Convergent TP53 loss and evolvability in cancer

Abstract

Cancer cell populations evolve by a stepwise process involving natural selection of the fittest variants within a tissue ecosystem context and as modified by therapy. Genomic scrutiny of patient samples reveals an extraordinary diversity of mutational profiles both between patients with similar cancers and within the cancer cell population of individual patients. Does this signify highly divergent evolutionary trajectories or are there repetitive and predictable patterns?Major evolutionary innovations or adaptations in different species are frequently repeated, or convergent, reflecting both common selective pressures and constraints on optimal solutions. We argue this is true of evolving cancer cells, especially with respect to the TP53 gene. Functional loss variants in TP53 are the most common genetic change in cancer. We discuss the likely microenvironmental selective pressures involved and the profound impact this has on cell fitness, evolvability and probability of subsequent drug resistance.

Keywords: TP53; cancer; convergence; drug resistance; hypoxia; stem cells.

Fig. 1

Recurrency of mutations in cancer showing TP53 in pole position These data were generated by the cBioPortal for Cancer Genomics (https://www.cbioportal.org/) [23, 24]. We selected the PanCancer Studies database and included only the curated set of non-redundant studies in adults, which included 8 different studies with worldwide cases [–32]. This dataset comprised 25,709 different pan-cancer samples specified by malignant cancer site (anatomical location or topography) and histology (morphology). The piechart shows the number of samples for each study included. Of note, the five most common cancer types were lung adenocarcinoma, colorectal adenocarcinoma, breast carcinoma, pancreatic adenocarcinoma and prostate adenocarcinoma and we displayed in the current figure the top 50 mutated genes. Sequencing data, from the 8 analysed studies, were obtained by using next-generation sequencing (NGS)/targeted sequencing, whole genome paired-end sequencing (WGS), and whole transcriptome/exome paired-end sequencing (WES). For further details about how the data was generated, curated and processed, please see original articles [–32] and the cBioPortal webpage

Fig. 2

Map of TP53 mutations in patients and Mole-rats We illustrate the mutational profile of TP53 using an in-house developed data visualisation tool built with Python to display somatic and germline mutations from cancer patients (https://tp53.isb-cgc.org/) [45] and germline data from rodent species [47, 48]. For the rodent data, two previous publications showed four different germline mutations or variants found in three different rodent species: S104N-Myospalax baileyi (M. b.), S104E-Microtus oeconomus (M. o.), R172K and R207K-Spalax judaei (S. j.) [47, 48]. In human, these TP53 variants correspond to S106N, S106E, R174K and R209K mutations, respectively. GRCh37/hg19 was used as our genome reference and NM_000546/NP_000537 as reference sequences for mutation annotations and protein domains. For the patient data, we included 4,299 samples from cases diagnosed with Li-Fraumeni Syndrome (LFS) and Li-Fraumeni-like Syndrome (LFL) (germline mutations) and 29,656 samples from general cancer patients (somatic mutations). We have filtered the database to only include samples with confirmed germline or somatic mutational status and with available genomic mutation annotation (GRCh37/hg19). The mutations are represented by discs at the codon position. The disc sizes and their distance from TP53 protein scheme are both proportional to the number of mutations. The most frequent alterations are annotated within each disc. The discs located above the protein scheme represent the somatic data for cancer patients, the ones immediately below refer to the germline mutations found in the LFS and LFL cases and the rodent data is displayed at the bottom of the plot. Mutations are coloured according to their effect (missense, frameshift-red, nonsense, silent, etc.) and TP53 protein structure is coloured highlighting its main domains (NP_000537). Even though, they are not frequent enough to be automatically highlighted (size of the discs and distance from TP53 protein scheme), we decided to display the ‘shared’ mutations between rodent and human cancer data for clarity purposes

Fig. 3

Surviving hypoxia: TP53 mutant selection and evolvability Figure illustrates multiple fitness impacts in cells surviving hypoxic intra-tumour environments via TP53 LOF. These phenotypic features compound to increase adaptive evolvability enabling both metastasis and drug resistance. EMT, epithelial-mesenchymal transition. The curved line is to indicate that EMT is a response to hypoxia (see text for explanation), but is only likely to happen if cells can survive hypoxia/acidosis-associated cell death. Which in turn is much more probable if TP53 signalling is aborted. Note that ocogenic or genotoxic stress can also select for TP53 LOF in the absence of hypoxia and with the same fitness benefits. However, the overall impact on evolvability will be less in the absence of hypoxia-driven EMT

Fig. 4

Evolutionarily ancient mechanisms of resistance in bacteria The three generic and evolutionarily conserved mechanims of escape are illustrated with black dots in cells representing mutations underpinning resistance. Phenotypic plasticity includes dormancy or proliferative quiescence as well as rapid adaptability of intracellular signalling networks