Abstract
Since its discovery in 1948, circulating DNA (cirDNA) analysis has demonstrated its clinical potential in oncology.My first thesis objective was to study the clonal heterogeneity of tumors from metastatic colorectal cancer (mCRC) patients by cirDNA analysis during their therapeutic management. Only patients with non-mutant RAS tumors can benefit from anti-EGFR therapies. Prior to anti-EGFR therapy, cirDNA analysis showed the detection of higher RAS/BRAF mutations (76%, n=119) than tumor tissue analysis (55%, n=89) as well as a high proportion of co-occurring mutations (45%) in mCRC patients (KPLEX2 study). We showed that the median data turn-around time was 12 days for tumor tissue analysis (n=83; 3-273 days) versus only 2 days for cirDNA analysis (n=121; 0-10 days). These results indicated that cirDNA could replace tumor tissue analysis in clinical practice.In the retrospective KPLEXR study, longitudinal analysis of mCRC patients treated with FOLFOX-dasatinib with or without cetuximab showed that 98% of treatment-refractory patients (n=42) were RAS/BRAF mutant at the end of the treatment. This result suggested that RAS/BRAF mutations induce resistance to anti-EGFR therapy. Quantitative and longitudinal analysis of cirDNA closely mirrored variations in blood CEA levels and tumor mass as observed by CT-scan. Moreover, 21% of RAS/BRAF mutations were found with allelic frequencies lower than 0.1% revealing the need of the use of an ultrasensitive method to detect point mutations before and during anti-EGFR treatments. These results highlighted the theranostic potential of cirDNA during anti-EGFR therapy.Refractory patient tumors to all standard therapies can be treated with regorafenib. However, there is currently no biomarker to predict benefit or resistance to regorafenib. Inthe prospective TEXCAN study we observed that the RAS/BRAF mutant status of heavily pretreated mCRC patients refractory to standard therapies does not influence the survival of these patients treated with regorafenib. On the other hand, we showed that it is possible to identify prior to treatment those patients who will benefit clinically from the drug. Indeed, we were able to establish thresholds for total cirDNA concentration (26 ng/mL plasma), mutant cirDNA concentration (2 ng/mL plasma) and allelic frequency (6%) that would predict which patients will have better overall survival. The cirDNA could be a biomarker for predicting patient tumor resistance or tolerance to regorafenib.It has long been recognized that apoptosis is the most relevant mechanism of cirDNA release since many studies have shown that cirDNA has a size profile around 150-180 bp. The second objective of my thesis was to evaluate the association between NETs (neutrophil extracellular traps) formation and cirDNA production in mCRC patients. We show that a significant fraction of cirDNA in mCRC patients is derived from the NETosis mecanism. This work redefines existing paradigms on the mechanisms of cirDNA release. NET degradation and resultant cirDNA release could explain the large variation of allelic frequency in the blood of mCRC patients. However, the full set of enzymes leading to the production of cirDNA by the in vivo degradation of NETs remains to be demonstrated.