Metagenomic analysis of colonic tissue and stool microbiome in patients with colorectal cancer in a South Asian population

Abstract

Background: The gut microbiome is thought to play an important role in the development of colorectal cancer (CRC). However, as the gut microbiome varies widely based on diet, we sought to investigate the gut microbiome changes in patients with CRC in a South Asian population.

Methods: The gut microbiome was assessed by 16s metagenomic sequencing targeting the V4 hypervariable region of the bacterial 16S rRNA in stool samples (n = 112) and colonic tissue (n = 36) in 112 individuals. The cohort comprised of individuals with CRC (n = 24), premalignant lesions (n = 10), healthy individuals (n = 50) and in those with diabetes (n = 28).

Results: Overall, the relative abundances of genus Fusobacterium (p < 0.001), Acinetobacter (p < 0.001), Escherichia-Shigella (p < 0.05) were significantly higher in gut tissue, while Romboutsia (p < 0.01) and Prevotella (p < 0.05) were significantly higher in stool samples. Bacteroides and Fusobacterium were the most abundant genera found in stool samples in patients with CRC. Patients with pre-malignant lesions had significantly high abundances of Christensenellaceae, Enterobacteriaceae, Mollicutes and Ruminococcaceae (p < 0.001) compared to patients with CRC, and healthy individuals. Romboutsia was significantly more abundant (p < 0.01) in stool samples in healthy individuals compared to those with CRC and diabetes.

Conclusion: Despite marked differences in the Sri Lankan diet compared to the typical Western diet, Bacteroides and Fusobacterium species were the most abundant in those with CRC, with Prevotella species, being most abundant in many individuals. We believe these results pave the way for possible dietary interventions for prevention of CRC in the South Asian population.

Fig. 1

Species accumulation curves of healthy and CRC patients for each sample type. Healthy stool samples demonstrated the highest OTU discovery, whereas healthy gut tissue samples showed the least OTU discovery

Fig. 2

Comparison of the faecal and tissue microbiome (alpha diversity) in healthy individuals, patients with CRC, with premalignant lesions and with diabetes. Alpha diversity in bacterial communities was calculated and presented by the number of observed OTUs and Shannon index in tissue samples (T) and stool samples (S), in the in those with CRC (C), healthy individuals (N), those with diabetes mellitus (M) and in those with premalignant lesions (PM). P values were calculated with Kruskal-Wallis test

Fig. 3

Comparison of the beta diversity of faecal and tissue microbiome in healthy individuals, patients with CRC, with premalignant lesions and with diabetes. A) The Principal Coordinates Analysis (PCoA) using weighted UniFrac distances indicating significant clustering between the stool subgroups of those with CRC (C), those with diabetes mellitus (M), healthy individuals (N) and those with premalignant leisions (PM), (p = 0.02). B) The analysis of similarities test (ANOSIM) for samples obtained from colonic tissue (T) and stools (S), in those with CRC (C), healthy individuals (N), those with diabetes mellitus (M) and in those with premalignant lesions (PM) was carried out to assess the differences in microbial composition. The microbial composition between gut tissue subgroups was consistent while stool samples exhibit significant dissimilarity (p < 0.0001) between the subgroups

Fig. 4

The relative abundance of Phyla of tissue samples of healthy individuals in comparison to patients with CRC. The relative abundance of different bacterial Phyla in tissue samples in healthy individuals (n = 18) and in patients with CRC (n = 18) was assessed. Phylum Firmicutes, Bacteroidetes, Proteobacteria and Fusobacteria were found to be highly abundant across the tissue samples

Fig. 5

The relative abundance of phyla in stool samples of patients with CRC, healthy individuals and in those with premalignant lesions. A) The differences in the relative abundance of different bacterial phyla between stool samples of healthy individuals (n = 50), those with CRC (n = 24), and in those with a pre-malignant lesion (n = 10) were compared. Phylum Firmicutes, Bacteroidetes, Proteobacteria, Actinobacteria and Fusobacteria were dominant across the stool subgroups. B) The differences in the abundance of the phyla Epsilonbacteraeota and Elusimicrobia were compared in stools between those with CRC (C), those with diabetes mellitus (M), healthy individuals (N) and those with premalignant lesions (PM). P-values were calculated using Dunn’s multiple comparisons test with Benjamini–Hochberg false discovery correction

Fig. 6

Significantly different genera in stool subgroups. The relative abundance of bacterial genera were compared in stool samples (S) in healthy individuals (S-N), pre-malignant (S-PM), patients with CRC (S-C) and patients with diabetes (S-M) using Dunn’s multiple comparisons test with Benjamini–Hochberg false discovery correction. Significantly different abundances of Christensenellaceae, Enterobacteriaceae, Mollicutes and Ruminococcaceae in pre-malignant stool were observed compared to other subgroups. Genus Romboutsia had significantly different abundance in healthy stool compared to CRC and diabetes subgroups

Fig. 7

The relative abundance of phyla in stool samples in healthy individuals and in patients with diabetes. The differences in the relative abundance of different bacterial phyla between stool samples of healthy individuals (S.N) (n = 50) and in patients with diabetes (S-M), were compared. Firmicutes, Bacteroidetes and Actinobacteria were found to be dominant in stool samples of individuals with diabetes

Fig. 8

A heatmap showing relative abundance of genera in tissue and stool samples. The type of sample and the subgroup are indicated with a color key. Dendrograms were produced with the (UPGMA) method based on Bray-Curtis distance