Structural segregation of gut microbiota between colorectal cancer patients and healthy volunteers

Abstract

Despite a long-suspected role in the development of human colorectal cancer (CRC), the composition of gut microbiota in CRC patients has not been adequately described. In this study, fecal bacterial diversity in CRC patients (n=46) and healthy volunteers (n=56) were profiled by 454 pyrosequencing of the V3 region of the 16S ribosomal RNA gene. Both principal component analysis and UniFrac analysis showed structural segregation between the two populations. Forty-eight operational taxonomic units (OTUs) were identified by redundancy analysis as key variables significantly associated with the structural difference. One OTU closely related to Bacteroides fragilis was enriched in the gut microbiota of CRC patients, whereas three OTUs related to Bacteroides vulgatus and Bacteroides uniformis were enriched in that of healthy volunteers. A total of 11 OTUs belonging to the genera Enterococcus, Escherichia/Shigella, Klebsiella, Streptococcus and Peptostreptococcus were significantly more abundant in the gut microbiota of CRC patients, and 5 OTUs belonging to the genus Roseburia and other butyrate-producing bacteria of the family Lachnospiraceae were less abundant. Real-time quantitative PCR further validated the significant reduction of butyrate-producing bacteria in the gut microbiota of CRC patients by measuring the copy numbers of butyryl-coenzyme A CoA transferase genes (Mann-Whitney test, P<0.01). Reduction of butyrate producers and increase of opportunistic pathogens may constitute a major structural imbalance of gut microbiota in CRC patients.

Figure 1

(a) Principal component analysis (PCA) scores plot based on the relative abundance of OTUs (98% similarity level). (b) PCoA (principal coordinate analysis) plot based on the weighted UniFrac metric. Blue circles represent healthy volunteers; red circles represent CRC patients.

Figure 2

Biplot of the RDA (redundancy analysis) based on the relative abundance of OTUs (98% similarity level). Constrained explanatory variables are indicated by red triangles: H represents healthy and C represents diseased. OTUs that have at least 5% of the variability in their values explained by the canonical axis are indicated by blue arrows. Upper right shows P-value of Monte Carlo Permutation Test. B, Bacteroides; P, Phascolarctobacterium; E, Escherichia; R, Ruminococcus.

Figure 3

Relative abundances of the 48 OTUs identified as key variables for the differentiation of microbiota of healthy volunteers and CRC patients. OTUs shown in black are the ones enriched in healthy volunteers and those in red are the ones enriched in CRC patients. Phylogenetic tree of the OTUs on the left is constructed with neighbor-joining method in ARB. For each OTU, the color intensity in each sample is normalized to represent its relative ratio in all samples. Individuals at the bottom are arranged according to disease status (healthy; stage I, II, III and IV of CRC) and age inside each stage (from youngest to oldest, color code: pink for age 40–49 years; orange for age 50–59 years; blue for age 60–69 years; green for age 70–79 years).