Cellular Plasticity in Cancer

Abstract

During cancer progression, tumor cells undergo molecular and phenotypic changes collectively referred to as cellular plasticity. Such changes result from microenvironmental cues, stochastic genetic and epigenetic alterations, and/or treatment-imposed selective pressures, thereby contributing to tumor heterogeneity and therapy resistance. Epithelial-mesenchymal plasticity is the best-known case of tumor cell plasticity, but recent work has uncovered other examples, often with functional consequences. In this review, we explore the nature and role(s) of these diverse cellular plasticity programs in premalignant progression, tumor evolution, and adaptation to therapy and consider ways in which targeting plasticity could lead to novel anticancer treatments. SIGNIFICANCE: Changes in cell identity, or cellular plasticity, are common at different stages of tumor progression, and it has become clear that cellular plasticity can be a potent mediator of tumor progression and chemoresistance. Understanding the mechanisms underlying the various forms of cell plasticity may deliver new strategies for targeting the most lethal aspects of cancer: metastasis and resistance to therapy.

Figure 1.. The contribution of lineage plasticity to tumor initiation.

Solid tumors are classified based on the organ in which they arise and their histological, molecular, and/or transcriptomic profiles. For example, primary tumors in the liver can be histologically classified as hepatocellular carcinoma (HCC) or cholangiocarcinoma (CC). While the cellular origins of divergent tumor types remain unclear, there are two general prevailing hypotheses. (A) One hypothesis proposes that different tumor types arise from different cells of origin. With respect to liver cancer, this would imply that HCC arises from hepatocytes, while CC’s are derived from cholangiocytes. (B) Alternatively, different tumor types may arise in a single organ through lineage plasticity, wherein distinct genetic or epigenetic events may induce a common cell-of-origin to acquire divergent malignant phenotypes. There is evidence for lineage plasticity in cancers of the esophagus (intestinal metaplasia), pancreas (acinar-to-ductal metaplasia), liver (biliary trans-differentiation), and lung and cervix (squamous metaplasia). See text for details.

Figure 2.. Epithelial-mesenchymal plasticity.

Epithelial cells are characterized by intercellular connections comprised of junctional proteins such as E-cadherin. Over the course of EMT, these cells lose these junctions and instead acquire the functional and morphologic phenotypes reminiscent of fibroblasts. These changes are orchestrated by a transcriptional rewiring that results in the silencing or repression of epithelial genes and a concomitant upregulation of mesenchymal genes. While this process classically represents a “complete EMT,” there is increasing evidence of partial EMT states, which are frequently defined by dual expression of epithelial and mesenchymal genes at both the transcriptional and protein levels. It is unclear whether these observed partial EMT states represent stable states or are transient intermediates along an EMT spectrum. While the mechanisms underlying partial EMT are still mostly unknown, there is evidence that relocalization of the junctional proteins plays a role during this process.

Figure 3.. Cellular plasticity and therapy resistance.

Acquired resistance to therapy is associated with a variety of histologic, molecular, and/or transcriptomic changes. (A) In some cases, resistance is mediated by a pre-existing subpopulation of tumor cells that are already intrinsically (and stably) resistant at the onset of therapy. Under these circumstances, therapy induces an initial response, but survival and eventual outgrowth of the resistant subpopulation results in tumor recurrence. Such resistance can occur through genetic or epigenetic means (i.e. mutations or epigenetic silencing of the target of therapy). (B) Alternatively, cancer cells may switch back and forth between drug-sensitive and drug-resistant state as a result of cellular plasticity programs. Under these circumstances, treatment would result in killing of cells in the sensitive state. Resistant cells returning to the sensitive state would also be killed, but any cells capable of stabilizing or “locking-in” the resistant state would have a selective advantage leading to recurrence. These two paradigms are exemplified by NSCLC’s that are driven by mutations in EGFR. In tumors with a pre-existing subpopulation of resistant cells (e.g. cells with a secondary EGFR mutation), initial response to EGFR inhibition (EGFRi) is followed by relapse, driven by the resistant subpopulation. In tumors where cancer cells cycle between a sensitive state and a resistant state, therapy will result in the outgrowth of tumor cells that manage to stably adapt the resistant state. As shown in the figure, this includes NSCLC’s that acquire a SCLC-like identity with neuroendocrine features after treatment. While such resistant tumors typically harbor the same EGFR mutation as the original (pre-treatment) tumor, they no longer depend on it for survival. This paradigm also likely applies to tumors that utilize epithelial-mesenchymal plasticity as a treatment escape mechanism.

Figure 4.. Epithelial-mesenchymal plasticity results in changes to the tumor microenvironment (TME).

As cells undergo EMT, their secretory patterns change, resulting in differences in components of the extracellular matrix (ECM) and ECM-modifying factors such as matrix metalloproteinases (MMPs). Cells that have undergone EMT also secrete higher levels of pro-angiogenic and pro-inflammatory cytokines like GM-CSF, IL6, and TNFα. These factors recruit immunosuppressive leukocyte populations into the tumor, which results in the exclusion of CD8+ T cells. Collectively, these TME-remodeling factors facilitate tumor cell invasion, metastasis, and immune evasion.