Dysregulation of saliva and fecal microbiota as novel biomarkers of colorectal cancer Authors: Jiamei Rong, Xiaocui Chen, Zhangqin Li, Bona Li, Yang Sun, Yinglei Miao Published: 2024 ABSTRACT The aim of this study was to investigate the biomarkers of salivary and fecal microbiota in Colorectal cancer (CRC). Initially, the study scrutinized the microbial community composition disparities among groups. Utilizing Lasso analysis, it sifted through operational taxonomic units (OTUs) to pinpoint distinctive features. Subsequently, by intersecting feature OTUs across groups, it curated a set of core-shared OTUs and devised a corresponding network. Concluding with functional enrichment analysis, the research delved into the divergent biological functions of these microbial communities within the studied groups. Analysis revealed higher bacterial diversity in saliva compared to feces, with distinct differences at both phylum and genus levels. Feces primarily contained Firmicutes, while saliva was dominated by Bacteroidetes and Proteobacteria. Notably, Escherichia-Shigella and Fusobacterium in feces and Streptococcus in saliva showed increasing abundance from average to adenoma to colorectal cancer. Specific dominant flora was identified within and between groups, including CRC and adenomas across different stages. Seventeen core shared OTUs were identified, and networks of shared OTUs were constructed for each group. Functional enrichment analysis highlighted distinct microbial community functions among the groups. This study’s findings on characteristic OTUs in saliva and fecal samples offer valuable insights for distinguishing between healthy individuals, adenoma patients, and those with colorectal cancer. This study identified distinctive OTUs in saliva and feces to distinguish between healthy individuals, adenoma patients, and those with CRC, offering a valuable diagnostic reference. Introduction Colorectal cancer (CRC) is one of the three common malignant tumors (1). It is estimated that the global CRC burden will increase by 60% in 2030, with more than 2.2 million new cases and nearly 1.1 million CRC deaths (2). At present, it is generally believed that the occurrence and development of CRC follows the sequence of “normal mucosa-adenoma-carcinoma”, and about 80%-95% of CRC develops from adenomatous polyp, which generally takes 5-10 years or even longer, providing an important time window for early diagnosis and clinical intervention (3). CRC screening methods in our population are not uniform and have their own limitations (4, 5). Therefore, in-depth study of the imbalance of intestinal microecology during the occurrence and development of CRC, and mining for targeted and sensitive biomarkers to monitor early CRC can lay an important theoretical basis and practical scheme for improving early diagnosis and treatment management strategies, which are of great significance for improving the quality of life of patients at high risk of CRC. More and more evidence shows that chronic inflammation, host genetic susceptibility and environmental factors are related to the progress of CRC (6). Environmental factors are very important to the composition and function of intestinal microorganisms, and changes in them cause alterations in host gene expression, metabolic regulation and local and systemic immune responses, thus affecting the development of cancer (7). In recent years, more and more studies have reported that the composition and diversity of intestinal flora play an important role in the occurrence and development of chronic liver disease, irritable bowel syndrome, inflammatory bowel disease, CRC and other multi-system diseases (8–10). Imbalance of intestinal flora is considered a potentially important cause of CRC (11). Different from the causal role of Helicobacter pylori in gastric cancer, the specific microorganism that causes CRC has not been determined. Therefore, the repeated identification and verification of cancer-causing microorganisms in patients with diseases is of great significance for understanding its role in the pathogenesis of CRC and finding the best treatment for CRC. The gastrointestinal tract begins in the mouth, and the oral microflora is the second largest microflora after the gastrointestinal tract because of its complex composition. In recent years, it has been found that CRC is not only related to intestinal microflora, but also closely related to in oral microflora (12, 13). However, few studies pay attention to the differences of oral microflora between adenoma and CRC patients and healthy people. There is evidence that there may be an “oral-intestinal axis” in the pathogenesis of CRC (14), that poor oral hygiene may cause cancer by changing the number of specific oral bacteria, and that periodontal disease may increase systemic inflammation, leading to immune disorders and changes in the intestinal flora, which may affect the occurrence of CRC (15). In this topic, we explore whether there is a potential microbial basis between oral microorganisms and CRC, and further explore this potential oral-intestinal axis to find new non-invasive biomarkers of CRC. At present, the role of intestinal microbiota in the occurrence and development of intestinal cancer has become more and more clear, but the details of how oral microbiota changes intestinal microbiota and affects the occurrence of intestinal cancer are still unclear. In this study, 16SrRNA high-throughput sequencing was used to analyze the microbial diversity and species abundance in normal, adenoma and CRC in saliva and feces, with a large sample size and strict exclusion criteria. Linking the two provides a theoretical basis for the early screening and pathogenesis of CRC from the perspective of salivary and intestinal microecology. Materials and methods Patient selection Patients who underwent colonoscopy in the Digestive Endoscopy Laboratory of the First Affiliated Hospital of Kunming Medical University from October 2020 to September 2022 and were pathologically diagnosed with adenoma and bowel cancer were selected, while healthy volunteers matching the age, gender and body mass index of patients with adenoma and bowel cancer were recruited as healthy control groups. A total of 122 normal samples, 122 adenoma samples and 117 colorectal cancer (CRC) samples were enrolled in this study. Saliva and fecal samples were collected separately for each sample. These samples were classified into 8 groups, specifically: group A: fecal of the normal group, group B: fecal of the adenoma group, group C: fecal of the CRC group, group D: saliva of the normal group, group E: saliva of the adenoma group and group H: saliva of the CRC group. The studies involving human participants were reviewed and approved by [the First Affiliated Hospital of Kunming Medical University]. The patients provided their written informed consent to participate in this study. All the procedures were conducted in accordance with the “Declaration of Helsinki”. Sample collection Saliva collection No brushing of teeth after getting up in the morning and fasting 2 hours before sampling. A sterile, enzyme-free collection tube was left in the mouth for at least 1 min and 2 mL of non-irritating saliva. After collection, freeze in a -80°C freezer within 2 hours. Fecal collection Fresh stool was collected and divided into about 50-100 mg into a sterile centrifuge tube and frozen in a -80°C refrigerator within 2 hours. DNA extraction and 16S sequencing The CTAB/SDS method was used to extract the total genome DNA in samples. PCR amplification was performed with a diluted genomic DNA template. Following the manufacturer’s recommendations, sequencing libraries were generated with NEBNext®Ultra™ IIDNA Library Prep Kit (Cat No. E7645). The library quality was evaluated on the Qubit@ 2.0 Fluorometer (Thermo Scientific) and Agilent Bioanalyzer 2100 system. Finally, the library was sequenced on an Illumina NovaSeq platform and 250 bp paired-end reads were generated. Operational taxonomic units distribution and alpha diversification analysis The valid sequence clusters of the samples were grouped into OTUs based on the sequencing data of 16S rRNA of each sample with a similarity threshold of > 98%. OTUs with relative abundance less than 0.01% in the sample were excluded, and the remaining OTUs were utilized for follow-up analysis. Then, an alpha diversity analysis was carried out, and the rarefaction curves and box plots were plotted. Beta diversity analysis The PCoA clustering plots based on all samples were utilized to illustrate the differences between the different subgroups in the fecal group and the saliva group. Subsequently, within-group and inter-group similarity analyses were performed separately for the fecal and saliva groups using ANOSIM analysis. Analysis of species composition diversity Furthermore, we observed differences in microbial community composition between the six groups at the phylum and genus level. The phylum horizontal bar stack (TOP10) and analysis of differences box-plot was plotted. Meanwhile, genus horizontal bar stack (TOP12) and analysis of differences box-plot was also plotted. Analysis of flora differences The Linear Discriminant Analysis (LDA) Effect Size analysis was carried out to find out which OTUs were causing the differences in the communities according to LDA score > 2 and p < 0.05. The analysis was performed between groups A and B, A and C, B and C, D and E, D and H, and E and H. Of these, groups A and B and D and E were analyzed to identify the specific dominant flora associated with adenomas in feces and saliva. The groups A and C and D and H were analyzed to determine the specific dominant groups of bacteria associated with CRC. The groups B and C and E and H were analyzed to determine the specific dominant groups of bacteria associated with CRC and adenomas. Fecal and saliva group samples of adenoma and early CRC were analyzed to identify specific dominant flora associated with adenoma and early CRC. Samples from the fecal and saliva groups of early and advanced CRC were analyzed to identify the specific dominant flora associated with early (stage I-II) and later stages (Stage III-IV) of CRC. Machine learning filtering feature OTUs The data of groups A and B, A and C, B and C, D and E, D and H, and E and H were randomly classified into training and validation sets in the ratio of 5:5. The Least Absolute Shrinkage and Selection Operator (LASSO) analysis was performed with the glmnet package (16) to build a diagnostic model and to filter the feature OTUs. The specific groupings were as follows: 244 cases each in groups A and B, and D and E, including 122 cases in the training set (A: 61, B: 61) and 122 cases in the validation set (A: 61, B: 61); 239 cases each in groups A and C, and D and H, including 119 cases in the training set (A: 62, C: 57) and 120 cases in the validation set (A: 60, C: 60); 239 cases each in groups B and C, and E and H, including 119 cases in the training set (A: 62, B: 57) and 120 cases in the validation set (A: 60, B: 60). Then, the AUC value of the receiver operating characteristic (ROC) curve was computed to assess the predictive accuracy of the model. Lastly, POD difference analysis was performed. Network analysis of shared OUT The OTUs that were expressed in at least 30% of the 6 groups of samples and had more than 300 reads in a single sample were identified as core-shared OTUs. Afterwards, the Spearman correlation coefficients between the core-shared OUTs were computed via the psych package and the core-shared OUT network was created. Analysis of flora and clinical features Pearson correlation coefficients were computed for different clinical features and microbial community genus levels. The top 20 genera with the highest relative abundance were selected for analysis, and the p < 0.05 relationship pairs were presented. Thereafter, statistical analyses of differences were performed for these 20 genera in different clinical subgroups (cancer sites and stage). Functional enrichment analysis For exploring the pathways involved in inter-group differences in microbial communities, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was performed via PICRUSt2 (17). The analysis was performed between groups A and B, A and C, B and C, D and E, D and H, and E and H. Statistical analysis All bioinformatics analyses were carried out in R language. The data of different groups were compared by Wilcoxon test. Results A total of 1567 common core OTUs were identified for alpha diversification analysis A total of 1675 common core OTUs were found in the 6 groups, of which 15 were OTUs specific to group A, 22 to group B, 38 to group C, 7 to group D, 6 to group E, and 12 to group H ( Figure 1A ). The rarefaction curves indicated that the sequencing data were sufficient to reflect the diversity and abundance of species in the samples ( Figure 1B ). And the diversity index of the saliva group was greater than that of the fecal group ( Figure 1C ). Alpha analysis of the indices revealed that the p-values for the observed OTU index, evenness index and PD index were less than 0.05 in the six groups ( Figures 1D–G ). The evenness index was greater in groups B and C than in groups E and F respectively, while the opposite was true for the observed OTU index and PD index ( Figures 1E, F ). The results of the within-group analysis of variance revealed that the observed OTU and PD indices were significantly higher in group C than in group A ( Figures 1E, F ), and that there were significant differences in the observed otu indices between groups B and C ( Figure 1E ). Meanwhile, there was a significant difference in the observed OTU index between group E and group H ( Figure 1E ). Species abundance and bacterial diversity in six groups. (A)Venn diagram shows the common or unique OTUs in different groups of samples. (B) The dilution curve of observed_features diversity index shows that the sequencing of the sample is enough to reflect the species diversity in the sample. (C) The diversity index of saliva group was greater than that of feces group.α-diversity is indicated by the Shannon (D), observed OTU (E), Evenness (F) and PD diversity (G) indices. A- healthy group feces, B- adenoma group feces, C- colon cancer group feces, D- healthy group saliva, E- adenoma group saliva, H- colon cancer group saliva, Core- common OTUs. ns, no significance; *: p < 0.05; **: p < 0.01; ***: p < 0.001; ****: p < 0.0001. Beta diversity analysis revealed differences between different subgroups of the stool group and the saliva group The differences between the different subgroups of the fecal and saliva groups were shown in Figures 2A, B . The results of the similarity analysis indicated that there were significant differences within (C VS B, C VS A and A VS B) and between groups in the fecal group ( Table 1 ). Similar results were seen in the saliva group ( Table 2 ). Principal Coordinate Analysis (PCoA) of saliva and fecal microbiota. β diversity can reflect the overall differences between different groups of samples. There are significant differences in β diversity between fecal (A) and saliva (B) among the three groups. ANOSIM analysis showed that the difference between fecal and saliva groups reached a significant level, and the P values in each group were less than 0.05. Table 1The results of ANOSIM in the fecal group.GroupRP-valueA_VS_B0.0110.002A_VS_C0.0140.001B_VS_C0.010.005ALL0.0150.001R: correlation coefficient. -1