Investigate tumour heterogeneity and genetic pathways in colorectal cancer

Background: Colorectal cancer is a major global public health problem that has predominantly been considered a genetic disease following a precise series of molecular events. These are characterized by sequential accumulation of genetic and epigenetic alteration in several oncogenes and tumour suppressor genes. Understanding of the genetic mechanisms that explain the initiation and evolution of colorectal cancer are key to improving risk prediction, prognostication and treatment. Aims: The aim of this study was to understand the basic principles of the molecular biology of colorectal cancer based on genomic, transcriptomic, and proteomic profiles analyses. Methods: DNA extracted from 147 formalin fixed paraffin embedded (FFPE) samples from 83 patients with colorectal cancer (first cohort) including (83 primary colorectal cancer (CRC), 22 matched liver metastases, 25 matched biopsies and 17 normal colon tissue) were screened for mutation in 26 genes (Trusight tumour kit) using Targeted Next Generation Sequencing. Additionally, exonuclease domain region of POLE and POLD1 were also screened using High Resolution Melting and Sanger Sequencing methods. These data used to: Investigate mutation profiles of CRCs among 83 primary samples. Investigate the difference between chromosomal instable-CRC (46 primary sample) and chromosomal and microsatellite stable-CRC (35 primary sample). Compare mutations in 26 genes of 25 paired biopsy samples and corresponding resection specimens. Investigate genetic discrepancies between 22 primary colorectal cancers and their respective metastases. Additionally, expression of a panel of six miRNAs (miR-20a, miR-21, miR-29a, miR-31, miR-92a and miR-224) was measured using RT-qPCR and protein expression of 20 genes was measured using Reverse Phase Protein Array (RPPA). In a second cohort including 81 primary CRC and their matched normal samples, expression of the six miRNAs and mRNA of six genes (SMAD4, PTEN, BCL2, TGFBRII, KLF4 and RASA1) targeted by the six miRNAs were measured using RT-qPCR. Additionally, expression of proteins of the targeted six genes was also measured using immunohistochemistry (IHC). Cell-free DNA (cfDNA) extracted from 16 blood samples (third cohort), which were taken from 5 CRC patients at different time points (pre- and post-surgery) were screened for mutations in KRAS, TP53, PTEN, SMAD4, BRAF and PIK3CA genes. Additionally, expression of the six miRNAs was measured using RTqPCR. Results: In the first part; investigating mutation profile of the first cohort 83 CRC showed high frequency of mutation in TP53 (75%), APC (57%) and KRAS (53%). Approximately 93% CRCs have mutation in at least one of APC/TP53/KRAS/BRAF/SMAD4/PIK3CA/PTEN/FBXW7 genes. Moreover, mutations were found in the exonuclease domain regions of POLE in 9.6% and POLD1 in 2.4%. Regarding biopsy vs resection, the mutant allele frequency was 1.03-fold higher in resection specimens than biopsies and there was no mutation in the biopsy specimens that were not seen in the resection specimens. In the second part; Comparison of CIN-CRC vs MACS-CRC, which were included in the first cohort CRCs showed similar mutation frequencies of mutation in all 28 genes except KRAS (41%CIN vs 68%MACS), POLE (15%CIN vs 2%MACS), GNAS (0%CIN vs 11%MACS). Statistically there was a significant difference (each p=0.01) which was lost following multiple testing correction. In the third part; comparison of primary CRC vs matched metastasis showed that a total of 61 non-synonymous somatic variations in 12 genes were found in primary 22 specimens whereas 60 were found in metastasis cases. The mutant allele frequency was 1.01-fold higher in primary than metastasis CRCs. Evaluated expression levels of six miRNAs and protein expression of other 20 genes, did not show any significant differences between primary CRC and matched metastasis. In the fourth part; Expression of the six miRNAs and mRNA and protein of the six targeted genes were tested in the second cohort 81 samples. Statistical analysis revealed significant increase in the expression level of miR-20a (p=0.04), miR-21 (p=0.01) miR-29a (p=0.03) and miR-31 (p=0.01) and decrease in the mRNA expression level of TGFBRII and RASA1 in tumour samples compared to normal tissues. IHC staining showed low expression level of SMAD4 in 51 (63%), PTEN 67 (83%), TGFBRII 65 (80%), RASA1 61 (75%) BCL2 47 (58%) and high expression of KLF4 36 (44%). High miR-21 and miR-224 expression were associated with low expression of TGFBRII. In addition, over expression of both miR-29a and miR-31 inversely correlated with RASA1. In the fifth part; Mutation in the cfDNA was detected in 5 cases. Two of these showed a loss of the mutant signal post-operatively. Whereas the mutant signal was persistent in the rest 3 of the cases for all the samples taken post-operatively. Although miRNAs expression was fluctuated between these time points, paired test showed a non-significant difference when comparing pre- and post-surgical miRNAs level. However, level of the cfmiRNAs is changed by more than 2 folds (upregulated) in the day of surgery compared to normal plasma as follow, miR- 20a in 1/5 (20%), miR-21 in 4/5 (80%), miR-29a in 3/5 (60%), miR-92a in 4/5 (80%) and miR-224 in 2/5 (40%). Conclusion: Investigation profiles of CRCs from both cohorts indicated that, different mutated genes and upregulated miRNAs, which are involved in different signalling pathways, may have roles in CRC carcinogenesis. Significant difference was neither noticed between MACS and CIN group and nor between primary and metastasis tumour. miRNAs from tissues and cfmiRNAs from plasma, can differentiate CRC from healthy group and have potential clinical value in early CRC detection. In addition to the resection specimens, the study found that it is acceptable to use biopsy material for predictive testing and cfNAs can be used for a variety of clinical and investigational applications.