Background: Circulating-tumour DNA (ctDNA) offers a non-invasive method to monitor cancer dynamics. Fragmentation features of cell-free DNA (cfDNA) in plasma have been used for signal enrichment and even sample classification. As urine is a peripheral fluid that truly can be collected “non-invasively”, attention is now turning towards features that could similarly improve its translational utility.
The study: We aimed to investigate heterogeneity in the size profiles of urinary cfDNA fragments from patients with urological and non-urological cancers.
Methods: Shallow-depth paired-end whole genome sequencing of urinary cfDNA was done for 79 bladder, 21 renal, 49 melanoma cancer samples, and 27 healthy controls. We (a) explored the urinary cfDNA fragmentation pattern of each sample, (b) assessed cfDNA size profiles in bladder cancer samples collected at 7 different time points during treatment, and (c) compared median fragment size distributions between healthy control, urological and non-urological malignancies.
Results: cfDNA size profiles showed a distribution ranging from 30-300, with a consistent peak at 81bp, and prominent 10bp periodic oscillations. Bladder and renal cancer profiles were similar to those from healthy controls (p>0.05). Also, no significant difference in fragmentation patterns could be observed at different time points for bladder cancers (p>0.05). However, in comparison to healthy controls, melanoma patients had a greater proportion of longer fragments (p<0.0001).
Discussion: Fragmentation profiles of urine cfDNA reveal a sub-nucleosomal structure hinting at enzymatic digestion. Whilst renal and bladder cancer samples showed similar distributions to healthy controls, it is noteworthy that all samples had low or even undetected ctDNA, as assessed by other methods. Melanoma patients had the late-stage disease and urine generally contained longer DNA. Further exploration of the factors contributing to heterogeneity in urinary cfDNA fragmentation, including e.g. sample collection procedures, could help us realise the potential of this biomarker.