First-line treatment with regorafenib in frail metastatic colorectal cancer (mCRC) patients has shown some benefit. To accurately identify such patients before treatment, we studied blood biomarkers and primary tumor molecules. We unveiled serum microRNAs (miRNAs), single-nucleotide polymorphisms (SNPs) in angiogenic-related genes, and Notch 1 expression as biomarkers associated with response or toxicity. MicroRNA array profiling and genotyping of selected SNPs were performed in the blood of fragile mCRC patients treated with regorafenib. Notch 1 and CRC-associated miRNA expression was also analyzed in tumors. High levels of miR-185-5p in serum, rs7993418 in the vascular endothelial growth factor receptor 1 (VEGFR1) gene, and Notch 1 expression in biopsies were associated with a favorable response to treatment. Serum levels of miR-126-3p and miR-152-3p and tumor expression of miR-92a-1-5p were associated with treatment toxicity, particularly interesting in patients exhibiting comorbidities, and high levels of miR-362-3p were associated with asthenia. Additionally, several miRNAs were associated with the presence of metastasis, local recurrence, and peritoneal metastasis. Besides, miRNAs determined in primary tumors were associated with tumor-node-metastasis (TNM) staging. The rs2305948 and rs699947 SNPs in VEGFR2 and VEGFA, respectively, were markers of poor prognosis correlating with locoregional relapse, a higher N stage, and metastatic shedding. In conclusion, VEGF and VEGFR SNPs, miRNAs, and Notch 1 levels are potential useful biomarkers for the management of advanced CRC under regorafenib treatment.

Cyclooxygenase 2 (COX2) has been implicated in cancer development and metastasis. We have identified several COX2-regulated inflammation-related genes in human colorectal cancer cells and shown that some of them play important roles in tumor progression. In this work, we have studied the COX2-regulated genes in the mouse colorectal cancer cell line CT26, to find that many are also regulated by COX2 over-expression. On the other hand, we generated a CT26 cell line expressing Gfp and Luciferase, to study tumor growth and metastasis in immunocompetent Balb/c mice. We then collected solid tissue, and blood samples, from healthy and tumor-bearing mice. Using the Parsortix® cell separation system and taking advantage of the fact that the tumor cells expressed Gfp, we were able to identify circulating tumor cells (CTCs) in some of the mice. We compared the mRNA expression levels of Ptgs2 and effector genes in the samples obtained from tumor-bearing or healthy mice, namely, tumor or healthy colon, Ficoll purified buffy coat, and Parsortix-isolated cells to find different patterns between healthy, tumor-bearing mice with or without CTCs. Although for genes like Il15 we did not observe any difference between healthy and tumor-bearing mice in Ficoll or Parsortix samples; others, such as Egr1, Zc3h12a, Klf4, or Nfat5, allowed distinguishing for cancer or CTC presence. Gene expression analysis in Ficoll or Parsortix processed samples, after liquid biopsy, may offer valuable diagnostic and prognostic information and thus should be further studied.