Análisis del microbioma-metiloma y estudio de asociación entre Fusobacterium nucleatum y marcadores moleculares en el cáncer colorrectal

Résumé

Objective: Study of the microbiome and methylome from intestinal tissue samples from CRC patients to establish associations between the CpG islands genetic methylation profile and the genus Fusobacterium relative abundance. Subsequently, an association study between Fusobacterium nucletarum relative quantification and the clinical and molecular characteristics in CRC, as well as the potential usefulness of F. nucletaum relative quantification in the diagnosis/screening of CRC, were carried out. Material and methods: 154 paired samples, 77 tumor tissue (T) and 77 peripheral non-tumor tissue (NT), corresponding to 77 patients with CRC, classified according CRC type (CC, SAC and hmMSI), location. (proximal, distal) and stage according TNM criteria. The microbiome study is carried out using NGS technology (IlluminaMiSeq System®). The CpG islands DNA methylation profile is carried out using the Infinium® HumanMethylation450 BeadChip array and the results are validated by pyrosequencing. MSI status is determined using the MSI Analysis System kit (Promega) and CIMP status with the SALSA® MS-MLPA® Probemix ME042-C1 CIMP technique (MRC-Netherlands). Gene mutations: KRAS, BRAF and PIK3CA are also determined by pyrosequencing. Finally, qPCR were carried out to total bacterial and F. nucletaum quantification. PCA (Principal Component Analysis) method, U Mann-Withey-Wilcoxon test and LDA analysis (“linear discrimination analysis”) were carried out to analize microbiome differences between T and NT samples Pearson correlation analysis between Fusobacterium genus relative abundance and genes methylation level was carried out with the R statscor function using RStudio IDE version 1.0.143. Genes methylation validation study, correlation analysis between F. nucletaum relative quantification and CpG island methylation percentages and for the different markers of interest was carried out using the SPSS version program. 21.0. Results: The diversity analysis of microbial populations between T and NT samples according the location determines specifically significant differences due to T samples of distal location, with less diversity, while there are no microbial diversity differences according to CRC type. The microbiome data PCA shows no differences between cases. Bacteroides, Eubacterium, Fusobacterium and Acinetobacter are the genera with the highest discriminatory power in the predictive model obtained by LDA method favorable to T samples. Only Lachnoclostridium appears as a genus in the predictive model in favor of NT samples. The validation study for the genes methylation level and Fusobacterium relative abundance has confirmed a positive correlation with PLCH1, mirLET7D, EGFR and COBL genes. The F. nucletaum relative cuantification has shown positive correlation only with ZNF788 gene. Were no differences for the quantification of F. nucletaum by sex or location. There are high F. nucletaum relative quantification levels with significant differences for proximal SAC/hmMSI CRC, larger tumors (T3, T4) and advanced stages (II, III, IV). There are also F. nucletaum relative quantification significant differences between MSS and MSI-H, CIMP-H and CIMP-L, KRAS and BRAF mutated states, while they are not found for PIK3CA. The multivariate analysis confirms that MSI status, CRC stage and CRC type are the variables with an significant association with F. nucletaum relative quantification. The F. nucletaum relative quantification has showed a diagnostic efficiency for CRC of 69%. The optimal cut-off point is established at a difference ≤ 10 cycles between F. nucletaum and total bacterial loads obtained by qPCR. The predictive model establishes that F. nucletaum relative quantification is an independent predictor for CRC (OR=2.23; 95% CI [1.51-3.30]) and for SAC/hm-MSI CRC versus CC (OR=2.00, 95%CI [1.22–3.28]). Conclusions: Bacteroides, Eubacterium, Fusobacterium and Acinetobacter are potential bacterial CRC biomarkers. PLCH1, mirLET7D, EGFR and COBL genes have shown a hypermethylated state associated with the abundance of Fusobacterium spp. in T samples, while only ZNF788 CpG methylation percentage has shown positive correlation with F. nucletaum relative quantification. Higher F. nucletaum level is significantly associated with SAC/hmMSI CRC type, advanced stage, MSI-H status, CIMP-H status and BRAF mutation, although multivariate analysis confirms the independent association with SAC/hmMSI CRC, advanced stages and MSI-H status. Due to F. nucletaum relative quantification good specificity, this parameter is proposed as an improving CRC screening/diagnosis candidate as non-invasive techniques, even being able to differentiate between CC and SAC/hmMSI.