The role of tumour heterogeneity and clonal cooperativity in metastasis, immune evasion and clinical outcome

Review Open Access Open Peer Review The role of tumour heterogeneity and clonal cooperativity in metastasis, immune evasion and clinical outcome Deborah R. Caswell1Email author and Charles Swanton1, 2 BMC Medicine201715:133 https://doi.org/10.1186/s12916-017-0900-y ©  The Author(s). 2017 Received: 1 February 2017Accepted: 22 June 2017Published: 18 July 2017 Open Peer Review reports Abstract Background The advent of rapid and inexpensive sequencing technology allows scientists to decipher heterogeneity within primary tumours, between primary and metastatic sites, and between metastases. Charting the evolutionary history of individual tumours has revealed drivers of tumour heterogeneity and highlighted its impact on therapeutic outcomes. Discussion Scientists are using improved sequencing technologies to characterise and address the challenge of tumour heterogeneity, which is a major cause of resistance to therapy and relapse. Heterogeneity may fuel metastasis through the selection of rare, aggressive, somatically altered cells. However, extreme levels of chromosomal instability, which contribute to intratumour heterogeneity, are associated with improved patient outcomes, suggesting a delicate balance between high and low levels of genome instability. Conclusions We review evidence that intratumour heterogeneity influences tumour evolution, including metastasis, drug resistance, and the immune response. We discuss the prevalence of tumour heterogeneity, and how it can be initiated and sustained by external and internal forces. Understanding tumour evolution and metastasis could yield novel therapies that leverage the immune system to control emerging tumour neo-antigens. Keywords Intratumour heterogeneityTumour progressionMetastasisLinear evolutionBranched evolutionCompetitive evolutionCooperative evolutionMutation burdenImmunotherapyAneuploidy tolerance Background In his 1958 essay, Foulds [1] explains that linear tumour progression, the orthodox view of tumour evolution at the time, is the theory that neoplasia advances through an orderly sequence from local invasion, to progressive lymph node invasion, to metastasis [1]. Foulds goes on to discuss tumour progression and metastasis in several cancer types and concludes that neoplasia is not linear, but a complex, ever-diverging route throughout tumour development [1]. This was one of the first articulate explanations of tumourigenesis as a multistep pathway that can progress, persist, or regress [1]. Nowell’s 1976 [2] cancer evolution model proposed that genomic instability drives branched evolutionary pathways from a clone of origin. Heppner [3] reviewed literature on the emergence of heterogeneity in tumours, and the challenges surrounding its study, emphasising that tumour cells exist in a society, and that the interactions between heterogeneous populations are as important as the subclones themselves [3]. Recent technological advances have allowed scientists to study tumour complexity in more detail. Following exome sequencing on multiple spatially separated samples obtained from primary carcinomas and metastatic sites, intratumour heterogeneity and parallel evolution of subclonal driver events was uncovered [4–7]. Multi-region sequencing has now been performed in many cancer types including breast, lung, colorectal, renal, oesophageal cancer and glioma, uncovering intratumour heterogeneity in all cancer types studied [4–15]. Drivers of heterogeneity Intratumoral heterogeneity exists in many forms, from somatic coding and non-coding alterations to epigenetic, transcriptomic and post-translational modifications [16, 17]. Intratumoral copy number heterogeneity also exists (see [16]). Many endogenous triggers of cancer genome instability contribute to intercellular heterogeneity (see [18]), some of which may be therapeutically exploitable. Defective DNA mismatch repair results in hypermutation and microsatellite instability, and mutations inhibiting the proofreading ability of DNA polymerases δ and ε increase base mismatches [19]. Evidence also exists that tumour cell dormancy contributes to tumour heterogeneity (see [20]). Recently, it has been uncovered that APOBEC (apolipoprotein B mRNA editing enzyme catalytic polypeptide) family members are endogenous drivers of tumour diversity in many tumour types [21]. These enzymes initiate DNA cytosine deamination [21, 22], and are a major source of subclonal cancer gene mutations in bladder, breast, head and neck squamous cancers, lung adenocarcinomas and lung squamous cell carcinomas [21, 23–26]. External factors such as cytotoxic therapy [27, 28], and patient factors such as genetic background [29, 30], can also influence tumour heterogeneity. Mechanisms mediating levels of genomic instability and tumour heterogeneity Cahill and Vogelstein [31] discussed the conflict between disadvantages and advantages of genomic instability in tumour evolution. They questioned how cancer cells are able to select for alterations driving genomic instability, which is, usually disadvantageous to the cell, can sometimes lead to cell death, and has no direct growth advantage [31]. Looking to basic studies of mutation rate and cellular fitness in bacteria, the authors reasoned that in stressful environments bacteria with higher overall levels of genomic instability eventually dominate the population because they can adapt [31]. This model can be applied to tumour populations, where genomic instability may be critical for tumour progression [31]. Most normal diploid cells negatively select against chromosomal instability (CIN). This is partially mediated by p53, which inhibits cell propagation after genome instability [32, 33]. CIN mouse models support the concept that low or moderate CfIN levels promote tumour formation, but excessive CIN suppresses tumour formation [34]. This is analogous to mutational meltdown and error-prone catastrophe in bacterial and viral genetics [34–37]. Aneuploid, specifically trisomic cell lines, grow poorly in vitro and as xenografts compared to genetically matched euploid cells [38]. Yet, following prolonged growth, aneuploid cells adapt by acquiring additional alterations correlating with improved fitness [38]. While aneuploidy was detrimental initially, over time it became more advantageous. In tumours, selection might favour the mitigation of excessive CIN to prevent cell autonomous lethality. Partial dysfunction of anaphase-promoting complex/cyclosome (APC/C) lengthens mitosis, allowing more time for correction of impending chromosome segregation errors. This permits tumour cells to fine-tune CIN during tumour evolution, navigating the delicate equilibrium between cell death as a result of too much or too little genomic instability [39]. In colorectal cancer, alterations in BCL9L promote tolerance of chromosome missegregation events, propagation of aneuploidy and genetic heterogeneity [40]. This tolerance (...truncated)