standards of high-performance liquid chromatography (HPLC) or the highest commercially available grade.

The meat and impurities attached to the surface of the bovine nasal cartilage were removed, washed, degreased with acetone, and air dried. Thereafter, 10 g of bovine nasal cartilage powder was weighed and suspended in 0.01 M EDTA buffer containing 0.01 M cysteine (in a ratio of 1:10, w/v) with pH adjusted to 6.5. A 1.5% composite enzyme solution (papain: trypsin = 1:2, w/w) was added to the mixture and allowed to react in a 65°C water bath for 22 h. Following centrifugation at 4000 rpm for 10 min, the supernatant was collected and precipitated with 1.25 volume of ethanol at 4°C for 12 h, followed by 2 volumes of ethanol containing 0.5% sodium acetate. The resulting solution was centrifuged at 4000 rpm for 10 min, and the precipitate was dissolved in 200 mL of 0.5 M NaCl solution containing 5% cetylpyridinium chloride (CPC). After centrifugation at 4000 rpm for 10 min, the precipitate was dissolved in 100 mL of 2 M NaCl: ethanol (100:15, v/v) and precipitated with three volumes of ethanol. The above steps were repeated, and the obtained sediment was centrifuged, dissolved in water, and then dialyzed against ultrapure water for 3 days. The solution was concentrated under reduced pressure and freeze-dried to obtain the CS samples.

To obtain components with different molecular weights, degradation conditions were screened using a controlled variable method. Briefly, 50 mg of CS was weighed accurately for each serving, and then suspensions were prepared with a material liquid ratio of 0.5% and transferred into stoppered vials. The concentration of H₂O₂ was maintained at 60 mM and that of Vc was changed from 0.5 mM to 60 mM. Conversely, the concentration of Vc was maintained at 20 mM and the concentration of H₂O₂ was changed from 0.5 mM to 80 mM. After sealing, the solutions were allowed to react for 3 h in a 60°C water bath. Upon reaction completion, the solution was concentrated, dialyzed at 4°C for 3 days using a dialysis bag with a molecular cut-off of 3,500 Da (excluding H₂O₂ and Vc), and freeze-dried to collect the degraded samples.

The carbazole sulfuric acid method was used to determine the glucuronic acid content (20). Using potassium sulfate as the standard, the barium chloride gel method was used to measure the sulfate content (21). Protein content was determined using the BCA kit method (22). The FT-IR spectra (KBr particles) of the polysaccharides (2 mg) were recorded in the range of 400–4,000 cm⁻¹ using a Nicolet 5,700 spectrometer (Thermo Fish Scientific Inc. Waltham, MA, United States). The average molecular weights of CS and its degradation components were determined using HPLC with a TSK Gel G4000SWXL dextran gel chromatographic column, differential refractive detector, mobile phase consisting of phosphate buffer brine with pH 6.0, flow rate of 0.3 mL/min, and injection volume of 20 μL.

The ability of CS and its degradation products to remove ABTS⁺· was determined using the method of Li et al. (23). The scavenging activity of 1-diphenyl-2-picrylhydrazine (DPPḤ) was determined using the Yang et al. method (24). The Fenton method was used to determine the scavenging potential of CS and its degradation products on hydroxyl radicals (19). The abilities of CS and its degradation products to reduce iron ions were measured using the method described by Unver et al. (25).

The 3-month-old C57BL/6 J mice were obtained from Xi’an Jiaotong University (Xi’an, Shaanxi, China). Mice were housed in an animal facility under standard environmental conditions (12/12 h light dark cycle, humidity 50 ± 15%, temperature 22 ± 2°C) and were provided with a standard feed (AIN-93 M). Mice were randomly divided into four groups: CON, LPS, CS-1 + LPS, and CS-2 + LPS (n = 10/group). CS-1 + LPS and CS-2 + LPS group mice were administered CS-1 or CS-2 (200 mg/kg/d, dissolved in physiological saline) via oral gavage over 64 consecutive days. Mice in the LPS, CS-1 + LPS, and CS-2 + LPS groups were intraperitoneally administered LPS (0.25 mg/kg/d, dissolved in physiological saline), while those in the CON group received an equivalent volume of physiological saline for 18 consecutive days. This dosage of LPS was chosen based on previous studies (26, 27). At the end of the experiment, the mice were then placed in an induction chamber and exposed to 4% isoflurane-oxygen mixture until they were unconscious, as confirmed by the lack of a pedal withdrawal reflex. Then, the mice were euthanized through cervical dislocation, after which both serum and brain samples were collected, and stored at −80°C for further analysis. The animal protocol was approved by the Animal Ethics Committee of Northwest A&F University. All experimental procedures were performed in strict accordance with the “Guidelines for the Care and Use of Experimental Animals” (8th edition; ISBN-10:0-309-15396-4).

Mouse brain tissue was fixed in 4% (v/v) paraformaldehyde and embedded in paraffin. Brain slices were cut into 5 mm slices using a slicer. Before staining, sections were dewaxed with xylene and rehydrated with each gradient ethanol (100, 90, 80, and 70%) for 5 min, and washed five times with PBS (pH 7.4) for 5 min each. The brain slides were stained using H&E.

Commercial ELISA kits (Mouse TNF- α Reagent kit, IL-1 β test kit, Shanghai Xinle Biotechnology Co., Ltd., China) were used to determine the levels of TNF-α and IL-1β in mouse serum. The levels of MDA and GSH in the brain tissue were measured according to the instructions of the Nanjing Jiancheng kits.

Total RNA was extracted from the brain tissue using the BIOZOL reagent (Hangzhou Bioer Technology Co., Zhejiang, China). RNA concentrations were determined using a NanoDrop 2000/2000C (Thermo Fisher Scientific; Waltham, MA, United States), and diluted to the same concentration with enzyme-free water. RNA was reverse transcribed into cDNA using the cDNA synthesis kit (US Everbright, China) according to the manufacturer’s protocol. After reverse transcription, the samples were diluted 5-fold with enzyme-free water, and then the reaction system was prepared with the 2× SYBR Green qPCR Master Mix (Suzhou Hengyu Biotechnology Co., Ltd., Jiangsu, China) for quantitative analysis in a LightCycler^® 96 (Roche Diagnostics, Switzerland) using the following condition: 95°C for 120 s, followed by 40 cycles of 95°C for 5 s and 60°C for 30 s. The mouse primers are listed in Table 1. Gapdh was used as an internal reference and relative gene expression was calculated using 2 ^{− △△Ct}.

Microbial community analysis was performed by Shenzhen Huada Gene Technology Service Co., Ltd. (Shenzhen, China). The total DNA of the microbial community was isolated from mouse fecal samples according to the E.Z.N.A. Stool DNA Extraction Kit (Omega Biotek, Norcross, GA, United States). The V3–V4 hypervariable regions of the bacteria 16S rRNA gene were amplified with primers 341F (5’-CCTACGGGNGGCWGCAG-3′) and 806R (5’-GGACTACHVGGGTATCTAAT-3′) by thermocycler PCR system (GeneAmp 9,700, ABI, United States). The PCR reactions were conducted using the following program: 3 min of denaturation at 95°C, 27 cycles of 30 s at 95°C, the 30 s for annealing at 55°C, and 45 s for elongation at 72°C, and a final extension at 72°C for 10 min. PCR reactions were performed in triplicate 20 μL mixture containing 4 μL of 5 × FastPfu Buffer, 2 μL of 2.5 mM dNTPs, 0.8 μL of each primer (5 μM), 0.4 μL of FastPfu Polymerase and 10 ng of template DNA. Genomic DNA was randomly fragmented, which will be selected by magnetic beads to purify PCR-amplified products and dissolved in an elution buffer to complete library construction. The fragment range and concentration of the library were checked using the Agilent 2,100 Bioanalyzer. Qualified libraries were sequenced on the DNBSEQ platform by MGI2000 according to the size of the inserted fragments. The filtered clean data were used for later analysis. Sequence splicing was performed using FLASH software (Fast Length Adjustment of Short reads, v1.2.11), and the paired reads were spliced into a sequence through overlapping relationships to obtain tags of hypervariable regions. The software USEARCH (v7.0.1090) was used to cluster the spliced tags into OTUs (Operational Taxonomic Units), compared with the database, and species annotation. Based on the OTU and annotation results, species difference analyses between groups were performed by the Kruskal test. R (version 3.1.1) with the “vegan” package was employed to conduct the Principal Coordinate Analysis (PCoA), and the “ggplot2” package was utilized for the visualization of the PCoA diagram. R (v3.1.1) “gplots” package was used to form a heatmap. R (v3.4.1) was used to make species histograms.

As previously described (29), the concentration of SCFAs in feces were detected using gas chromatography (GC). The standard curve was constructed with the standard mixture of different concentrations of SCFAs. Briefly, 0.20 g feces were homogenized with 1 mL of MilliQ water and 0.15 mL 50% H₂SO₄ (w/w). The mixture was added with 1.6 mL of diethyl ether, and the samples were incubated on ice for 20 min. Centrifuge at 8000 rpm for 5 min to obtain supernatant, and filter through 0.2 μm filter (Branch billion Lung Experimental Equipment Co., Ltd., Tianjin, China) into clear GC vials. The standards of SCFAs included acetate, A116173; propionate, P110445; butyrate, B11se0438 (Aladdin Bio-Chem Technology Co., LTD, Shanghai, China). The concentration of SCFAs in fecal was analyzed using a GC-2014C gas chromatograph (Shimadzu Corporation, Kyoto, Japan), equipped with a DB-FFAP capillary column (Agilent Technologies, Wilmington, DE, United States) and flame ionization detector. The following heating program: initial temperature was employed: initially 50°C for 3 min, 10°C/min to 130°C, 5 ° C/min to 170°C, 15 ° C/min to 220°C, where it was maintained with maintenance for 3 min. The injector temperature was 250°C and the detector temperature was 270°C.

In vivo data are presented as mean ± SEM of at least six independent experiments, and in vitro data are presented as mean ± SD. Significant differences among means were evaluated using one-way ANOVA followed by the Tukey post-hoc test for multiple comparisons in GraphPad Prism 8.0 software (GraphPad Software Inc., San Diego, CA, United States). A p-value of less than 0.05 was considered statistically significant.

CS was extracted from the nasal cartilage of Shaanxi yellow cattle using a complex enzyme method (papain: trypsin = 1:2). The resulting CS sample appeared as a white powder, with yield of 7.50 ± 0.34%. In this study, CS degradation was achieved using the H₂O₂-Vc oxidation method; this method was selected owing to its mild degradation conditions and simplicity, resulting in minimal structural damage to the glycan chains. By changing the concentrations of H₂O₂ and Vc, CS components with different molecular weights were prepared.

Maintaining the concentration of H₂O₂ (60 mM) and varying the concentration of Vc (0–60 mM), as shown in Figure 1A, resulted in a gradual decrease in the molecular weight of the sample with increasing Vc concentration. When the concentration of H₂O₂ was 60 mM and that of Vc was 20 mM, the degree of degradation reached its maximum, and the molecular weight decreased to 14.7 kDa. The concentration of Vc (20 mM) was maintained while that of H₂O₂ was varied from 0 to 80 mM, as illustrated in Figure 1B. At an H₂O₂ concentration of 40 mM, the molecular weight decreased to 16.2 kDa. With a further increase in the concentration of H₂O₂, the molecular weight stabilized at approximately 15 kDa. When the concentration of H₂O₂ was increased from 60 to 80 mM, the change in molecular weight was not significant, possibly because the concentration of H₂O₂ reached an optimal range.

Since the molecular weight of CS before degradation was 98.7 kDa, in order to further explore the effects of molecular weight on the biological activity of CS, we prepared medium molecular weight segments (CS-2, 50.9 kDa), which produced when the concentration of H₂O₂ was 7 mM and Vc was 7 mM. In addition, the commercially available CS standard (CS-S, 80.8 kDa) was used as a control for subsequent studies. The physicochemical properties of CS, CS-S, and their degradation components, CS-1 and CS-2, are shown in Table 2. No significant changes were found in the sulfate and uronic acid contents of the degraded CS-1 and CS-2, and CS.

Fourier transform infrared (FT-IR) spectroscopy is a valuable tool for assessing the structure of polysaccharides and obtaining data on their functional groups and chemical bonds. FT-IR spectroscopy was conducted using CS and its degradation components within the range of 4,000 cm⁻¹ to 500 cm⁻¹ (Supplementary Figure S1). The peaks around 1,238 cm⁻¹ and 856 cm⁻¹ were derived from the tensile vibration of sulfate S=O in the axial position and the bending vibration of C-O-S, respectively, indicating the presence of several spectral bands associated with sulfate esters. The signals of O-H stretching vibration and C=O asymmetric stretching vibration of N-acetylaminogalactose and glucuronic acid were visible at 3441 cm⁻¹ and 1,639 cm⁻¹. Furthermore, the symmetric stretching vibration of glucuronic acid COO⁻ and the stretching vibration of C-O within COOH were identified at 1410 cm⁻¹ (30). These results indicate that H₂O₂ and Vc did not damage the main functional groups involved in the degradation of polysaccharides.

The ability of CS and its degradation components to scavenge DPPḤ free radicals is shown in Figure 2A. CS and its degradation components exhibited DPPḤ free radical scavenging ability in a concentration-dependent manner. At 30 mg/mL, CS-1 and CS-2 exhibited clearance rates of 79.92 ± 0.84% and 58.32 ± 4.85%, respectively, surpassing those of CS by factors of 1.83 and 1.07, respectively. As shown in Figure 2B, 30 mg/mL of CS-S, CS, CS-1, and CS-2 exhibited an ABTS⁺· clearance rate of 40.97 ± 0.02%, 44.62 ± 0.10%, 58.27 ± 0.03%, and 52.04 ± 0.01%, respectively. Notably, CS-1 showed a significant improvement in clearance rate for ABTS⁺· compared to CS. As shown in Figure 2C, CS and its degradation components displayed good ferrous reduction abilities, which gradually increased with increasing sample concentration. At 30 mg/mL, CS-1 exerted the strongest reduction ability, surpassing that of CS by a factor of 1.07 and that of CS-2 by 0.07. Of note, 30 mg/mL CS-S, CS, CS-1, and CS-2 significantly differed in their ability to clear ·OH, with values of 61.00 ± 1.68%, 63.85 ± 2.29%, 90.84 ± 0.35%, and 86.46 ± 3.44%, respectively (Figure 2D).

By utilizing the natural exploratory behavior of mice in novel and diverse environments, the Y-maze test can effectively measure the spatial work abilities of experimental animals. The animal time-processing axis is shown in Figure 3A. Based on the Y-maze test (Figure 3B), the percentage of spontaneous alternation in mice in the LPS group significantly decreased compared to that of mice in the CON group (p < 0.05), indicating a notable reduction in autonomous and working memory abilities after LPS treatment. This trend was reversed in the CS-1 and CS-2 groups (p < 0.05), suggesting that LMWCS can effectively improve the spatial working memory of LPS-induced mice.

During the period of the novel object recognition test, the learning and memory abilities of mice were evaluated based on their innate preference for novelty. As shown in Figure 3C, mice in the LPS group had a significant reduction in the discrimination index of new and old objects (p < 0.05). Compared with the LPS group, mice in both the CS-1 and CS-2 groups had a significant increase in the discrimination index of new objects (p < 0.05), with more evident improvement effects observed for mice in the CS-2 group.

The Barnes maze was employed to evaluate spatial learning and memory. As shown in Figure 3D, the exploration time of the Barnes target box was significantly reduced in the LPS group (p < 0.05). Intriguingly, mice in the CS-1 and CS-2 groups had significantly longer exploration time than mice in the LPS group (p < 0.05), with a duration 3.87- and 2.76-fold longer than that in the LPS group. As illustrated in Figure 3E, compared to the CON group, the Barnes target quadrant exploration time in the LPS group was significantly reduced (p < 0.05), and the exploration time in the CS-1 group was significantly higher than that in the LPS group (p < 0.05), with no significant difference found relative to the CON group. The exploration distances of the CS-1 and CS-2 groups were significantly higher than that of the LPS group (p < 0.05) (Figure 3F), indicating that LMWCS effectively improved spatial memory in LPS-induced mice. No significant differences were observed between the CS-1 and CS-2 interventions.

To reveal the potential neuroprotective effects of LMWCS supplementation against LPS-induced damage, H&E staining was performed to observe the morphology and distribution of neuronal cells in the hippocampus and cortex of mice. As illustrated in Figure 4A, the arrangement of neurons in the LPS-induced mouse brain was loose and disordered, and neurons disappeared or underwent nuclear pyknotic morphological changes. However, LMWCS supplementation significantly ameliorated these morphological abnormalities and reduced neuronal loss. To determine the impact of LMWCS on LPS-induced systemic and brain inflammatory responses, the expressions of inflammatory mediators, such as IL-1β and TNF-α were further detected. TNF-α is one of initial factors leading to the cascade reaction of inflammatory mediators and plays a positive feedback role in amplifying inflammation (31). In addition, IL-1β is closely related to the formation of acute neuroinflammation in vivo. IL-1β in the brain of rodents will cause the rapid activation of astrocytes and microglia, which are the main immune cells in the brain, and their activation is the key feature of neuroinflammation (32). As depicted in Figures 4B,C, compared to the CON group, the levels of IL-1β and TNF-α significantly increased in the serum of mice in the LPS group (p < 0.05), while those of these inflammatory mediators in the serum of mice in the CS-1 and CS-2 groups were obviously reduced (p < 0.05). CS-1 and CS-2 also substantially downregulated the expressions of inflammation-related genes in mouse brain tissue, including the mRNA levels of Il-1β and Tnf-α (p < 0.05) (Figures 4D,E). In terms of reducing oxidative stress in the brain, as depicted in Figures 4F,G, dietary supplementation with CS-1 and CS-2 enhanced the levels of GSH in brain tissue and inhibited the production of the lipid peroxidation product, MDA (p < 0.05). This result indicates that LMWCS regulated the oxidative stress state, as well as dampened neuroinflammation, thereby mitigating LPS-induced learning and memory disorders in mice.

H&E staining (Figure 5A) revealed a normal colon tissue structure in CON mice, with no signs of mucosal, crypt, or glandular damage. No evidence of inflammatory cell infiltration and abundant goblet cells was found. In contrast, the colon tissues of mice in the LPS group were disordered and characterized by mucosal and submucosal erosion, severe ulceration, bleeding, and granulation tissue formation. The crypt structure was disordered with cryptitis and crypt abscess formation. The glandular arrangement was irregular and marked by atrophy, infiltration by numerous inflammatory cells and neutrophils, and a decrease in the number of goblet cells. After LMWCS intervention, compared to the LPS group, the CS-1 and CS-2 groups showed varying degrees of improvement in lesions. Mucosal lesions were only observed in the mucosa and submucosa, with reduced damage to the crypts and glands, reduced inflammatory cells, and increased goblet cells.

The SCFA content in the intestinal microflora was determined using GC–MS, and the effects of LMWCS on LPS-induced intestinal microflora metabolism in mice were explored. As shown in Figure 5B, compared to the CON group, the acetic acid content in the colon of the LPS group was significantly reduced (p < 0.05). Following supplementation with CS-1 and CS-2, the acetic acid content significantly increased (p < 0.05). In particular, CS-1 + LPS had a better improvement effect, exhibiting a 3.88-fold higher effect than that of LPS mice. As depicted in Figure 5C, the propionic acid content in the LPS group was significantly lower than that in the CON group (p < 0.05). After supplementation with LMWCS, the propionic acid content in the CS-1 + LPS and CS-2 + LPS groups was significantly higher than that in the LPS group (p < 0.05), with no significant difference between CS-1 and CS-2. As shown in Figure 5D, the butyric acid content in the colon of mice in the CS-1 + LPS and CS-2 + LPS group was significantly higher than that of mice in the LPS group (p < 0.05). Surprisingly, supplementation with CS-1 increased the level of isobutyric acid (p < 0.05), while CS-2 did not (Figure 5E). There was no significant difference in the valeric acid and isovaleric acid content in the CS-1 + LPS and CS-2 + LPS groups compared with the LPS group (Figures 5F,G).

The Venn plot was generated to analyze the similarity and overlap in OTU composition among the different treatment groups. All groups had the same 624 OTUs. The unique OTU numbers in the CON, LPS, CS-1 + LPS, and CS-2 + LPS groups were 21, 24, 18, and 12, respectively (Figure 6A). Therefore, treatment with CS-1 and CS-2 can significantly affect the OTU composition of the intestinal microbiota in mice altered by LPS treatment. As shown in the Principal Coordinate Analysis (PCoA) diagram (Figure 6B), the gut microbiota structure significantly differed among the CON, LPS, CS-1, and CS-2 groups, indicating that dietary supplementation with LMWCS led to alterations in the β-diversity of intestinal microorganisms in mice. The changes in the composition of the gut microbiota at the family level in each group are shown in Figure 6C. At the family level, the gut microbiota is mainly composed of Lachnospiraceae, Erysipelotrichaceae, Coriobacteriaceae, Ruminococcaceae, Porphyromonadaceae, and Desulfovibrionaceae. The mice in the LPS group had markedly increased enrichment of Porphyromonadaceae, Erysipelotrichaceae, and Coriobacteriaceae (associated with gastrointestinal disease) and decreased relative abundance of Lacnospiraceae and Ruminococcaceae (SCFAs-producing) compared to the control group, while the change that was restored by the LMWCS treatments. Compared to treatment with LPS, dietary supplementation with LMWCS significantly increased the relative abundances of Sutterellaceae, Streptococcaceae, Eubacteriaceae, and Lachnospiraceae, while reducing the relative abundance of Erysipelotrichaceae and Porphyromonaceae, indicating that CS-1 and CS-2 can maintain the steady-state of the microbiota to varying degrees. Furthermore, the abundances of Streptococcus, Eisenbergiella, Vampirovibrio, Coprococcus, Enterococcus and Lachnoanaerobaculum were significantly increased (p < 0.05) in the intestines of mice in the CS-1 and CS-2 group, whereas those of Parabacteroides and Mycoplasma were significantly decreased (p < 0.05) (Figures 6D,E).

To assess the potential relationship between the distinct bacteria and amelioration of LPS-induced cognitive deficits in mice administered CS-1 and CS-2, a correlation analysis was performed based on the Spearman correlation coefficient. As shown in Figure 6F, Parabacteroides, the dominant genus in LPS group showed a negative correlation with GSH content but a positive correlation with MDA content. Streptococcus, Vampirovibrio, Burkholderia, Enterococcus and Aquabacterium were negatively associated with serum TNF-α levels, while Parabacteroides, Porphyromonas and Olsenella showed the positive correlation. Vampirovibrio, Coprobacillus, Lachnoanaerobaculum and Aquabacterium showed positive correlations with Barnes target hole exploration time, while Mycoplasma was negatively associated. These results indicate that CS-1 and CS-2 alleviate LPS-induced learning and cognitive impairment in mice by regulating the gut microbiota.