Analysis Of Mucosa-associated Microbiota In Colorectal Tumor

Background:Colorectal tumor is one of the common digestive tumor around the world,with the colorectal cancer to be the third most common cancer and the fourth most common cancer cause of death globally.The colorectal cancer also accounts for a relatively high proportion for the cancer cause of death in China,and with its increasing prevalence every year colorectal tumor poses severe threats to people's health.Colorectal tumor is closed linked to genetic predisposition,so-called western lifestyle and environmental factors.The gut microbiota is proved to be connected with colorectal tumor and there are two separate hypotheses considering the relationship between the gut flora and colorectal tumor:"Alpha-Bugs" theory and "driver-passenger"modelAim:To analysis the mucosa-associated gut microbiota of different stages of colorectal tumor and find out the differences and biomarkers of microbiota with colorectal tumor.Method:The data of microbiota sequences in this research came from a published paper in the journal of Nature Communication on October,2015 and was downloaded from SRA in NCBI.Data was public.After de-noising,the sequences were clustered into OTU using the UPARSE with 95%similarity cut-off and the chimeras were detected and removed using UCHIME.We selected the representative sequence of each OTU and mapped them with Greengene database in order to get the taxonomy annotation for each OTU,then assigned to phylotypes with different taxonomic levels.a diversity index(including chaol index,simpson index and shannon index),PCoA,Permutational Multivariate Analysis of Variance(PERMANOVA)and clustering analysis were carried out in R,? diversity(including Bray-Curtis distances,weighted UniFrac distances and unweighted UniFrac distances)were calculated in QIIME.a diversity and the relative abundances of bacterial taxa of different taxonomic levels between groups were compared using Kruskal-Wallis rank test and Wilcoxon rank test and p values were adjusted with Benjamini-Hochberg and considered significant if the adjusted-p<0.1.LEfSe was used to identify the biomarkers among groups.Outcomes:A total 160 patients were recruited in this study,with tumor-free colon(n=61),with confirmed histology of colorectal adenomas(n=47),and with invasive adenocarcinomas(n=52),which we defined as normal,adenoma and adenocarcinoma,respectively.There are no significant differences for clinical parameters among three groups but ages(60.13±5.99,67.32±8.80,67.85 ± 13.18,p<0.001).A total 1012335 reads were retrieved after sequence processing,with 6327 reads per sample.There were 338391 reads and 5547 reads per sample in the normal group,327500 reads and 6968 reads per sample in the adenoma group,346444 reads and 6662 reads per sample in the adenocarcinoma group.A total 605 OTUs and 63 OTUs per sample were found with clustering using 95%similarity,and there were 445 OTUs in the normal group,417 OTUs in the adenoma group,482 OTUs in the adenocarcinoma group.We found out that there were significant differences in the chaol index among three groups(p=0.0006),but no significant differences in either the simpson index or the shannon index(simpson,p=0.22;shannon,p=0.2).PCoA showed that the adenocarcinoma group could be separated from the other two groups,while the samples in the normal group and adenoma group were too close to be separated,and the overlap could be observed among three groups.PERMANOVA analyses indicated that there were significant differences between the composition of microbiota in the three groups.Bacterial phylotypes were assigned to each OTU into all taxonomic levels(phylum,class,order,family,genus and species)and most of OTUs could be annotated into genus level.12 phyla were found among the three groups and the majority belonged to Firmicutes,Bacteroides,Actinobacteria,Proteobacteria.We got 94 families and the taxa with higher average relative abundance were Enterobacteriaceae,Bacteroidaceae,Ruminococcaceae and Fusobacteriaceae.We also got 148 genera,and the taxa with higher average relative abundance were Bacteroides,Escherichia,Fusobacterium and Streptococcus.There were no statistical differences among all taxa in the whole 6 taxonomic levels between the normal and adenoma groups using Wilcoxon rank test.When comparing the normal and adenocarcinoma group,the phyla of Actinobacteria(p=0.01)and Proteobacteria(p=0.001)were significant higher in the normal group,and the phyla of Firmicutes(p=0.07)and Fusobacteria(p=0.008)were significant lower in the normal group.There were also 16 genera with statistical difference in the above two groups,among which the Campylobacter,Dialister,Eikenella,Fusobacterium,Lactobacillus,Leptotrichia,Mogibacterium,Parvimonas,Peptostreptococcus and Prevotella were higher in the adenocarcinoma group,while the Acidomonas,Blautia,Escherichia,Faecalibacterium,Pseudomonas and Sphingomonas were lower in the same group.21 genera were considered to be statistical difference between the adenoma and adenocarcinoma group and there were Acidomonas,Acinetobacter,Bulleidia,Campylobacter,Desulfovibrio,Dialister,Enhydrobacter,Escherichia,Fusobacterium,Leptotrichia,Mogibacterium,Odoribacter,Oscillospira,Parvimonas,Peptostreptococcus,Porphyromonas,Prevotella,Pseudomonas,Sphingomonas,Staphylococcus,Treponema.We used the LEfSe to figure out the biomarkers among different stages through the development of colorectal tumor.The biomarkers of the normal group constituted of Actinobacteria,Actinomycetales,Moraxellaceae and Sphingomonas.The biomarkers of the adenoma group included Alphaproteobacteria,Bacillales,Escherichia coli,Proteobacteria,Pseudomonas and Staphylococcus.The biomarkers in the adenocarcinoma group comprised of Bacilli,Bacteroides_fragilis,Firmicutes,Fusobacterium,Streptococcaceae,Parvimonas and Prevotella.We performed the cluster analysis using the unweighted UniFrac distance matrix of all samples and noticed that not all samples from the same group could be clustered into one cluster.Conclusions:We demonstrated the differences in mucosa-associated gut microbiota among the distinct stages of the development of colorectal tumor with our work.The composition of microbiota in the adenocarcinoma group was significant different from the normal and adenoma group,with higher number in microbiota of potential tumorigenesis like Bacteroides fragilis and Fusobacterium,and lower number in microbiota of possible antitumor like Blautia and Faecalibacterium prausnitzii.The microbiota composition between the normal and adenoma group,however,was so close that it was hard to distinguish using PCoA,which meant there maybe some confounding factors in our study.We selected the biomarkers of three groups and there were completely disparate,which indicated the mucosa-associated microbiota changed heavily among the development of colorectal tumor and microbiota could play a vital role in the colorectal tumorigenesis.The relative abundance of Escherichia coli was higher in adenoma group compared with adenocarcinoma group,we proposed that the Escherichia coli exerted influence on the early phase of colorectal tumorigenesis and eventually substituted by other more suitable microbiota around the development of colorectal tumor,which illuminated the hypothesis of "driver-passenger" model.