Metastasis, the primary cause of cancer deaths (1), occurs through a multistep cascade (2), beginning with the detachment of cancer cells from the primary tumor, which is followed by local invasion into the surrounding tissue (3), intravasation into blood or lymphatic vessels (4), survival of circulating tumor cells (CTCs) (5), extravasation (6), colonization of a distant organ (7), and culminating in the growth of a secondary tumor. Critical cellular traits are acquired at each step of the metastatic cascade as cells transition through distinct states in which they gain new phenotypic and functional features (8, 9). While lineage differentiation was once considered unidirectional and irreversible, cell identity is now known to be far more plastic than previously assumed. Cellular plasticity is the capacity of cells to undergo reprogramming and alter their fate and identity in response to intrinsic or extrinsic factors (10). For example, over the past few decades, researchers have identified reprogramming factors capable of dedifferentiating mature cells into pluripotent stem cells (11-15). Moreover, cellular plasticity allows tumor cells to transition between the different cell types to overcome specific challenges (16).

In terms of cancer cell morphology, epithelial-to-mesenchymal transition (EMT), a reprogramming process by EMT-TFs, which promotes loss of epithelial cell-cell contacts and acquires mesenchymal features, has long been recognized for its critical role in cancer progression and metastasis (17). On the other hand, the adherent-to-suspension transition (AST) is an emerging reprogramming process that facilitate metastatic dissemination by allowing adherent solid tumor cells to escape anchorage dependence and primary tumor sites. Recently, it was discovered that a combination of four defined factors orchestrates the AST plasticity in cancer cells, thereby reprogramming their anchorage dependency and fostering metastatic dissemination (18, 19). This review primarily delves into the recent insights into the biology of EMT and AST plasticity, and their relationship during cancer metastasis.

During EMT, cells from the epithelial state, characterized by cell-cell junctions and polarity, shift to the mesenchymal state, which loses cell-to-cell adhesions and acquires migratory and invasive behavior. Initially identified in the context of embryonic development where it takes part in gastrulation, EMT was subsequently implicated in various human diseases including fibrosis and cancer progression (20). EMT can be induced by pleiotropic signaling factors, including TGFβ, Wnt, Notch, or inflammatory cytokines. These then trigger the expression of specific transcription factors, notably SNAI1/2, TWIST1, and ZEB1/2, which are collectively referred to as EMT-TFs (21). EMT is often confirmed by testing for pronounced changes in the expression of epithelial markers, such as E-cadherin, occludins, and cytokeratin, as well as mesenchymal markers like N-cadherin and vimentin.

The activation of EMT-TFs plays critical roles during the invasive process (Fig. 1A). Jing et al. initially reported the significance of EMT-TFs in metastasis, identifying an essential role for Twist1 in the metastasis of breast cancer cell lines (22). Furthermore, reducing Snail expression inhibited metastasis development in PyMT breast cancer model (23) and deletion of genes encoding Zeb1 impaired metastasis in pancreatic cancer models (24). Hence, the loss of cell-cell adhesion molecules through EMT-TFs enables cancer cells to dissociate from their primary tumor. This subsequently triggers the formation of stress fibers, which then regulate cofilin and actin, thus facilitating the dynamic reorganization of the cytoskeleton and cell cycle (25, 26).

The tumor niche’s microenvironment plays a crucial role in inducing EMT and initiating metastasis. Cancer-associated fibroblasts (CAF) support tumor cells by secreting extracellular matrix and matrix metalloproteinases, promoting migration, invasion, and angiogenesis. Additionally, they induce EMT in cancer cells by activating autocrine and paracrine secretion of TGFβ (27). Following the initial stages of local invasion, cells that undergone EMT have been shown to be located at the invasive front and begin to penetrate the blood or lymphatic vessels. Cells that manage to enter the circulation are referred to as CTCs, only a portion of which manage to survive and extravasate into the secondary organ (Fig. 1B). Overexpression of Twist1 in mouse skin squamous cell carcinoma (SCC) promotes tumor invasion and intravasation of tumor cells into the bloodstream, producing CTCs with EMT traits (28). Furthermore, the EMT process facilitates the release of CTCs by promoting angiogenesis, as shown by the expression of the angiogenic factor and EMT target gene VEGF-A, in CTCs isolated from breast cancer patients (29). EMT also induces protein hydrolase, such as matrix metalloproteinases (MMPs), to enhance the migration of tumor cells (30). Analyses of patient blood suggest that MMPs production by CTCs could serve as a functional marker for cells possessing metastasis-enabling properties (31, 32). Unlike epithelial cells, which undergo anoikis upon the loss of cell-extracellular matrix (ECM) signals during intravasation, EMT-associated survival signals enables CTCs to acquire anoikis resistance and evade cell death (33).

Nevertheless, recent reports have cast doubt on the indispensability of EMT in cancer metastasis (34). The collective invasion of tumor cells, coupled with sustained expression of E-cadherin or EMT-independent phenotypes, has been linked to heightened metastatic capacity and unfavorable prognosis (35-37). Moreover, clustered CTCs, which arise from oligoclonal tumor cell aggregates harbors epithelial properties and exhibit greater aggressiveness in metastasis than single CTCs (38-45). The contribution of EMT in metastasis has been further questioned by several studies. In the first study, Fsp-1 GFP-reporter lines were used to monitor the expression of mesenchymal markers in a MMTV-PyMT breast cancer model, which revealed that GFP-positive cells, indicative of EMT, did not significantly contribute to metastatic outgrowths (46). Another study found that targeted loss of either Snail or Twist in the pancreas did not hinder cancer dissemination or metastasis of pancreatic ductal adenocarcinoma (PDAC) (47). Collectively, these findings prompted the field to question the necessity of EMT for metastasis and suggested the need for an alternative paradigm to explain the metastatic cascade that do not involve EMT.

Until now, the mechanisms governing the reprogramming of anchorage dependency, a key morphological shift from adherent solid tumor cells to suspended CTCs during cancer metastasis, have remained unexplored. The question of what regulates the detachment of cancer cells from the matrix at the primary tumor site has remained unanswered. Recently, a novel theory was introduced showing that adherent-to-suspension transition (AST) in cancer cells can be induced by a combination of defined hematopoietic factors, which were critical for the dissemination of CTCs. This discovery reveals the role of hematopoietic mimicry of solid tumor cells and suggests AST factors as potential therapeutic targets for anti-metastatic drug development.

Analyses of gene expression profiles across a broad spectrum of adherent and suspension-type cells revealed transcription factors exclusively expressed in hematopoietic-like suspension cells, which were termed AST factors. The ectopic expression of AST factors, including ikaros zinc finger family 1 (IKZF1), nuclear factor erythroid 2 (NFE2), interferon regulatory factor 8 (IRF8), and BTG anti-proliferation factor (BTG2), has been shown to directly convert adherent cells into suspension-type cells (Fig. 2A) by modulating cell-matrix adhesion and extracellular matrix (ECM) organization. Although these four AST factors play a role in blood cell development (48-52), they do not induce hematopoietic lineage differentiation when expressed in cancer cells. Following cellular detachment from the matrix, cells typically undergo a form of programmed cell death or anoikis, however, AST-TFs promote anoikis resistance simultaneously. Furthermore, AST factors evoke global suppression of genes related to focal adhesion assembly and integrin-ECM interaction by inhibiting the effectors of Hippo pathway, YAP and TEAD. This is consistent with previous reports implicating YAP/TEAD in facilitating cell-ECM interaction in cancer cells though upregulation of integrin and ECM-related genes (53-56). The physiological loss of anchorage dependency leads to increased intracellular reactive oxygen species (ROS) and anoikis (57). AST factors confer anoikis resistance by reducing ROS levels through the upregulation of HBA1/2 gene expression, which is otherwise found primarily in red blood cells (RBCs). A recent report demonstrated that CTCs disseminated from primary tumor alter their expression of hemoglobin genes to overcome ROS-induced cell death (58). These data suggest AST is a physio-pathologically relevant cellular process that governs the metastatic dissemination and survival of CTCs.

Through single-cell RNA sequencing analysis of paired samples from primary tumors and blood collected from de novo metastatic patients, as well as fatpad-to-lung metastasis mouse model, increased expression of the four AST factors was observed in CTCs compared to their counterparts in the primary tumor. The downregulation of YAP/TEAD and the induction of globin genes in CTCs were also recapitulated, underscoring the key mechanisms involved in the AST process. The critical role of AST factors in the metastatic dissemination is supported by results showing that blocking AST via shRNA impedes both the dissemination of CTCs from the primary tumor and their metastatic capability without altering primary tumor growth. Blocking the AST-induced reprogramming of anchorage dependency with immunomodulatory drugs (iMIDs) like lenalidomide and pomalidomide, which selectively degrade IKZF1(59, 60), significantly suppresses metastasis. These studies demonstrate the potential for developing therapeutic interventions that target AST factors as anti-metastatic agents (Fig. 2B).

Epithelial cells maintain cell-cell adhesions via cellular structures like tight junctions (TJs), adherence junctions (AJs) and desmosomes. EMT evokes the loss of cell-cell adhesion molecules, typically via the downregulation of E-cadherin, which is one of the critical components of the AJs (61). In contrast, AST involves the reprogramming of anchorage dependency through the dissociation of integrin-matrix adhesion rather than cell-cell adhesions, raising the question regarding the relationship between the modulation of cell-ECM interaction via AST and cell-cell interaction via EMT. According to recent reports, AST can be induced in both epithelial and mesenchymal cancer cells without altering the expression of EMT markers, such as E-cadherin, N-cadherin, or vimentin (Fig. 3) (62). Moreover, epithelial cells in an AST-induced suspended state maintained similar expression level and cell surface localization of E-cadherin between cell-cell junctions, suggesting that AST factors could reprogram anchorage dependency via mechanisms distinct from those of EMT. The independency of AST from EMT theory offers an alternative mechanism to explain metastasis that are seemingly unrelated to EMT. Cancer cells undergoing EMT, however, acquire highly migratory and invasive properties through ECM remodeling, characterized by increased ECM synthesis, deposition, and remodeling enzymes like MMPs (63). Thus, exploring how EMT-induced restructuring of the ECM affects the cell-matrix dissociation triggered by AST warrants further investigation.

During development, EMT-driven mesenchymal-like cells reclaim their epithelial phenotype through mesenchymal-to-epithelial transition (MET), a reverse process of EMT (64). In tumor metastasis, the mesenchymal state induced by EMT promotes the local invasion of cancer cells from the primary tumor, leading to the intravasation of individual CTCs with increased survival signals preventing anoikis. Despite the increase in invasiveness resulting from the loss of epithelial traits, the reversion of EMT through MET is required for the outgrowth of cancer cells after they colonize distant organs (Fig. 3). While EMT-MET plasticity has typically been considered a dichotomous switch, the field have reframed the concept of EMT as existing along a continuum of intermediate or hybrid EMT states characterized by diverse levels of both epithelial and mesenchymal phenotypes that express combinations of epithelial and mesenchymal cell markers (65, 66). This broader concept of EMT plasticity suggests cancer cells can acquire specific adaptations to challenges they encounter during both dissemination and colonization (67). A screen of a large panel of cell surface markers in skin and mammary primary tumors showed multiple tumor subpopulations linked to different EMT spectra, ranging from epithelial to complete EMT, as well as hybrid EMT states (68). Cancer cells seldom undergo complete EMT, opting instead for partial EMT. It is these partial EMT cells that contribute to lung metastasis, whereas cells that undergone full EMT much less efficiently colonize the lung (69).

The plasticity of AST has been demonstrated in various cellular context. The reversible nature of shifting AST to SAT (suspension-to-adherent transition) have been shown via the Tet-ON system, which allowed for doxycycline-induced expression of the AST factors. We observed that cells driven into suspension by doxycycline-induced AST factor expression reverted to an adherent morphology upon doxycycline withdrawal, confirming that AST factor expression level underlies AST-SAT plasticity. Furthermore, corroborating the relevance of AST plasticity with additional physiological evidence, a notable increase in AST factor expression was observed in melanoma CTCs, followed by their subsequent repression in the metastatic lesions of the lung. These findings demonstrate that AST factor expression induces hematopoietic mimicry in cancer cells, governing the dissemination of CTCs from the primary tumor site. Subsequently, the spatial and temporal modulation of anchorage dependency afforded by AST plasticity permits the restoration of adhesive properties that facilitate the colonization of secondary organs (Fig. 4). We expect deeper insights into AST plasticity to prove crucial in the development of novel drugs that can elucidate and target the specific stages of metastasis progression that depend on the AST-SAT spectrum.

Described initially as a pivotal occurrence in embryonic morphogenesis, the roles of EMT in diverse physiological and pathological processes, including cancer metastasis, are now well-established. The AST, which was initially introduced into the context of cancer metastasis, is now appreciated not only in the context of morphological changes required for the dissemination of cancer metastasis but also in situations unexplained by EMT. The molecular mechanisms and dynamics that regulate EMT- and AST-related cancer cell heterogeneity should be investigated to understand how rare population of primary tumor cells generates CTCs and succeeds in metastasis. Given that metastatic cancer cells undergo repeated cycles of losing and gaining specific cell fates as they advance through the multi-step process of the metastatic cascade, there are significant challenges associated with targeting the plasticity of cellular reprogramming during cancer progression. Cancer metastasis requires the upregulation of EMT and AST during the dissemination phase, as well as their suppression upon colonization. Although reports have highlighted the importance of such flexibility in the EMT and AST during metastasis, the factors that determine specific EMT and AST status of cancer cells in vivo remain unclear, as do the means of identifying which step to inhibit to effectively block metastasis. Furthermore, due to the presence of wide spectrum of EMT and AST states, determining the extent to which cells should be reversed remains uncertain (70), underscoring the need for careful characterization in the development of anti-EMT or anti-AST agents.

Despite significant progress in identifying the AST phenomenon and understanding the cellular reprogramming of anchorage dependency leading to the dissemination of CTCs, several questions remain unanswered. Considering the identification of extrinsic factors that induce EMT in cancer progression, we wonder what stimulants and signaling factors in the metastatic microenvironment of the primary tumor site can induce the expression of AST factors. Is AST-SAT plasticity a binary process, or does it involve intermediate states akin to the metastable EMT hybrid states? How do individual AST factors orchestrate the conversion of adherent primary tumors into suspended CTCs?

AST occurs broadly via two sequential steps: 1) a spontaneous dissociation of cell-matrix adhesion through a global suppression of integrin and ECM-related gene expression, followed by 2) the acquisition of anoikis resistance via globin gene induction. Building on previous reports that identified the roles of the four AST factors in determining blood cell lineage differentiation, it would be interesting to determine how each AST factor acts either independently or synergistically to drive each process. During B-cell development, IKZF1 acts as a transcriptional repressor, suppressing an extra-lineage transcription network that includes YAP and TEAD to regulate the gene expression related to cell-ECM interactions (71). In addition, deficiency and/or mutation of IKZF1 leading to its dysfunction or mislocalization via increased cell-stroma adhesion and decreased drug sensitivity is associated with a poor prognosis in acute lymphoblastic leukemia (ALL) (72-75). These suggest IKZF1 could be primarily involved in the initial step of AST by disrupting cell-ECM adhesion, whereas NFE2, which serves as key coordinator of globin gene expression (76), could play a potential role in achieving anoikis resistance during the second step of AST.

This review underscores the significance of cellular plasticity in both EMT and AST during cancer metastasis. The dynamic modulation of cancer cell plasticity in both cell-cell and cell-ECM networks is crucial in the cellular adaptations necessary for overcoming the diverse challenges throughout the metastatic cascade. A deeper understanding of the EMT and AST-related plasticity that regulates cellular behavior and morphology will likely guide the development of novel therapeutic strategies for overcoming metastasis and prolong patient survival.

This work was supported by grants from the National Research Foundation of Korea (2020M3F7A1094077, 2020M3F7A1094089, 2021R1A2C1010828, 2020R1A4A1019063, 2018R1C1B6004301) to HWP, and by the Brain Korea 21 FOUR Program to H.D.H.

The authors have no conflicting interests.