Alterations of Fungal Microbiota in Patients With Cholecystectomy

Abstract

Increasing evidence suggests a high risk of gastrointestinal postoperative comorbidities (such as colorectal cancer) in patients with postcholecystectomy (PC). Although previous studies implicated the role of fungi in colon carcinogenesis, few reports focused on the fungal profile in patients with PC. We enrolled 104 subjects, including 52 patients with PC and 52 non-PC controls (CON), for fecal collection to detect the fungal composition by an internal transcribed spacer (ITS) 1 rDNA sequencing. Data showed that Candida (C.) glabrata and Aspergillus (A.) Unassigned were enriched, and Candida albicans was depleted in patients with PC. In addition, postoperative duration was the main factor to affect the fungal composition. Machine learning identified that C. glabrata, A. Unassigned, and C. albicans were three biomarkers to discriminate patients with PC from CON subjects. To investigate the fungal role in colon carcinogenesis, the subjects of the PC group were divided into two subgroups, namely, patients with PC without (non-CA) and with precancerous lesions or colorectal cancer (preCA_CRC), by histopathological studies. C. glabrata was found to be gradually accumulated in different statuses of patients with PC. In conclusion, we found fungal dysbiosis in patients with cholecystectomy, and the postoperative duration was a potent factor to influence the fungal composition. The accumulation of C. glabrata might be connected with carcinogenesis after cholecystectomy.

Keywords: Candida; Candida glabrata (C. glabrata); carcinogenesis; cholecystectomy; fungal microbiota.

Figure 1

Fungal diversity and composition in PC and CON subjects. (A) Fungal alpha diversity based on Chao1 and Shannon index dealt with base 2 logarithm (log2, Shannon_2 index); ns, not significant. (B) Alpha rarefaction curve. (C) Beta diversity based on Bray–Curtis distance. (D) Fungal composition of top 20 species in CON and PC groups. Visualization was performed on the Circos website (http://circos.ca/). The color of the left outer and right inner bars represents groups (blue, CON; red, PC), and the color of the left inner bands and right outer bars represents fungal species. CON, non-PC control subjects; PC, postcholecystectomy. The left inner values represent the relative abundance of specific fungal contents in the total fungal community. The right inner values represent the sum value of these species in the CON or PC group.

Figure 2

Enriched fungal ASVs and fungal biomarkers in patients with PC. (A) Enriched fungal ASVs in the PC group. EdgeR package was used for comparative analysis. The difference between the two groups is shown as a Manhattan diagram. Point shape indicates ASV enriched, depleted, or not significant in the former group compared with the latter one. Point color indicates fungal phylum. Point size indicates the abundance of ASV. CPM, count per million. (B) Machine learning based on random forest analysis to identify the fungal biomarkers. Fungal profiles from half of the subjects (N_CON = 26, N_PC = 26, randomly) were utilized for the model of machine learning. (C) Discrimination of patients with PC from CON subjects based on the fungal biomarkers. The receiving operational curve (ROC) analysis was performed on the rest of the subjects (N_CON = 26, N_PC = 26). Area under curve, AUC. The line colors represent the fungal biomarkers from the results of machine learning. Blue, Candida sglabrata; red, Aspergillus Unassigned; black, Candida albicans; CON, non-PC control subjects; PC, postcholecystectomy; ASVs, amplicon sequence variants.

Figure 3

Environmental factors and clinical indexes associated with fungal microbiota. (A) Redundancy analysis (RDA) of environmental factors correlated with the fungal profile. The red arrow represents environmental factors and the blue represents fungal phyla. Results of the permutation test on all axes showed that pseudo-F equaled 2.2 with a P-value of 0.014. DM, diabetes mellitus; BMI, body mass index; HBP, high blood pressure; FLD, fatty liver disease; CHD, coronary heart disease; HLP, hyperlipemia; symptom, PC patients with gastrointestinal symptoms, including acid reflux, indigestion, bellyache, and diarrhea. Duration, years after patients with PC undergoing cholecystectomy. (B) Environmental factors fitting analysis. Only factors with significant correlation with sample clustering were noted as colored arrows in the panel. (C) RDA of clinical indexes correlated with the fungal profile. The red arrow represents clinical indexes and the blue represents fungal phyla. Results of the permutation test on all axes showed that pseudo-F equaled 1.8 with a P-value of 0.026. WBC, white blood cell; NE, neutrophilic granulocyte; ALP, alkaline phosphatase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; HB, hemoglobin; PLT, platelet count; GGT, glutamyl transpeptidase; ALB, albumin; DBIL, direct bilirubin; TBIL, total bilirubin. (D) Clinical indexes fitting analysis. Only factors with significant correlation with sample clustering were noted as colored arrows in the panel.

Figure 4

Fungal diversity and composition in PC patients with or without precancerous lesions and colorectal cancer. (A) Alpha diversity in patients with PC, displayed as chao1 index and Shannon index dealt with base 2 logarithm (log2, Shannon_2 index); ns, not significant. (B) Alpha rarefaction curve of two groups; (C) beta diversity based on Bray–Curtis distance. (D) Fungal composition of top 20 species in CON, non-CA (N = 43) and preCA_CRC (N = 9) groups. Visualization was performed on the Circos website (http://circos.ca/). The left inner values represent the relative abundance of specific fungal contents in the total fungal community. The right inner values represent the sum value of these species in the CON or PC group. CON, non-PC control subjects; non-CA, PC patients without precancerous lesions or colorectal cancer; preCA_CRC, PC patients with precancerous lesions or colorectal cancer.