During cellular turnover, cells release intracytoplasmic and nuclear contents which can be detected in culture media52. This process has allowed various in vitro models to be used to study cfDNA biology2,4,7,52 and to assess drug response4,29,30. Using this model, we previously reported cfDNA release from UM cells into conditioned media21. We further demonstrated the presence of dose-dependent cfDNA release during drug treatments, including beta-adrenoceptors-blocker and mifepristone on UM and skin melanoma cells, both potential adjuvant antitumor drugs29,30. In the present study, we used a panel of cancer cells (colorectal cancer, lung cancer, and UM cells) and different cytotoxic agents that mimic anticancer therapeutic agents to study changes in the kinetics and fragmentation of shed cfDNA. We brought evidence that the levels and fragmentation patterns of released cancer cfDNA depend on the cytotoxic treatment. These findings suggest that, depending on the anticancer therapy applied, we could predict the characteristics of the released cfDNA, which may guide the choice of suitable analytical method of these DNA fragments.
Circulating tumor DNA (ctDNA)-based liquid biopsy has emerged as an important tool to detect and monitor cancer53 that allows for the real-time analysis of tumor cells or their products using bodily fluids14. Various studies have reported that ctDNA circulating in the bloodstream reflects tumor burden from primary and metastatic sites and heterogeneity by showing a spectrum of mutations that correlate with tumor size54. As a result, ctDNA seems to be a highly specific biomarker for non-invasive monitoring, and its fluctuations may help to guide personalized therapeutic decision-making55. Although ctDNA as a biomarker has been widely studied in clinical reports56, our understanding of its etiology remains limited. Important knowledge gaps remain on how cfDNA release is influenced by anti-cancer treatments, which has major implications for clinical interpretation of ctDNA levels in patients undergoing treatment. We undertook the present study to address this gap.
Although the biology of cfDNA release is not fully understood, studies have generally concluded that cfDNA is mainly a product of apoptosis6. To get more insights on the mechanisms governing cfDNA generation, we applied a battery of cytotoxic conditions to cancer cells that are similar to key effector mechanisms of several anticancer therapeutic drugs, including agents that are inducers of apoptosis, necrosis, anti-proliferative (i.e. cell cycle blocker), and senescence. In summary, we found that the levels and fragment sizes of mutant cfDNA were associated with the mechanism of cytotoxicity. In agreement with our findings, it has been reported that cytostatic treatment (i.e. cell cycle progression interference) alters cfDNA kinetics. More specifically, a cfDNA increase was observed upon the cytostatic treatment. This increase might be associated with the preparation for cell division4. Tumor cells that fail a transition from synthesis (S phase) to mitosis (M phases) within the cell cycle might undergo apoptosis and/or necrosis, processes that are known as a source of cfDNA4.
Radiotherapy is one of the most common cancer treatments and is the standard of care to treat UM lesions49,57. Irradiated cells undergo mitotic catastrophe, a form of cell stress that precedes apoptosis, necrosis, senescence, and autophagy50. However, little is known about the effect of irradiation in cfDNA release. A previous study conducted in head and neck squamous cell carcinoma and non-small lung cancer cells reported that cfDNA is modulated by a treatment-induced cell senescence58. Other reports identified a correlation between cfDNA levels and clonogenic survival, suggesting that cfDNA accurately quantified cell survival upon irradiation59. These observations are in line with our conclusions from the present study. We also noticed that the levels of cfDNA in the different cancer cells following treatments varied substantially. This variation may be due to the carried specific mutations in the respective cancer cell type. For instance, in irradiated cells, we observed a discrepancy in the amount of cfDNA release between HT29 and HCT116 cancer cells. This discrepancy could be due to the fact that HCT116 cells harbor a wild-type TP53 whereas HT29 cells have mutated TP53. As TP53 favours radiosensitivity60, it may also be determinant in the differential release of the cfDNA. However, this comparison was not possible in UM (OMM2.5 and MP41), in which TP53 mutations are uncommon61, or A549 lung cancer cells that have been reported as wild-type TP5362. Further studies are therefore required to analyze the effect of specific mutations, e.g., TP53, in cfDNA release.
In addition to cfDNA levels, the length of cfDNA fragments may also have prognostic value37,38.
In both physiological and pathological conditions, a fragment length of ~ 167 bp has been suggested as a result of apoptosis, reflecting the length of DNA wrapped around a nucleosome, 147 bp, plus a stretch of DNA, 20 bp63. In contrast, necrosis has been associated with high-molecular-weight DNA. Therefore large fragment sizes of about > 10,000 bp have been found8. However, the pattern of cfDNA fragmentation in cancer patients remains poorly characterized. While some studies have shown that cfDNA isolated from individuals with cancer is longer compared to those obtained from healthy subjects, others have reported an opposite observation11,38,51,64. More studies, including additional methodologies are warranted and are in progress by our group to clarify this matter. Aside from this progress, our data show that the sizes of releases cfDNA depend on the anticancer treatments.
On the one hand, we show that anticancer treatments determined the levels and sizes of cancer cell-derived cfDNA. On the other hand, encapsulated DNA in extracellular vesicles such as exosomes is another major source of cfDNA, but which was not examined in the present study.
Studies targeting this specific reservoir of cfDNA are needed to complete the scheme as they will provide further information about the mechanism of cfDNA release from cancer cells65. In addition, the next logical step is to apply the analytical strategies we used in this in vitro study to our established in vivo model21 by analyzing mutant cfDNA in the circulation as a preclinical step.
Our observations will help to better understand the release pattern of cfDNA under anticancer therapeutic strategies. This approach will be beneficial in leading to a better understanding of cfDNA role in cancer development and how best to utilize this indicator clinically as a biomarker of treatment response and tumor progression.