Professor Julian Downward, Lung Cancer Group
Can circulating free tumour DNA be detected in preclinical cancer models prior to detection of tumours by imaging?
A number of genetically engineered mouse models of cancer are being employed to address this question, including models of KRAS p53 mutant lung adenocarcinoma, KRAS p53 mutant pancreatic ductal adenocarcinoma and EGFR p53 mutant lung adenocarcinoma.
A particular difference between cancers in genetically engineered mouse models and in humans is that in the mice the tumour is artificially driven, often by a defined recombination event, while tumours in patients are caused by diverse mutation events. This potentially provides an easier route to detection of tumour-derived DNA in the model system by analysis of the recombination product. In the case of the lung cancer models, this occurs in a small number of cells in the lung that are exposed to intratracheally delivered Cre recombinase-expressing adenovirus.
A number of polymerase chain reaction (PCR)-based approaches have been used to study these rearrangements of tumour DNA in the plasma. Both digital PCR and quantitative PCR protocols have been optimised for this purpose.
In the KRAS p53 mutant lung adenocarcinoma model, rearrangement of the p53 locus can readily be quantified in circulating free DNA (cfDNA) from plasma, occurring at about 5–10% of the level of the unrecombined allele when a high tumour burden is present. When the appearance of recombined p53 is followed over time in repeated blood samples, it can first be picked up as early as four weeks after tumour initiation.
When micro CT scanning is carried out, the very earliest that any lesions are radiographically detectable is at six weeks post-infection, when they are about 1mm in diameter. These data suggest that analysis of cfDNA in plasma may enable detection of tumours before they can be visualised by high-resolution imaging technologies, providing hope that circulating tumour DNA may have potential as an early detection methodology.
What processes are involved in the release of circulating free tumour DNA and is it possible to promote its appearance?
While the presence of circulating tumour DNA (ctDNA) has been reported in plasma and serum from patients with many advanced cancers and can be seen in preclinical cancer models, the mechanisms accounting for its appearance in the circulation are poorly understood. A major source of ctDNA is likely to be the death of cells from the tumour mass, either by apoptosis, leading to nucleosome-sized fragments of around 200 base pairs, or necrosis, leading to larger fragments of up to 20,000 base pairs. In addition, live cells may be released from the tumour mass into the circulation as circulating tumour cells, whose subsequent death could release DNA, although this probably only contributes a minor fraction of the ctDNA. Poorly defined active mechanisms for DNA release from live tumour cells may also exist, including the release of exosomes.
In some cancers, circulating free DNA (cfDNA) levels have been shown to correlate with inflammatory markers, but it is unclear whether ctDNA is also elevated in these cases. Release of DNA from tumour cells to the circulation may involve the action of macrophages, either on necrotic or apoptotic tumour cells and cell debris.
The mechanisms of ctDNA release are being investigated in the experimentally tractable systems described above. As part of studies investigating immunological components of the response of tumours to therapeutic intervention, preclinical lung cancer models have been crossed onto backgrounds defective in various components of the adaptive and innate immune systems. In addition, specific inflammatory cells can be targeted either with antibodies or other agents. These systems should enable a better understanding of the molecular and cellular processes involved in the release of ctDNA. This will help to identify factors that might impede ctDNA detection, and might also lead to identification of interventions that would promote release of cfDNA, which might perhaps aid early clinical detection of tumour development.
Is it possible to use circulating free tumour DNA as a screening tool in patients at risk of developing pancreatic or lung cancer?
The digital polymerase chain reaction (PCR) methodology employed successfully in preclinical models is highly sensitive, giving close to single-molecule detection. This methodology has been adapted for use in human clinical plasma samples, focusing particularly on the detection of KRAS and EGFR mutations. It has also been compared with targeted sequencing of panels of genes in human plasma samples.
In the case of pancreatic cancer, where KRAS mutation frequencies approach 100%, digital PCR has shown very good concordance in the identification of KRAS mutations between circulating free DNA and tumour biopsy.
To explore the potential of single-molecule KRAS mutation analysis of circulating tumour DNA for use in the early detection of pancreatic and lung cancer, plasma samples from both early-stage cancer patients and individuals considered to be at high risk of developing these cancers are being analysed. Subsequent work on patient samples will study the presence of other mutations and changes in DNA methylation, in particular with a view to finding independent supporting evidence for the presence of early-stage tumours that could be used to increase the specificity of any plasma DNA-based tests.