Introduction

Osteoarthritis is a painful, disabling, and slowly developing degenerative disease that affects about 240 million people globally [1,2,3]. Chondroitin sulfate (CS) is a dominant family of sulfated linear polysaccharides that exist ubiquitously both on the cell surfaces and in the extracellular matrices of the human body [4,5,6]. Previous studies have indicated that CS is the most abundant structural component in the cartilage of the human joint tissues [4, 7, 8]. In this regard, following the recommendations of the European League Against Rheumatism (EULAR) and the European Society for Clinical and Economic Aspects of Osteoporosis, Osteoarthritis, and Musculoskeletal Diseases (ESCEO), CS has widely been used as a symptomatic slow-acting drug (SYSADOA) or a dietary supplement for the treatment and prevention of osteoarthritis [9, 10]. However, the use of CS for the management of osteoarthritis is still under debate since it cannot be absorbed after oral intake due to its polyanionic nature and large molecular weight [11,12,13,14]. Gut microbiota has recently been proposed to play a pivotal role in the metabolism of drugs and nutrients [15,16,17]. Nonetheless, how CS is degraded by the human gut microbiota has not been fully characterized [18].

Results and discussion

To address this issue, we first investigated the degradation profiles of CS by the human gut microbiota from 23 individuals using in vitro anaerobic fermentation (Fig. 1A). We found that each human gut microbiota was characterized by a unique capability for CS degradation (Fig. 1B; Supplementary Figure S1 and S2). For example, the gut microbiota of donor T25 only utilized about 20% of the original CS while that of donor T32 utilized more than 70% of the original CS in the culture medium (Fig. 1B; Supplementary Figure S2). However, on average, the human gut microbiota degraded and fermented about 40% of the original CS within 48 h (Fig. 1B; Supplementary Figure S2). This suggested that CS was a readily degradable polysaccharide for most individuals’ gut microbiota and thus might be used as a microbiota-accessible carbohydrate (MAC).

Fig. 1
figure 1

Degradation of CS by the human gut microbiota. Experimental design (A). Relative CS content in the culture medium (B). Heatmap of the concentrations of different SCFAs in the culture medium (C). TLC shows the degradation of CS by the gut microbiota of donor T32 (D). UPLC-MS/MS analysis of CSOSs in the culture medium of donor T32 (E). Total ion chromatograms showing the elution profiles of CSOSs in the culture medium of donor T32 at different time points (F). Part of the figure was created with BioRender.com. *p < 0.05; ***p < 0.001

Full size image

Short-chain fatty acids (SCFAs) are major fermentation products of MACs in the human gut [19]. With the help of high-performance liquid chromatography (HPLC), we found that CS fermentation by the human gut microbiota produced a significant amount of SCFAs that were dominated by acetate, propionate, and butyrate (Fig. 1C). Although succinate, lactate, and isovalerate were also detected in the medium, they were produced at a much smaller amount (Fig. 1C). Thin layer chromatography (TLC) and ultra-performance liquid chromatography (UPLC)-mass spectrometry (MS)/MS further confirmed that CS was degraded by the human gut microbiota to produce a series of unsaturated CS oligosaccharides (CSOSs) with a degree of polymerization (dp) ranging from 2 to 8 (Fig. 1D–F; Supplementary Figure S2 and S3). However, it should be noted that although different CSOSs were produced at the very first 24 h, the tetrasaccharide (udp4), hexasaccharide (udp6), and octasaccharide (udp8) were further degraded by the gut microbiota to produce disaccharide (udp2) as the fermentation continued (Fig. 1E, F; Supplementary Figure S3). Moreover, udp2 was the only degradation product left in the culture medium after 36 h (Fig. 1E, F; Supplementary Figure S2 and S3).

Due to the polyanionic nature and large molecular weight, CS could not be directedly absorbed after oral intake [11,12,13]. However, previous human and animal studies have well demonstrated that CS could be absorbed in the form of udp2 after oral intake [11, 14]. In view of the above results, our study suggested that the gut microbiota might to some extent be able to contribute to the absorption of CS in the intestine by degrading it into oligosaccharides. However, more detailed studies are needed to verify this possibility.

Using 16S rRNA gene amplicon high-throughput sequencing and bioinformatics analysis, we found a significant change in the composition of the human gut microbiota before and after fermentation (Fig. 2A; Supplementary Figure S4). Besides, some genera were enriched upon cultivation in the medium containing CS as the major carbon source (Supplementary Figure S4 and S5), suggesting the possibility that they might have degraded and utilized CS as a substrate for their growth. We next wondered which bacteria were responsible for CS degradation in the human gut microbiota. To answer this question, we isolated a total of 586 bacterial strains with a potential CS-degrading capability from all 23 human fecal samples using the well-established enrichment culture method (Fig. 2B, C; Supplementary Figure S6 and Table S1). 16S rRNA gene-based phylogeny suggested that these fecal isolates belonged to 48 different species of bacteria (Fig. 2C; Supplementary Figure S7 and Table S1). This implied that the CS-degrading capability might be widely distributed among the human gut microbiota. Additionally, distinct species of bacteria were isolated from the fecal samples of different individuals (Fig. 2C; Supplementary Figure S6). This indicated that each individual was characterized by its own unique bacteria for CS degradation, a further argument for the individualized metabolism of CS by the human gut microbiota.

Fig. 2
figure 2

Isolation of CS-degrading bacteria from the human gut microbiota. Changes in the structure of the human gut microbiota before and after fermentation. PCA analysis (A). Experimental design for the isolation of CS-degrading bacteria from the human gut microbiota (B). Heatmap of the number of CS-degrading bacteria isolated from all 23 human fecal samples (C). Part of the figure was created with BioRender.com

Full size image

With all the 48 species of intestinal bacteria in hand, we next sought to investigate and compare their degrading capabilities. Bacteroides salyersiae, Bacteoides finegoldii, Bacteroides xylanisolvens, Bacteroides thetaiotaomicron, and Bacteroides ovatus were identified as top 5 degraders for CS in the human gut microbiota (Supplementary Figure S7). Preceding studies have shown that B. finegoldii, B. xylanisolvens, B. thetaiotaomicron, and B. ovatus are well-recognized CS-degraders in the human gut [20]. In line with previous results [17, 20], we found that these bacteria could degrade approximately 26 to 59% of the original CS within 72 h (Supplementary Figure S7). However, the amount of CS consumed by these bacteria was relatively smaller as compared to that consumed by B. salyersiae, a proficient CS-degrader identified in the present study (Fig. 3; Supplementary Figure S7–S9). Even in the very first 24 h, up to 85% of the original CS was successfully degraded by B. salyersiae (Supplementary Figure S7). This made B. salyersiae a potent CS-degrader within the gut microbiota of the 23 individuals. Besides, CS degradation by B. salyersiae produced the highest amount of CSOSs as compared to that produced by the four well-established CS-degraders (Fig. 3F; Supplementary Figure S10–S12). Additionally, B. salyersiae grew very well in both the liquid and solid culture medium containing CS as the major carbon source (Figs. 3A and 4A, B). The major degradation product produced by B. salyersiae was udp4 and the major fermentation product produced by B. salyersiae was propionate (Fig. 3B–E; Supplementary Figure S12).

Fig. 3
figure 3

Degradation of CS by B. salyersiae CSP6. Colony forming units (CFUs) of B. salyersiae CSP6, B. finegoldii B36-12, and B. thetaiotaomicron E1-7 grew in the culture medium containing CS as the major carbon source (A). TLC showing the degradation of CS by B. salyersiae CSP6 (B). Concentrations of different SCFAs in the culture medium of B. salyersiae CSP6 (C). UPLC-MS/MS analysis of CSOSs produced by B. salyersiae CSP6 (D). Total ion chromatograms showing the elution profiles of CSOSs in the culture medium of B. salyersiae CSP6 at different time points (E). Comparison of the amount of CSOSs produced by B. salyersiae CSP6, B. finegoldii B36-12, B. xylanisolvens B33-17, B. thetaiotaomicron E1-7, and B. ovatus B33-4 at 72 h (F). *p < 0.05; **p < 0.01; ***p < 0.001

Full size image
Fig. 4
figure 4

Cell morphology and genomic analysis of B. salyersiae CSP6. Colony morphology of B. salyersiae CSP6 on the plate (A). Transmission electron microscope (TEM) analysis of the cell morphology of B. salyersiae CSP6 (B). B. salyersiae CSP6 was a round-shaped bacterium with a cell size of about 800 to 1000 nm (diameter). COG function analysis of the genome of B. salyersiae CSP6 (C). Analysis of the CAZymes in the genome of B. salyersiae CSP6 (D). A proposed model for understanding the degradation of CS by B. salyersiae CSP6 based on genomic analysis (E). Part of the figure was created with BioRender.com

Full size image

To exclude the possibility that the CS-degrading capacity was specific to the particular isolate of B. salyersiae obtained in the present study, we further tested the CS-degrading capacity of another human gut bacterium B. salyersiae FL17. Although this strain was not isolated from the aforementioned 23 individuals using the enrichment culture method, it was found to have the same CS-degrading capacity as compared to the one isolated in the present study (Supplementary Figure S13). Both strains degraded CS to produce CSOSs in the culture medium (Supplementary Figure S13). Besides, both strains utilized an equal amount of CS during fermentation (Supplementary Figure S13). Altogether, our study suggested that the CS-degrading capacity of B. salyersiae identified in the present study was possibly a general characteristic of this species.

Given that B. salyersiae was a potent species for CS degradation in the 23 human fecal samples tested in our study, we then wondered how CS was degraded by this specific gut anaerobe. To answer this question, we sequenced the whole genome of B. salyersiae. Bioinformatics analysis suggested that the genome size of B. salyersiae was 5,561,372 bp and the G + C content in the genome was 41.95% (Supplementary Figure S14). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis and clusters of orthologous groups (COG) function analysis indicated that most of the identified genes in B. salyersiae were involved in the carbohydrate transport and metabolism pathway (Fig. 4C; Supplementary Figure S14), suggesting that this bacterium might be skilled at degrading and metabolizing dietary polysaccharides. In this sense, we next analyzed the carbohydrate-active enzymes (CAZymes) in the genome of B. salyersiae. A total of 334 genes were identified as responsible for the expression of different classes of CAZymes including glycoside hydrolases (GHs), glycosyltransferases (GTs), polysaccharide lyases (PLs), carbohydrate esterases (CEs), carbohydrate-binding modules (CBMs), and auxiliary activities (AAs) in this bacterium (Fig. 4D).

The CAZymes that target the glycosidic bonds in CS from the Bacteoides spp. have been extensively studied [20]. Based on previous results [21,22,23], our genomic analysis suggested that PL29, PL8_3, and GH88 were candidate CS-metabolizing enzymes in B. salyersiae (Supplementary Table S2–S4). Besides, genes coding the starch utilization system (Sus) proteins were also identified in the genome of B. salyersiae (Supplementary Table S2–S4). The Sus proteins in Bacteoides spp. have been well-demonstrated to capture and transport different glycans [20, 23, 24]. Altogether, based on these bioinformatics results, we tentatively put forward a possible model for understanding the degradation of CS by B. salyersiae (Fig. 4E). However, although we have clearly demonstrated that CS was degraded by B. salyersiae, the CAZymes activities of this bacteria shown in the model have not been confirmed using genetic or biochemical approaches in the present research. Future studies are therefore warranted to further investigate the detailed molecular mechanisms involved in the degradation of CS by B. salyersiae.

As aforementioned, CS degradation by B. salyersiae produced a significant amount of udp4 in the culture medium (Fig. 3; Supplementary Figure S12). In this light, we next wondered if the produced udp4 could be further utilized by other anaerobes in the gut. To address this question, a spent medium assay was conducted to screen the candidate bacterium that could utilize the udp4 produced by B. salyersiae in the culture medium (Fig. 5A; Supplementary Figure S15). Of all the 54 different species of bacterial isolates tested, only Bacteroides stercoris was found to have the capability for udp4 utilization (Fig. 5; Supplementary Figure S16). This suggested that the udp4-utilizing capability might be very specific to only a few species of anaerobes in the human gut.

Fig. 5
figure 5

Cross-feeding interactions between B. salyersiae and B. stercoris. Experimental design of the spent medium assay (A). TLC showing the utilization of udp4 by B. salyersiae CSP6, B. stercoris P22-5, and B. stercoris P22-28 (B). B. salyersiae CSP6 was used as a negative control for the assay. The utilization of udp4 was monitored from 12 to 96 h using UPLC-MS/MS analysis. C Total ion chromatograms showing the elution profiles of udp4 in the culture medium of B. salyersiae CSP6 (D), B. stercoris P22-5 (E), and B. stercoris P22-28 (F) at different time points. A proposed model for understanding the cross-feeding interactions between B. salyersiae and B. stercoris (G). Part of the figure was created with BioRender.com. **p < 0.01; ***p < 0.001

Full size image

B. stercoris utilized about 40% of udp4 in the spent medium within 48 h (Fig. 5B–F; Supplementary Figure S16 and Figure S17). Besides, fermentation of udp4 by B. stercoris produced significant amounts of acetate and propionate (Supplementary Figure S16). Nonetheless, it should be noted that B. stercoris itself was not a good CS-degrader as it only degraded about 20% of the original CS in the culture medium even after 72 h (Supplementary Figure S7). Collectively, these results indicated that the udp4 produced by the primary degrader B. salyersiae had the potential to serve as a “public goods” molecule for the growth of B. stercoris, a secondary CS-degrader that was skilled at utilizing CSOSs but not CS (Fig. 5G).

Our study does not rule out the possibility that other CSOSs such as udp2, udp6, and udp8 might also be able to mediate the cross-feeding interactions between the primary and secondary CS-degraders in the human intestine. This could be the subject of future research. The ability of B. stercoris to utilize udp4 produced by B. salyersiae suggested that some bacteria might have co-evolved to work synergistically to degrade CS in the human gut. The cross-feeding interactions between B. salyersiae and B. stercoris identified in the present study provided a framework for understating the degradation and metabolism of CS in the human gut.

The evolution of the metabolic cross-feeding interactions between specific bacteria in the human intestine has been proposed to be driven by different factors [25,26,27]. Our study demonstrated that the udp4 produced by the primary degrader during the metabolism of CS could serve as a “public goods” nutrient for the growth of the secondary degrader in the same niche. In accordance with previous results [28,29,30], our findings reinforced the notion that intermediate oligosaccharides produced during the degradation of complex carbohydrates had the potential to drive the multi-species symbiotic cross-feeding in the human gut. These results opened a new window for understanding the modulatory effect of CS on the human gut microbiota. This is critically important since in vivo studies have well demonstrated that the gut microbiota is deeply involved in the pathogenesis of osteoarthritis [31,32,33] and that changes in the composition of gut microbiota are determining factors for achieving the therapeutic effect of CS in vivo [34,35,36].

Conclusions

Taken together, in the present research, we comprehensively investigated the detailed degradation profiles of CS by the gut microbiota from 23 healthy individuals and illustrated that B. salyersiae was a potent species for CS degradation in the human intestinal microbiota (Fig. 6). Our study suggested that B. salyersiae was potentially a keystone species for CS degradation in the human intestine. The udp4-based cross-feeding interactions identified in our study provide insights into the metabolism of CS by the human gut microbiota, which has promising implications for the development of medical and nutritional therapies for osteoarthritis.

Fig. 6
figure 6

A schematic diagram illustrating the degradation of CS by the human gut microbiota. Each gut microbiota was characterized by a unique capability for CS degradation. CS was readily degraded and fermented by specific anaerobes in the gut to produce SCFAs and CSOSs. These bacteria included B. salyersiae, B. finegoldii, B. xylanisolvens, B. thetaiotaomicron, and B. ovatus. B. salyersiae was identified as a potent bacterium for CS degradation in the present study. The udp4 produced by the primary degrader B. salyersiae sustained the growth of the secondary degrader B. stercoris. B. salyersiae and B. stercoris might have co-evolved to work synergistically to degrade and utilize CS in the human gut. The figure was created with BioRender.com

Full size image