3.1. Clinical characteristics of subjects
We enrolled 15 patients with CP and 15 HCs in the study. The average age of CP patients was 46.07 ± 13.93 years, and HCs averaged 48.07 ± 16.09 years. No significant differences in age or gender were observed between the patients and the HCs (p=0.719, p>0.99). The relevant clinical data for each individual is presented in Table 1.
3.2. Changes in gut microbiome composition in patients with CP
Based on the abundance of genes in each sample, we randomly selected varying sample sizes to construct and visualize dilution curves for Core and Pan genes. The curve gradually flattens as the sample size increases, indicating that the sequencing sample size is adequate for the analysis requirements (Fig. 1A). And gene number differences analysis indicated that the number of unique genes in the CP group and HC group was 468,334 and 143,060, respectively (Fig. 1B).
Table 1. Clinical characteristics of subjects
|
Total (n = 30)
|
CP (n = 15)
|
HC (n = 15)
|
P values
|
Gender, n (%)
|
|
|
|
>0.99
|
Female
|
13 (43)
|
6 (40)
|
7 (47)
|
|
Male
|
17 (57)
|
9 (60)
|
8 (53)
|
|
Age, Mean ± SD
|
47.07 ± 14.82
|
46.07 ± 13.93
|
48.07 ± 16.09
|
0.719
|
Abbreviation: CP= Craniopharyngioma, HC= healthy control.
Subsequently, we conducted anosim analysis at the genus and species levels. The results indicated that the discrepancy between the two groups was significantly greater than the intra-group (p=0.015; p=0.018) (Fig. 1C, D). Additionally, at both the genus and species levels, we plotted a Bray-Curtis-based Principal Coordinates Analysis (PCoA) score map. This map revealed a clear separation between the CP group and the HC group (Fig. 1E, F). Further, at both the genus and species levels, based on the analysis of species discrepancies between the two groups, we generated species abundance clustering heatmaps for groups and samples (Fig. 2A, B, C, D).
Furthermore, to identify species with significant differences in abundance between the two groups, we conducted an LDA Effect Size analysis. Subsequently, we created a species evolutionary branching diagram and an LDA value distribution histogram (Fig. 3A, B). Ultimately, in the CP group, the results indicated an up-regulation of Eubacteriales, Bacillota, Roseburia, and others, while Bacteria were down-regulated. The overall results clearly indicated a significant difference in the composition of gut microbes between the CP group and the HC group.
To characterize the functional variations in the gut microbiome, we conducted metagenomic functional annotation of the Kyoto Encyclopedia of Genes and Genomes (KEGG) module. The results revealed that the most significant variation between the two groups was related to metabolic pathways (Fig. 3C). The most pronounced aspects included Carbohydrate metabolism, Amino acid metabolism, and Metabolism of cofactors and Vitamins. The unique gene content for each aspect was 79,972, 56,671, and 42,275, respectively. Furthermore, the unique genes associated with Energy metabolism and Glycan biosynthesis and metabolism exceeded 30,000.
3.3. The difference in fecal metabolism between the CP group and HC group
Considering that the genetic composition of the gut microbiota in the CP group and HC group differs significantly, particularly in metabolism, untargeted LC-MS metabolomics were employed to analyze the metabolomic profiles of fecal samples from both groups. Fecal samples were analyzed using LC-MS in positive ion mode (POS) and negative ion mode (NEG), respectively. PLS-DA was employed to build a relationship model between metabolite expression and sample class to differentiate the metabolic profiles among the groups. The model was verified without overfitting and can be applied for further data analysis (Fig. 4A, B).
Subsequently, volcano maps were plotted to depict the overall distribution of differential metabolites between the two groups (Fig. 4C, D). Additionally, to visually identify the up-regulated and down-regulated metabolites with significant differential multiples, we generated match maps of the top 20 metabolites based on the differential metabolites from each group of comparative analyses (Fig. 4E, F). As shown in the diagram, in the positive ion mode, Silibinin, Lysopc 20:0, LysoPC 20:2, and other substances were significantly up-regulated, while Ricinine, EMH, Estazolam, and other substances were markedly down-regulated. However, in the negative ion mode, Tauroursodeoxycholic acid, JWH 018 N-pentanoic acid metabolite, Xanthosine, and other metabolites were clearly up-regulated, whereas (+/-)10(11)-EpDPA, Rifampicin, beta-Nicotinamide mononucleotide, and other metabolites were significantly down-regulated.
Metabolites with similar metabolic patterns (up-regulated or down-regulated) may share similar functions or belong to the same pathway regulating the same metabolic process. Consequently, we conducted cluster analysis focusing on metabolites with comparable metabolic patterns (Fig. 5A, B). To better illustrate the correlation and degree of association between different metabolites in the samples, we generated chord diagrams based on the correlation coefficients of the varying metabolites between the two groups (Fig. 5C, D).
Additionally, bubble plots of the enriched KEGG pathways and regulatory interaction network diagrams for each group of differential metabolites were generated based on the enriched pathways of differential metabolites (Fig. 5E-H).
3.4. Correlation analysis of fecal microbiome and metabolism
In subsequent study, we explored the connection between modified microbiota and metabolic changes, conducting a metabolome-microbe correlation analysis. The results indicated that in positive ion mode, the quantity of the metabolite β-Cortolone exhibited a positive correlation with the abundance of the three microorganisms: Aquipluma, Blastopirellula, and Tamilnaduibacter. Simultaneously, in negative ion mode, the quantity of Ne-(1-Carboxymethyl)-L-lysine was positively correlated with the abundance of the same three microorganisms. We visualized the correlation between microorganisms and metabolites through heatmaps, chord diagrams, and Sankey diagrams, enhancing the intuitive understanding of the results (Fig. 6A-F).