Abstract
Background The mammalian circadian clock is a critical regulator of metabolism and cell division. Work in model systems indicates that disrupting the circadian clock systemically by environmental or genetic means promotes cancer, and that multiple oncogenes can in turn disrupt the circadian clock. However, whether the circadian clock is disrupted in primary human cancers is unknown.
Results Here we used transcriptome data from mice to define a signature of the mammalian circadian clock based on the co-expression of 12 genes that form the core clock or are directly controlled by the clock. Our approach can be applied to samples that are not labeled with time of day and were not acquired over the entire circadian (24-h) cycle. We validated the clock signature in circadian transcriptome data from humans, then developed a metric we call the delta clock correlation distance (ΔCCD) to describe the extent to which the signature is perturbed in samples from one condition relative to another. We calculated the ΔCCD comparing human tumor and non-tumor samples from The Cancer Genome Atlas and six independent datasets, discovering widespread dysregulation of clock gene co-expression in tumor samples. Subsequent analysis of data from mouse clock knockouts suggested that clock dysregulation in human cancer is not caused solely by loss of activity of clock genes.
Conclusions Our findings suggest that the circadian clock is dysfunctional in a wide range of human cancers. In addition, our approach opens the door to using publicly available transcriptome data to quantify clock function in a multitude of human phenotypes.
Abbreviations
- ΔCCD
- delta clock correlation distance
- MAD
- median absolute deviation
- TCGA
- The Cancer Genome Atlas
- ZT
- zeitgeber time