Improving survival rates in cancer patients rests on our ability to serially monitor and understand how cancers evolve and acquire resistance to therapy. Circulating tumour DNA, found in the bloodstream of patients has been used to track genomic evolution quite effectively by capturing somatic changes such as mutations and copy number alterations. However, tumours also adapt to therapeutic pressure by altering their transcriptome through ‘non-genomic evolution’. It is extremely challenging to investigate these changes using invasive serial tumour biopsies. Hence, it would be ideal to utilise circulating tumour DNA to dynamically monitor the transcriptome of cancer cells in vivo.
Here, we present a novel framework to track tumour gene expression changes by analysing circulating tumour DNA in serial blood samples called SNIPER (Serial Non-Invasive Plasma gene Expression Reconstruction). The regions surrounding the transcriptional start sites of actively expressed genes are under-represented in circulating tumour DNA as without the protection of nucleosomes, these fragments are rapidly digested by plasma nucleases. SNIPER comprehensively maps the genomic locations of circulating DNA fragments from low-coverage whole genome sequencing data (<10x) and exploits coverage patterns related to nucleosomal organisation at transcription start sites to infer gene expression.
We have applied SNIPER to 180 plasma samples from patients with breast cancer, melanoma, lung cancer and haematological malignancies. First, we demonstrated how our method can characterise distinct tumour-specific transcriptional profiles which are concordant with TCGA RNA-seq profiles for the different cancer types. Then, through the application of SNIPER across serial plasma samples, we identified adaptive changes in key oncogenic signalling pathways associated with therapeutic resistance to CDK4/6 inhibitor therapy in breast cancer and MAPK inhibitor therapy in melanoma. Finally, we detected lineage switching in EGFR-mutant non-small lung cancer that trans-differentiated to a small cell phenotype following exposure to EGFR-targeted therapy as a non-genetic mechanism of therapeutic escape.
An enduring challenge in cancer management is the ability to monitor adaptive responses to therapy in real time. Our results show that SNIPER can provide a tractable approach for clinical monitoring of transcriptional adaptation in patients following cancer treatment, offering the opportunity for real-time interventions to curtail emerging therapeutic resistance.