The concept that tumours are essentially heterogeneous and contain a range of subpopulations of cells with different metastatic potential is not new (Fidler et al. Cancer Res, 1978). Intratumour heterogeneity of target molecules is part of cancer evolution and is associated with phenotypic diversity that leads to the emergence of drug resistance. Thus far, the majority of clinical and therapeutic decisions are based on characteristics of individual tumour-tissue samples (primary tumour and/or metastasis), which may neither be representative of the entire tumour burden nor real-time assessments. Thus, the accurate genomic characterisation of tumours, considering also intra- and interpatient complexities, is of absolute importance for biomarker and drug development as well as for tailoring therapy.
The recent study by Gerlinger et al., published in The New England Journal of Medicine, investigated intratumour heterogeneity through the analysis of genomic alterations of multiple tumour regions. Four patients with metastatic renal-cell carcinoma were sampled before and after treatment with everolimus, a mTOR inhibitor, and their primary tumours and matched metastases were mapped.
The authors, while using exome sequencing, chromosome aberration, and ploidy profiling, confirmed the heterogeneity at the intratumour (different regions into the same tumour mass) and also intrapatient (primary tumour vs. matched metastases) levels in all samples assessed. There were genetic aberrations that were ubiquitous throughout tumour specimens (e.g. von Hippel–Lindau gene mutation), whereas other aberrations were showed to be heterogeneously distributed.
For example, about 65% of detected somatic mutations were not present across every tumour location. Some mutations were found to be unique to a certain tumour area, suggesting that mutations change over time (clonal evolution). Furthermore, a prognostic gene-expression signature revealed both good and poor outcomes from different biopsies within the same tumour. Additionally, the authors reported evidence of “convergent phenotypic evolution”, in which similar phenotypes are revealed by different genomic alterations or, in other words, different tumour regions abrogate different mutations of the same pathway. Interestingly, evolutionary biology explains the convergent phenotypic phenomena in which unrelated species respond to similar selection pressures through the development of similar traits (e.g. wings of birds, bats and insects) (Steiner et al. Mol Biol Evol, 2009).
The data regarding intratumour heterogeneity explored by Gerlinger et al. suggest that an individual biopsy, which is the current standard tool for assessing patient’s genomics, is not representative of the entire tumour bulk. This has important implications that may refine our understanding of tumour biology and the mechanisms that lead to therapeutic resistance. Identifying the driver or tumour-initiating mutations (representing the “trunk” of the evolutionary tree) and also divergent mutations (representing the “branches” of the tree) should provide more information to fight cancer. The remaining challenge is to select which target molecules we are going to shut down; this will indeed affect the design of future clinical trials and biomarker discovery programmes.