Baseline clinical data of recruited CD patients
A total of 38 CD patients (19 patients in the active stage and 19 patients in the remission stage) were finally enrolled in the present study (Table 1), as noted previously20. The indexes, including age, gender, body mass index, lesion location, duration of disease, and history of life (diet, smoking, and alcohol use), were similar in these individuals (all P > 0.05). The patients in the active stage had a higher nutritional risk than their counterparts in the remission stage (P = 0.0300). Moreover, the levels of C reactive protein, hemoglobin, albumin, prealbumin, and the erythrocyte sedimentation rate were significantly different between the two groups (all P < 0.05). However, the levels of procalcitonin, interleukin (IL)−1, IL−6, IL−8, IL−10, and white blood cells were similar (all P > 0.05, Table 1).
Alterations of ileum microbiota abundance and diversity in CD patients
The alpha diversity was evaluated by Chao and Shannon diversity indexes between the active and remission stages in CD patients. The Chao index indicated the abundance of ileum microbiota, while the Shannon index was performed to show the diversity of the microbiota. Compared with the remission stage, Chao and Shannon diversity indexes were higher in the active stage, while no significant differences were observed (all P > 0.05, Fig. 2A, 2B). PCoA was also performed to evaluate the beta diversity of ileum microbiota. On the phylum level, PC1 and PC2 accounted for 28% and 12.49% of the variance, respectively. A similarity analysis did not reveal a statistically significant difference between these two groups (P = 0.0770, Fig. 2C). Moreover, PC1 and PC2 accounted for 19.86% and 10.4% of the variance on the genus level. A significant difference was observed in the analysis of similarity (P = 0.0010, Fig. 2D).
Structural alterations of ileum microbiota in CD patients
The ileum microbiota taxonomic profiles between active and remission stages of CD were analyzed on the phylum and genus levels. On the phylum level, the ileum microbiota was partitioned into four main enterotypes based on the CH index, including Proteobacteria, Firmicutes, Bacteroidetes, and Firmicutes_1, respectively. Among them, ileum microbiota in the active stage of CD was mostly of the type Firmicutes_1, while in the remission stage, it was mostly of the type Firmicutes (Fig. 3A, 3B). Moreover, seven dominant phyla were identified in all samples. The relative abundances of Bacteroidetes, Actinobacteria, Fusobacteria, and Synergistota were higher in the active stage of CD, while the converse were observed for Firmicutes, Proteobacteria, and Patescibacteria (Fig. 3C, 3E). The only significant difference in the dominant phyla was Bacteroidetes (P = 0.0007). Furthermore, Desulfobacterota, Unclassified_k__norank_d__Bacteria, and Myxococcota were significantly different between the active and remission stages (P = 0.0002, P = 0.0004, and P = 0.0490, respectively) (Fig. 3D).
On the genus level, the ileum microbiota was also partitioned into two main enterotypes based on the CH index, these being Streptococcus and Escherichia-Shig. Among them, ileum microbiota in the active stage of CD was mostly of the type Streptococcus, whereas in the remission stage, it was mostly of the type Escherichia-Shigella (Fig. 4A, 4B). Streptococcus and Bacteroides were the most significantly different genera enriched by LEfSe in the active and remission stages of CD, respectively (Fig. 4C). The main genera with significant differences included Veillonella, Lactococcus, Klebsiella, Unclassified_f__Enterobacteriaceae, and Blautia (P = 0.0002, P = 0.0039, P = 0.0019, P = 0120, and P = 0.0339, respectively), which were more abundant in the remission stage of CD. Conversely, Ruminococcus_torques_group, Ralstonia, and Collinsella were more abundant in the active stage (P = 0.0002, P = 0.0101, and P = 0.0029, respectively). These microbiota were also significantly different between the two stages. Interestingly, the relative abundance of Lactobacillus was higher in the remission stage with statistical difference (Fig. 4D−4F).
Alterations of serum bile acids in CD patients
The total concentrations of serum primary and secondary bile acids in the active stage of CD were higher than those in the remission stage. A significant difference was observed in serum secondary bile acids between the active and remission stages of CD (P = 0.0006), but the difference was not statistically significant regarding serum primary bile acids (P = 0.4526, Fig. 5A).
Concerning serum primary bile acids, the proportions of cholic acid, chenodeoxycholic acid, and taurochenodeoxycholic acid were higher in the active stage of CD, while the opposite were observed for glycocholic acid, glycinechenodeoxycholic acid, and taurocholic acid. Significant differences were observed regarding cholic acid, chenodeoxycholic acid, glycocholic acid, glycinechenodeoxycholic acid, and taurocholic acid (P = 0.0005, P = 0.0006, P = 0.0046, P = 0.0090, and P = 0.0130), respectively. As for serum secondary bile acids, the proportions of glycolithocholic acid, glycideoxycholic acid, glycoursodeoxycholic acid, and taurolithocholic acid were higher in the active stage of CD, while the opposite were observed for deoxycholic acid, ursodeoxycholic acid, lithocholic acid, taurodeoxycholic acid, and tauroursodeoxycholic acid. However, the differences were not statistically significant (Fig. 5C).
Alterations of serum amino acids in CD patients
Compared with the active stage of CD, the total concentrations of serum essential and non-essential amino acids were higher in the remission stage, but no statistical difference was observed between these two stages (P = 0.1668 and P = 0.1628, Fig. 5B).
Regarding serum essential amino acids, the proportions of valine, phenylalanine, methionine, threonine, and isoleucine were higher in the active stage of CD, while the opposite were observed for leucine, lysine, tryptophan, and histidine. The only significant difference in essential amino acids between active and remission stages was isoleucine (P = 0.0010). As for serum non-essential amino acids, arginine, serine and aspartic acid proportions were higher in the active stage of CD, and significant differences were observed regarding arginine and serine (P < 0.0001 and P = 0.0430). Conversely, the proportions of glycine, tyrosine, alanine, proline, glutamate, and asparagine were lower in the active stage of CD than in the remission stage. The serum non-essential amino acids with a significant difference between the active and remission stages were glycine, tyrosine, and asparagine (P = 0.0400, P = 0.0150, and P = 0.0030, respectively) (Fig. 5D).
Correlations among ileum microbiota, serum metabolites, and clinical factors in CD patients
The correlations among the key phyla, serum metabolites and clinical factors were evaluated in patients with CD. Interestingly, Desulfobacterota, Unclassified_k__norank_d__Bacteria, and Bacteroidetes were all positively associated with arginine (R = 0.3668, P = 0.0235; R = 0.3515, P = 0.0305 and R = 0.3352, P = 0.0397, respectively) and CDAI score (R = 0.5761, P = 0.0002; R = 0.5066, P = 0.0012 and R = 0.4331, P = 0.0066, respectively). Conversely, glycine (R=-0.3494, P = 0.0316; R=-0.4723, P = 0.0028 and R=-0.4413, P = 0.0055, respectively), citrulline (R=-0.3921, P = 0.0149; R=-0.3371, P = 0.0385 and R=-0.3228, P = 0.0481, respectively), glycinechenodeoxycholic acid (R=-0.3802, P = 0.0185; R=-0.3444, P = 0.0343 and R=-0.5360, P = 0.0005, respectively), and prealbumin (R=-0.5659, P = 0.0002; R=-0.4868, P = 0.0019 and R=-0.5887, P = 0.0001, respectively) showed negative correlations (Fig. 6A, 6C).
Furthermore, the key genera which were more abundant in the active stage of CD were evaluated. Ralstonia, Ruminococcus_torques_group, and Bacteroides were positively associated with glycolithocholic acid (R = 0.5424, P = 0.0004; R = 0.6898, P = 0.0002 and R = 0.6086, P = 0.0001, respectively), glycideoxycholic acid (R = 0.3834, P = 0.0175; R = 0.5180, P = 0.0009 and R = 0.4815, P = 0.0022, respectively), and CDAI score (R = 0.4372, P = 0.0061; R = 0.5494, P = 0.0004 and R = 0.4204, P = 0.0086, respectively), whereas they were negatively correlated with citrulline (R=-0.4035, P = 0.0120; R=-0.3479, P = 0.0323 and R=-0.3946, P = 0.0142, respectively) and prealbumin (R=-0.4764, P = 0.0025; R=-0.5738, P = 0.0002 and R=-0.5570, P = 0.0003, respectively) (Fig. 6B, 6D).
Regarding the key genera which were more abundant in the remission stage of CD, CDAI score, the only clinical factor, was negatively correlated with Veillonella (R=-0.5794, P = 0.0001), Unclassified_f__Enterobacteriaceae (R=-0.3709, P = 0.0219), Klebsiella (R=-0.5075, P = 0.0011), Rothia (R=-0.4687, P = 0.0030), and Streptococcus (R=-0.3898, P = 0.0156, Fig. 6B, 6D).
The potential correlations between serum metabolites and clinical factors were also investigated. CDAI score was positively associated with cholic acid (R = 0.3982, P = 0.0133), chenodeoxycholic acid (R = 0.4647, P = 0.0033), and arginine (R = 0.5052, P = 0.0012), both of which were more abundant in the active stage of CD. Conversely, CDAI score was negatively correlated with asparagine (R=−0.5644, P = 0.0002), glycine (R=−0.3703, P = 0.0221), tyrosine (R=−0.4131, P = 0.0099), and taurocholic acid (R=−0.3427, P = 0.0352), which were more abundant in the remission stage. However, prealbumin was positively associated with asparagine (R = 0.4838, P = 0.0021), glycine (R = 0.3993, P = 0.0130), tyrosine (R = 0.3951, P = 0.0141), taurocholic acid (R = 0.3505, P = 0.0310), glycocholic acid (R = 0.4228, P = 0.0082), and glycinechenodeoxycholic acid (R = 0.4553, P = 0.0041), while the opposite was observed for arginine (R=-0.3623, P = 0.0254, Fig. 6E).