Epithelial guardians against cancer: integrating cell competition and extrusion in early tumor control

Abstract

Epithelial tissues maintain homeostasis by employing sophisticated mechanisms to promote clearance of precancerous cells, thus preventing tumorigenesis. Two key processes—cell competition and extrusion—collectively termed epithelial defense against cancer (EDAC), enable normal epithelial cells to outcompete or physically expel transformed counterparts. This review integrates recent advances in EDAC, detailing critical proteins, biomechanical cues, and their interplay with immune surveillance. We quantify cellular proportions essential for EDAC efficacy and examine cancer models, focusing on pancreatic, colorectal, and mammary cancers. By linking molecular mechanisms to clinical potential, we position EDAC as a promising frontier for cancer prevention and therapy.

Introduction

Epithelial defense against cancer (EDAC) is an emerging paradigm that describes how epithelial tissues actively prevent tumor initiation by integrating two complementary processes: cell competition and extrusion as early, strategies to eliminate transformed cells from epithelial tissues [1]. These tissues—lining organs such as the intestine, breast, and pancreas—are constantly exposed to oncogenic insults, including somatic mutations and microenvironmental stressors that threaten their structural and functional integrity [2]. To preserve homeostasis, epithelia deploy intrinsic quality-control mechanisms capable of recognizing and eliminating aberrant cells before malignant transformation occurs [3].

EDAC has gained attention for its role as a first-line, non-immunological barrier to tumor initiation, acting through a sequential process of fitness-based recognition (cell competition) and mechanical clearance (extrusion) coordinated by tightly regulated cell–cell interactions and biomechanical signals [3]. By coupling the recognition of oncogenic changes with directed extrusion and programmed cell death, EDAC suppresses the clonal expansion of potentially tumorigenic cells [4]. In this review, we dissect the molecular and biomechanical underpinnings of EDAC, highlight its experimental modeling in vitro and in vivo, and examine how inflammation and tissue context modulate its efficacy. We also explore the translational potential of EDAC in the early interception of epithelial cancers, positioning it as a critical framework for understanding field cancerization, tumor suppression, and therapeutic targeting in precancerous lesions.

EDAC and cellular competition: fundamentals and experimental models

Epithelial tissues maintain structural integrity and functional homeostasis despite persistent genotoxic insults, such as ultraviolet radiation, tobacco exposure, or chronic inflammation—and the resulting accumulation of somatic mutations. Among evolutionarily conserved surveillance programs, cellular competition emerges as a critical non-immune mechanism to safeguard tissue fitness. Here, epithelial cells assess relative fitness through short-range cell-cell interactions, selectively eliminating less fit or potentially oncogenic neighbors [5]. This “winner-loser” paradigm acts as a proactive layer of tumor suppression, enabling normal cells to expel those harboring early oncogenic insults before clonal expansion occurs [3, 6]. Here, we explore how cellular competition functions as a tumor-suppressive barrier and its role within the broader EDAC, a coordinated, immune-independent process of oncogenic cell recognition and removal.

Biological basis of EDAC and its relationship with cell competition

Cellular competition acts as an intrinsic defense, purging pre-tumor clones before they dominate tissue. Somatic mutations in oncogenes, such as TP53 (tumor protein p53), NOTCH1 (Notch receptor 1), and KRAS (Kirsten rat sarcoma viral oncogene homolog), arise from aging, environmental exposures, or inflammation [7]. However, these mutations do not guarantee cancer; mutant cells must outcompete neighboring normal cells or other mutants for survival and space. Without a selective advantage, they are often eliminated [5]. Notably, “false starts”—clonal expansions that arise but fail to disrupt tissue function—occur in histologically normal epithelia, such as Barrett’s metaplasia and colorectal mucosa, where driver mutations affect up to 1% of cells [8, 9]. This suggests early tumorigenic events are common but often halted by competing clones, positioning cellular competition as a dynamic gatekeeper in cancer evolution [10].

Field cancerization and competitive dynamics

The concept of field cancerization, initially proposed by Slaughter et al. in 1953, describes the presence of somatic mutations or genetic aberrations in histologically normal tissue that precede the development of neoplasia [11, 12]. Recent studies have shown that cell clones with mutations such as TP53, NOTCH1, PIK3CA, APC, or KRAS may be present subclinically in the epithelia of patients without cancer, suggesting that the onset of carcinogenesis is an early and silent process. For example, more than 50% of esophageal squamous epithelium in older adults may contain mutations characteristic of cancer without evidence of dysplasia [13]. This phenomenon has also been described in tissues such as the skin [14], oral cavity [15], stomach [16], intestine [12], and colon [17]. Despite this mutational accumulation, not all altered epithelial fields progress to cancer, indicating the existence of surveillance mechanisms that restrict malignant transformation. Among these, the recently described mechanism EDAC (epithelial defense against cancer) acts to preserve homeostasis by actively eliminating potentially oncogenic clones.

The conceptual integration between field cancerization and EDAC suggests that, as fields accumulate mutations, processes such as cellular competition and apical extrusion can contain the expansion of transformed clones. This could explain why many cancerous fields remain stable or regress to a non-cancerous state. In murine models of esophageal carcinogenesis, mutant clones of TP53 and NOTCH1 have been observed to coexist with normal epithelium and successfully eliminate emerging microtumors through EDAC, without immune involvement or activation of classical apoptosis [18].

These findings suggest that the clonal architecture of adjacent healthy tissue may influence the fate of preneoplastic lesions and help explain the low incidence of cancer, despite the presence of high mutational loads in normal epithelia [9, 19, 20]. This highlights that EDAC may not only suppress initial oncogenic events, but also regulate interclonal competition within genetically altered fields. Its breakdown, therefore, may mark the transition from benign mosaicism to pre-malignant dominance.

Mechanisms of epithelial protection through cellular competition

To combat cancer progression, epithelia deploy the EDAC, a non-immune surveillance system. EDAC integrates mechanical extrusion of pre-neoplastic cells with their death via anoikis, a programmed cell death triggered by loss of matrix attachment. Extrusion occurs apically, pushing cells toward the lumen for clearance, or basally, which may enable deeper tissue invasion [21]. Healthy neighboring cells initiate this removal through mechanisms independent of immune recognition [3]. EDAC thus acts as a protective mechanism, eliminating newly transformed or potentially harmful cells.

Molecular mechanisms of EDAC

EDAC operates across epithelial types—intestinal, mammary, esophageal—through conserved molecular pathways. In models of pancreatic ductal adenocarcinoma (PDAC), immune-like recognition pathways, such as interactions between MHC-I (major histocompatibility complex class I) molecules and the inhibitory receptor LILRB3 (leukocyte immunoglobulin-like receptor B3) trigger caspase-dependent elimination of transformed cells; disrupting LILRB3 in MDCK (Madin-Darby Canine Kidney) cells reduces clearance by ~ 70% [9]. Similarly, the receptor tyrosine kinase EphA2 (Ephrin type-A receptor 2) drives repulsion of RasV12-expressing cells in PDAC-derived HPAF-II cultures, with silencing cutting elimination by 50% [12]. In MCF-10A (human mammary epithelial) cells, protein kinase C iota (PKCι) regulates junctional tension; its inhibition shifts extrusion basally, potentially increasing invasiveness [21]. Sequential mutations also shape fitness: in MDCK cells, PI3K (phosphoinositide 3-kinase) activation in RasV12 backgrounds impairs elimination by 30–50%[22]. These protein-driven pathways highlight EDAC’s role in blocking early neoplastic progression through coordinated recognition, extrusion, and cell death.

Molecular basis of epithelial defense against cancer: from recognition to expulsion

EDAC eliminates transformed cells through a coordinated molecular cascade, triggered by subtle fitness differences between normal and aberrant cells. Healthy epithelial cells detect oncogenic changes, initiating a sequence of recognition, signaling, cytoskeletal remodeling, and cell death that safeguards tissue integrity [22, 23]. This section explores the molecular architecture of EDAC, highlighting how epithelia sense and expel oncogenic threats, with key stages—recognition, extrusion, and elimination—outlined in Fig. 1.

Fig. 1.

Cell fate outcomes following epithelial defense against cancer (EDAC) activation in cancerized epithelial fields. Illustration of potential trajectories for mutant epithelial cells within cancerized fields. A single oncogenic mutation may result in a false start, where the altered cell fails to expand due to lack of fitness advantage. In some cases, clonal expansion creates a field of cancerization composed of morphologically normal but genetically altered cells. This field may activate successful EDAC, resulting in apical extrusion and clearance of transformed cells. Alternatively, EDAC may fail due to impaired recognition, altered mechanical cues, or inflammation, permitting clonal persistence and progression toward malignant transformation. In advanced stages, inverse EDAC may occur, wherein extrusion is basal rather than apical, potentially facilitating invasion and metastasis

Mutations triggering recognition

Epithelial carcinogenesis begins with mutations that confer selective advantages to cells [24]. In vitro models, such as MDCK cells and human mammary epithelial lines, identify oncogenes—Ras, Src, Cdc42 (cell division control protein 42 homolog), YAP1 (Yes-associated protein 1), and ERBB2—as key triggers of extrusion when mutant cells are surrounded by normal epithelium [25–30]. Notably, RasV12, prevalent in colorectal and pancreatic cancers, serves as a model for EDAC pathways [31, 32]. In monolayer cultures, normal cells recognize and apically expel RasV12-transformed cells [3, 28]. These cells undergo metabolic shifts, such as increased glucose uptake and reduced mitochondrial respiration, driven by PDK4 (pyruvate dehydrogenase kinase 4) overexpression under hypoxia, sensitizing them to elimination [1].

Beyond RasV12-driven paradigms, EDAC mechanisms have also been demonstrated in response to other oncogenic alterations, including mutations in TP53, APC, and overexpression of ERBB2. For instance, loss-of-function mutations in TP53 have been shown to induce necroptotic cell death in transformed cells when surrounded by normal neighbors, highlighting that EDAC is not restricted to Ras pathways [4]. In breast epithelial models, ERBB2 overexpression triggers luminal extrusion, resembling EDAC processes that prevent oncogenic expansion in mammary ducts [29]. Similarly, APC truncation or loss, frequently observed in colorectal cancer, alters extrusion polarity, shifting it from apical to basal and promoting invasive phenotypes [33]. These examples underscore that EDAC is a broad, conserved response across diverse oncogenic contexts, extending beyond the RasV12-MDCK model to clinically relevant pathways in pancreas, breast, and colorectal epithelia.

Differential recognition between benign stress and oncogenic transformation

EDAC depends on the ability of normal cells to selectively recognize and respond to neighboring transformed cells, without triggering responses against cells with transient physiological alterations or benign stress. This fine-grained discrimination is achieved through the integration of specific molecular, metabolic, and mechanical signals, which allow the detection of irreversible transformations associated with tumor risk.

At the metabolic level, for example, cells transformed by RasV12 exhibit distinctive changes, such as increased glucose uptake and alterations in mitochondrial respiration, mediated by PDK4 overexpression [1]. These modifications are not present in cells subjected to reversible stress, allowing them to be differentiated at very early stages.

Regarding immunophenotypic recognition, normal cells can detect the overexpression of MHC-I molecules in transformed cells. This interaction with the LILRB3 receptor in normal cells activates the SHP2–ROCK2 pathway, generating mechanical tension directed toward the apical extrusion of the abnormal cell [34].

Another key mechanism involves the secretion of sphingosine-1-phosphate (S1P) by transformed cells. This molecule binds to the S1PR2 receptor in neighboring cells, triggering contraction of the actomyosin ring and promoting extrusion of the transformed cell without compromising the integrity of the epithelium [35, 36].

Likewise, EphA2 (Ephrin type-A receptor 2), frequently overexpressed in oncogenic clones such as those induced by RasV12, activates intracellular signaling pathways dependent on Rho family GTPases (RhoA, Rac1, Cdc42), which reorganize the cytoskeleton to facilitate the expulsion of the affected cell [31, 37].

Finally, physical-mechanical sensing mechanisms also participate in this process. Proteins such as FBP17 (Formin Binding Protein 17) and Cdc42 [38–40] act as mechanosensors capable of identifying changes in membrane stiffness and cellular tension, characteristics characteristic of malignant transformation but absent in merely stressed cells .

Together, these systems allow the epithelium to differentiate cells that pose a real threat to tissue homeostasis from those that are merely experiencing temporary stress. This discriminatory capacity is essential for EDAC to act selectively and efficiently, promote clearance of transformed cells at very early stages without interfering with normal tissue function.

Molecular mediators of extrusion

A central player, EPLIN (Epithelial Protein Lost In Neoplasm), strengthens the actin cytoskeleton, forming actomyosin contractile rings around RasV12 cells; inhibiting EPLIN reduces apical extrusion by up to 60% [24]. EPLIN also coordinates neighboring cells, assembling contraction machinery via Cav-1 (caveolin-1), PKA (protein kinase A), myosin II, and filamin A [24]. The plakin-plectin-EPLIN complex modulates microtubule dynamics through α-tubulin acetylation, with disruption impairing extrusion [6]. Upstream, Rab5 (Ras-related protein Rab-5), a GTPase, boosts extrusion by internalizing E-cadherin at the normal-mutant cell interface, doubling rates in MDCK models [41]. PDK activity aids detachment from the extracellular matrix, accelerating expulsion [42]. Thus, Rab5 links recognition to extrusion, driving E-cadherin turnover, EPLIN stabilization, and actomyosin contraction [42]. These steps are detailed in Fig. 2.

Fig. 2.

Stepwise molecular coordination of epithelial defense against cancer (EDAC) in RasV12-transformed epithelial cells. Sequential stages of EDAC in epithelial cells carrying the oncogenic RasV12 mutation. (a) Recognition. Transformed cells are identified by surrounding healthy cells through juxtacrine and mechanotransductive signaling. These include interactions between MHC-I (major histocompatibility complex class I) and LILRB3 (leukocyte immunoglobulin-like receptor B3), binding of S1P (sphingosine-1-phosphate) to S1PR2 (S1P receptor 2), EphA2 (Ephrin type-A receptor 2) activation, and membranedeformation sensing via FBP17 (formin-binding protein 17). (b) Signaling and extrusion. Cell recognition activates extrusion via Rab5 (Ras-related protein Rab-5)-mediated endocytosis of E-cadherin and downstream recruitment of EPLIN (epithelial protein lost in neoplasm), which stabilizes the actomyosin ring. Additional components include cav-1 (caveolin-1), myosin II, filamin A, and vimentin, supported by the paracrine factor ADAMDEC1 (ADAM-like decysin-1) secreted by healthy neighbors. Rho-family GTPases—Rac1, Cdc42 (cell division control protein 42 homolog)—modulate cytoskeletal tension through the ROCK (Rho-associated protein kinase) pathway. (c) Anoikis. Once extruded, transformed cells typically undergo anoikis. However, some evade cell death through accumulation of Annexin A2 (ANXA2) and S100A10, which suppress reactive oxygen species–induced apoptosis and promote extracellular survival. (d) EDAC failure and basal extrusion. Disruption of extrusion machinery—via knockdown of EPLIN, ADAMDEC1, S1PR2, or mutations in APC (adenomatous polyposis coli)—impairs clearance and promotes basal extrusion, potentially enabling invasion and metastasis

Extrusion signals

Normal cells detect transformed neighbors via direct contact and mechanotransduction [43]. Transformed cells express S1P (sphingosine 1-phosphate), which binds S1PR2 (sphingosine-1-phosphate receptor 2) on adjacent cells, triggering actomyosin contraction at the basolateral domain for apical extrusion [35, 44]. In RasV12 cells, LILRB3 interacts with MHC-I, activating the SHP2 (Src homology region 2-containing protein tyrosine phosphatase 2)–ROCK2 (Rho-associated protein kinase 2) axis to generate extrusion forces via filamin [34]. EphA2, a tyrosine kinase, stimulates Rho GTPases—RhoA, Rac1, and Cdc42—to remodel cytoskeletons in surrounding cells [45]. Mechanical sensors, such as Cdc42 and FBP17, detect local tension, aiding recognition [39].

Cell adhesion and detachment

Apical extrusion requires loss of matrix anchorage. Reduced adhesion eases extrusion, often preceding anoikis, a programmed cell death triggered by detachment [21]. FAK (Focal Adhesion Kinase) sustains matrix-dependent survival; its downregulation promotes death post-detachment [21]. In cells lacking S1PR2, aberrant FAK activity impairs extrusion, a defect reversible by pharmacological inhibition [23, 46]. E-cadherin, a core component of adherens junctions, maintains epithelial integrity; its loss or extracellular truncation promotes extrusion in Cdc42-mutated cells [27]. In RasV12 models, E-cadherin downregulation shifts extrusion basally, aiding invasion [38]. Src mutations enhance myosin II and FAK via MAPK (mitogen-activated protein kinase) signaling, influencing outcomes [47]. Competitive success varies by context: Notch1 mutations in esophageal models drive apical micro-tumor elimination [18], while YAP1 acts as a winner or loser, with expulsion rates of 55–70% in MDCK co-cultures with Src or KRAS [28].

Cytoeskeletal remodeling

Extrusion relies on cytoskeletal changes in both transformed and neighboring cells. RasV12 cells gain size and stiffness via F-actin and myosin II accumulation [3, 26]. Adjacent cells upregulate filamin and vimentin, generating tensile force for extrusion; losing either protein, with vimentin dependent on filamin, reduces efficiency [26]. The Rho–ROCK and PKC (protein kinase C) pathways regulate filamin dynamics, while myosin II controls contractility [3]. Normal cells secrete ADAMDEC1 (ADAM-like decysin-1), inducing filamin and EPLIN in neighbors to support reorganization and extrusion [48].

Key signaling pathways

Oncogenes, including small GTPases (Ras, Cdc42) and tyrosine kinases (Src, ERBB2), trigger cascades culminating in extrusion [22]. Cdc42, elevated in H-RasV12 cells, is pivotal; its loss impairs extrusion, while activation triggers it [27, 38]. The ERK/MAPK pathway contributes but requires support [38]. In HaCaT (a spontaneously immortalized human keratinocyte line) and MDCK models, LILRB3 or AltR (alternative receptor, a signaling molecule enhancing extrusion) on neighbors engages SHP2–ROCK2 to drive extrusion [34]. EPLIN directs apical extrusion in MDCK cells, but its suppression shifts it basally [12]. In KRAS-transformed CFPAC-1 cells, basal extrusion aids invasion [24]. ANXA2 (Annexin A2) and S100A10 (S100 calcium binding protein A10) block apoptosis in some RasV12 cells, raising metastatic risk [25]. Rab5 enhances extrusion, while COX-2 (cyclooxygenase-2)/PGE2 (prostaglandin E2) signaling suppresses it by 60% in HT-29 models [4, 23]. S1PR2 and ADAMDEC1 increase efficiency by 40–50% in PDAC (pancreatic ductal adenocarcinoma) contexts [27, 28].

Anoikis and EDAC

Once extruded, ransformed cells typically undergo anoikis [49].Some evade this via autophagy, degrading components for energy, as seen in K-RasV12 cells catabolizing S1P for basal extrusion [36, 50] ANXA2 and S100A10 shield RasV12 cells from ROS (reactive oxygen species)-induced apoptosis, enabling multilayered structures in early carcinogenesis [50]. E-cadherin loss promotes EMT (epithelial-mesenchymal transition)-like behavior and anoikis resistance in mammary models [47, 51, 52] amplified by ERBB2 [53]. In PDAC, S1PR2-deficient cells extrude basally, aiding metastasis [46]. Early PDAC and CRC (colorectal cancer) models show normal cells clearing 70–90% of RasV12 cells at a 1:10 ratio via LILRB3 and EPLIN [24, 34]. In advanced CRC, a 1:2 ratio cuts clearance by 50%, suggesting transformed cells may overpower normal ones [43, 54].

Perspectives

EDAC balances protection and risk: apical extrusion clears pre-malignant cells via anoikis, yet basal extrusion may enable invasion. Its outcome hinges on extrusion direction and anoikis sensitivity, varying by context and stage. Future studies should integrate mechanical, metabolic, and cytoskeletal controls with transcriptional and epigenetic shifts, and explore EDAC’s interplay with immune surveillance to uncover therapeutic potential [49].

Apical vs. basal extrusion: determinants and risks in epithelial defense against cancer

EDAC typically protects tissues by apically extruding transformed cells, yet under specific molecular and mechanical conditions, extrusion shifts basally, bypassing containment and potentially aiding invasion or dissemination [36, 43]. This section examines the factors directing extrusion—cell polarity, cytoskeletal tension, oncogenic load, and tissue context—highlighting EDAC’s dual role in either suppressing or promoting early tumor progression.

EDAC’s dual role across cancer stages

In early tumorigenesis, EDAC eliminates transformed cells apically into the lumen, where they undergo anoikis, a programmed cell death triggered by matrix detachment [6]. In advanced lesions, however, extrusion may shift basally, allowing mutant cells to infiltrate stroma and spread [34, 43]. In MDCK models, RasV12-transformed cells extrude apically, but suppression of Cdc42 or ROCK redirects them basally, boosting invasive potential [47]. Co-expression of RasV12 with loss of Scribble, a polarity regulator, drives basal delamination in up to 90% of mutant cells, underscoring Scribble’s role in directing extrusion [55]. Basal extrusion is enhanced by Src mutations, alongside CDCP1 (CUB domain-containing protein 1), lipid raft enrichment, STAT3 (signal transducer and activator of transcription 3) activation, and upregulation of MMPs (matrix metalloproteinases) [33]. Actomyosin contraction dynamics dictate direction: apical contraction causes basal ejection, while perturbations in APC (adenomatous polyposis coli) or its absence favor apical extrusion [33]. In human MCF7 breast cancer cells, the EMT transcription factor Snail activates RhoA, shifting extrusion basally and increasing motility [37]. In 3D cultures, K-Ras-transformed MDCK cells form cyst-like structures, with basal extrusion seeding secondary cysts [56].

Role of APC and S1P2 in extrusion direction

APC, a tumor suppressor and cytoskeletal scaffold, guides extrusion polarity by aligning microtubules and remodeling actin in both extruding and adjacent cells [6]. Its truncation or loss, common in colorectal cancer, promotes basal extrusion and invasive behavior [41]. Similarly, reduced expression of S1P2 (sphingosine-1-phosphate receptor 2), critical for apical expulsion, yields apoptosis-resistant cell clusters that extrude basally, marking failed extrusion events [44]. In ductal adenocarcinomas, low S1P2 levels correlate with chemoresistance and invasiveness, suggesting defective EDAC contributes to malignancy [36, 44]. Thus, EDAC can shift from a protective barrier to a pathway for metastasis.

Collective extrusion and EMT

Tumor cells escape epithelia individually or collectively. Disruption of E-cadherin, a key adhesion molecule, boosts EMT and metastatic potential [57]. In Src-transformed MDCK models, cell clusters undergo collective extrusion, retaining junctions while reducing E-cadherin and increasing vimentin, aligning with partial EMT [51]. This competitive-driven invasion allows extruded cells to evade anoikis and gain migratory traits [58, 59].

Cellular proportions and EDAC efficiency

EDAC efficacy depends on the ratio of transformed to normal (T: N) cells (Table 1). At low T: N ratios (1:100), RasV12-expressing cells extrude apically with 80–90% efficiency [26, 27, 38]. As ratios rise (1:50 to 1:30), efficiency drops to 40–60%, varying by oncogene and context [24, 28, 34, 35, 39, 55, 60, 61]. At moderate densities, transformed cells form contractile aggregates above the epithelial plane, reflecting mechanical repulsion by healthy neighbors [62]. Surrounded by other RasV12 or temperature-sensitive Src mutants, extrusion falters, highlighting EDAC’s reliance on a non-transformed environment [47, 63]. RasV12 cells achieve ~ 84% extrusion at 1:100, falling to 40–50% at 1:50 [26, 28, 35, 38], while v-Src reaches 80% [55]. Human Caco-2 cells, a human colorectal adenocarcinoma line widely used to model intestinal epithelial differentiation and barrier function, show similar ~ 80% rates, suggesting conservation across species and tissues [64].

Table 1.

In vitro models of epithelial defense against cancer: oncogene-specific extrusion efficiency and molecular mediators

Model	Proportion (T: N)	Oncogene	Extrusion efficiency	Time	Main mediators	References
MDCK	1:50	K-Ras(G12V) or v-Src	40–60% / apical	14–24 h	Vimentin, Filamin	Chiba et al. [28]
	1:50	YAP	10–50% / apical	14–24 h
	1:50	RasV12	40–45% / apical	24 h	EPLIN, Cav-1, Myosin II, PKA	Ohoka et al. [24]
	1:50	RasV12	50% / apical	24 h	S1P	Yamamoto et al. [35]
	1:50	Scr	—	—	—	Anton et al. [60]
	1:50	RasV12	50% / apical	16–24 h	FBP17	Kamasaki et al. [39]
HaCaT	1:30	RasV12	40% / apical	16 h	SHP2-ROCK	Ayukawa et al. [34]
MDCK	1:50	RasV12	45% / apical	24 h
	1:50	RasV12	50% / apical	24 h	EPLIN	Kasai et al. [61]
	1:50	RasV12	50% / apical	24 h	Scribble	Kohashi et al. [55]
	1:50	RasV12 + loss of Scribble	90% / basal	–	Scribble
	1:100	Src	Majority (no% )	post-activation	Cdc42, ROCK	Kajita et al. [26]
	1:50	Cdc42	90%	8 h	MAPK	Grieve et al. [27]
	1:50	RasV12	~ 45% / apical	24–52 h	JNK
	1:100	Scr	80%	12–24 h	Myosin II, MAPK	Kajita et al. [25]
	1:100	Cdc42	90%	8 h	MAPK	Grieve et al. [27]
	1:100	RasV12	80%	24 h	Myosin II, FAK	Fujita et al. [45]
	1:100	Src	Majority	post-activation	Cdc42, ROCK
	1:100	Knockdown Mahjong	~ 45% / apical	24–52 h	JNK
	1:10	RasV12	~ 45% / apical	21 h	EphA2	Porazinski et al. [45]
Caco-2	1:50	RasV12	~ 80%	24 h	N-WASP	Wu et al. [64]

Summary of in vitro epithelial models evaluating epithelial defense against cancer (EDAC) in response to diverse oncogenic stimuli. The table includes the transformed-to-normal (T: N) cell ratio used in mosaic cultures, the oncogene introduced (e.g., RasV12, YAP, Src, Cdc42), the efficiency and direction of cell extrusion (apical or basal), experimental time points, molecular effectors implicated in extrusion regulation, and literature references. Models primarily employ MDCK cells (Madin–Darby Canine Kidney), a well-characterized epithelial monolayer system with apical-basal polarity; HaCaT cells, a spontaneously immortalized human keratinocyte line; and Caco-2 cells, a human colorectal adenocarcinoma line widely used to model intestinal epithelial differentiation and barrier function

Implications for altered tissues

EDAC excels in unperturbed epithelia but weakens in dysplastic, inflamed, or genetically altered tissues. Models assume clear mutant versus wild-type distinctions, yet real tissues often harbor subclinical mutations in phenotypically normal clones [65]. This overlap blurs EDAC’s effectiveness, especially under chronic stress, where truly healthy cells are scarce.

Perspectives

The transition from apical to basal extrusion marks EDAC’s shift from tumor suppression to potential invasion. Key determinants—APC, S1P2, and T: N ratios—shape this balance. Collective extrusion and EMT-like behaviors highlight how EDAC can be subverted. Advanced in vivo models are needed to study EDAC failures in precancerous or inflamed tissues, decode cues drivingextrusion polarity, and address anoikis resistance, potentially revealing strategies to reinforce EDAC in early cancer stages.

Factors modulating epithelial defense against cancer: inflammation and microenvironment

The efficacy of EDAC varies dynamically, shaped by the tissue microenvironment. Biomechanical properties, inflammatory mediators, and intercellular signals can either strengthen or weaken this surveillance mechanism. Transformed epithelial cells often exhibit increased stiffness and contractility, aiding extrusion via actomyosin tension [29, 30]. However, chronic inflammation—driven by diet or cytokine dysregulation—may reduce EDAC effectiveness, allowing mutant clones to persist [66–70]. This section explores how external signals influence EDAC, revealing a context-dependent interplay of epithelial mechanics, inflammation, and cell competition.

Biomechanical forces in EDAC

Biomechanical cues critically regulate EDAC. In MDCK and MCF-10A cultures, transformed cells show stiffness up to 1.5 times higher than normal counterparts, triggering actomyosin contractility to promote extrusion [29, 30]. Intracellular calcium waves from neighboring cells boost extrusion efficiency by ~ 30% [30]. Intracellular calcium waves from neighboring cells boost extrusion efficiency by ~ 30% [31]. In mammary epithelial models, mislocalization of vinculin, a key mechanotransducer, redirects ~ 40% of apical extrusions basally, raising invasion risk [21]. These findings highlight biomechanical forces’ dual role—supporting defense or enabling escape—based on direction and magnitude.

Immune-like signaling and inflammatory interference

Though independent of classical immune activation, EDAC engages immune-like recognition. In pancreatic ductal adenocarcinoma (PDAC), interactions between LILRB3 on transformed cells and MHC-I on neighbors drive efficient extrusion, mimicking innate immune clearance [9]. Chronic inflammation, however, undermines this process. In murine PDAC models, upregulation of COX-2 and subsequent PGE2 production in RasV12-mutated and normal cells curbs apical extrusion by up to 60% [66]. Chronic inflammation, prevalent in colorectal and pancreatic cancer-prone tissues, heightens oncogenic risk by altering EDAC [67–69]. Diet-induced inflammation, such as from ω6 fatty acids metabolized to arachidonic acid, shifts cytokine profiles and suppresses extrusion, increasing RasV12 cell retention by 40% in the small intestine and 20% in the pancreas in mouse models [70].

Context-dependent effects of inflammation on EDAC

Inflammation’s impact on EDAC varies by oncogene and signaling context. In MDCK models, YAP (Yes-associated protein)-transformed cells secrete PGE2 via COX-2, enhancing EDAC through the EP2–cAMP–PKA axis, which promotes E-cadherin internalization and extrusion [71]. Conversely, K-RasV12 and v-Src-transformed cells under similar signals exhibit reduced apical extrusion [71]. These contrasting effects underscore how oncogene-specific pathways determine whether inflammation bolsters or impairs EDAC.

Environmental and microenvironmental modulation

Beyond intrinsic mutations, EDAC responds to external inputs—inflammation, mechanical stress, and cell density—reshaping epithelial competition. These factors can amplify fitness differences to reinforce EDAC or disrupt it by altering polarity, junctional tension, or signaling. In PDAC, colorectal cancer, and mammary carcinoma models, the biomechanical-immune axis integrates structural and signaling cues from the tissue niche, critically influencing EDAC fidelity.

Perspectives

Biomechanical forces, immune-like surveillance, and inflammation shape the adaptability of EDAC. Increased stiffness and calcium signaling enhance apical extrusion, yet chronic inflammation—driven by diet or tumor-related cues—can weaken this defense, promoting transformed cell survival and invasion. Future research should unravel molecular pathways linking inflammation to EDAC disruption, evaluate reversibility, and explore interventions to enhance EDAC for prevention or therapy in early epithelial cancers.

Evidence of epithelial defense against cancer in human tissues and models

While most insights into EDAC stem from in vitro and murine systems, emerging evidence suggests similar processes operate in human epithelia. Apical extrusion of transformed cells is documented in monostratified tissues like pancreatic ducts and colonic crypts, positioning EDAC as an early tumor suppressor [24, 34, 35, 48, 62, 72]. Its role in stratified epithelia, such as oral mucosa or skin, remains less clear. This section synthesizes findings from organoid, animal, and human studies, mapping EDAC’s spatial and temporal activity, limitations across epithelial types, and relevance at different cancer stages.

Cancer models and mechanistic insights

In vitro models, particularly MDCK cells, elucidate EDAC mechanisms, highlighting RasV12 mutations and regulators like EphA2, EPLIN, Rab5, and ADAMDEC1 in driving apical extrusion. These processes involve actomyosin contractility, cytoskeletal remodeling, and paracrine signaling, modulated by S1PR2 and the Annexin A2/S100A10 complex for cell recognition and anoikis (programmed cell death due to matrix detachment) [24, 35, 41, 48, 62]. However, these models lack the tumor microenvironment’s complexity, including immune surveillance, stromal interactions, and chronic inflammation.

Murine models offer a broader perspective, capturing systemic and inflammatory contexts. In colorectal, pancreatic, and breast cancer models, EDAC is influenced by chronic inflammation, with COX-2, LILRB3, and MHC-I signaling playing key roles [34, 66]. In mammary epithelia, basal extrusion—driven by PKCι and mechanical effectors like vinculin—signals a shift toward EMT and early invasion [73]. These models bridge molecular mechanisms to physiological outcomes, as summarized in Table 2.

Table 2.

Experimental evidence supporting epithelial defense against cancer mechanisms in vivo and in vitro models

Experimental model	Cancer type	Cells used	Mutation/Condition	Mechanism studied	Clinical implication	Reference
In vivo (mouse)	Colorectal (MC38)	HaCaT RasV12 (human), MC38 (mouse)	RasV12 activation; Lilrb3l deletion	MHC-I / LILRB3 recognition suppresses EDAC	Therapeutic target: LILRB3 inhibition may restore EDAC	Ayukawa et al. [34]
	Pancreatic	Pancreatic ductal cells	Oncogenic RasV12; chronicinflammation	COX-2 protects transformed cells from elimination	Chronic inflammation impairs EDAC, promoting early tumorigenesis	Sato et al. [66]
	Breast	MCF-10 A	aPKCi overexpression	PKCi promotes basal extrusion and EMT features	Basal extrusion may mark EMT onset and invasion potential	Villeneuve et al. [73]
In vivo / In vitro	Pancreatic	Normal pancreatic cells, MDCK	ANXA2/S100A10 accumulation	Apoptosis suppression via ANXA2/S100A10 accumulation	Biomarker of apoptosis-evading tumor cells in the microenvironment	Ito et al. [72]
In vitro	Not specified	MDCK	RasV12 and Src	Rab5 enhances cell extrusion	Rab5 as a regulator of extrusion	Saito et al. [41]
	Early transformation model	MDCK	RasV12	EPLIN promotes apical extrusion	Loss of EPLIN or Cav-1 as early biomarkers	Ohoka et al. [24]
	Early transformation model	MDCK	Src	Collective extrusion mediated by actomyosin	Collective extrusion may support metastasis	Moitrier et al. [51]
	Not specified	MDCK	RasV12 with EphA2 depletion	EphA2 promotes recognition and contraction via Ephrin-A	EphA2 as a therapeutic enhancer of EDAC	Hill & Hogan [62]
	Early transformation model	MDCK	RasV12 with ADAMDEC1 knockdown	ADAMDEC1 secreted by normal cells induces extrusion via filamin and EPLIN	ADAMDEC1 as a biomarker and therapeutic candidate	Yako et al. [48]

Summary of in vivo and in vitro experimental models that provide mechanistic evidence for epithelial defense against cancer (EDAC) in gastrointestinal, pancreatic, and mammary epithelial tissues. The table includes the cancer type modeled, cell lines used, genetic or oncogenic perturbations, mechanistic pathways investigated, and translational implications of each finding. Mechanistic pathways characterized in these models span recognition signals (such as MHC-I–LILRB3, S1P–S1PR2, and EphA2), cytoskeletal and mechanical regulators (EPLIN, filamin A, myosin II, vimentin), immune-like and inflammatory modulators (COX-2/PGE2, SHP2–ROCK), survival and anoikis resistance pathways (Annexin A2/S100A10), and extrusion polarity controllers (Scribble, APC, PKCι, vinculin). Experimental outcomes define EDAC as a tunable system whose failure—via genetic alterations or inflammatory modulation—may permit basal extrusion, field cancerization, or transition to invasion. The table highlights how mechanistic discoveries in model systems guide potential therapeutic or diagnostic applications in early-stage epithelial cancers

EDAC in human tissues

EDAC is most evident in monostratified epithelia of the pancreas, colon, and breast, where it likely acts as an early tumor-suppressive barrier. Yet, direct human evidence is limited. For instance fallopian tube (uterine tube or oviduct) epithelia with TP53 mutations lack vimentin, a cytoskeletal protein linked to EDAC, questioning its reliability as a universal biomarker [74].

A critical dimension of EDAC research is the difference between monostratified and stratified epithelial architectures, as these structural contexts influence how transformed cells are recognized and expelled. In monostratified tissues (e.g., colon, pancreas), EDAC predominantly manifests as apical extrusion followed by anoikis. In contrast, stratified epithelia such as oral mucosa, esophageal epithelium, or skin may require additional coordination between basal stem-like cells and suprabasal layers to achieve effective elimination. Recent evidence from esophageal models demonstrates that mutant clones can extrude microtumor-like structures into the lumen, a process consistent with EDAC-like surveillance in stratified tissues [65]. Moreover, reviews on epithelial interfaces propose that EDAC principles—such as fitness sensing and contractile extrusion—could extend to stratified systems, where mechanical and cellular heterogeneity may modulate EDAC outcomes [47]. Although direct in vivo evidence in oral mucosa is limited, parallels with esophageal models and organotypic cultures support the notion that EDAC mechanisms are adaptable and could play a similar tumor-suppressive role in stratified epithelia [26]. This suggests that expanding EDAC studies to stratified tissues may uncover unique dynamics of epithelial surveillance with clinical relevance.

Gaps across cancer stages

EDAC research primarily focuses on early tumorigenesis, such as a single transformed cell extrusion within healthy epithelia. Its role in intermediate states, like epithelial dysplasia, or established cancers, remains underexplored. In vitro models excel at dissecting pathways but oversimplify inflammatory and metabolic contexts. Murine studies suggest that chronic inflammation and microenvironmental changes reduce extrusion rates, thereby promoting the retention of transformed cells and clonal expansion [34, 66]. These findings underscore the need to study EDAC across progressive disease stages.

Perspectives

Advancing EDAC’s relevance to human cancer requires moving beyond monocultures. Organoids, 3D co-cultures, and immune-competent models can probe EDAC in physiologically relevant settings. Investigating underrepresented stratified epithelia and later cancer stages will clarify whether EDAC persists, is subverted, or can be reactivated therapeutically. Identifying reliable biomarkers—such as EPLIN, filamin, or S1PR2—in human tissues will boost translational potential. Mapping EDAC’s activity across cancer types and stages may uncover opportunities for early intervention, prophylaxis, or enhanced tissue surveillance.

Clinical implications and future directions of epithelial defense against cancer

EDAC redefines early tumor suppression, offering translational opportunities for cancer prevention and precision oncology. If EDAC failure signals the shift from mutation to malignancy, its molecular components could serve as risk biomarkers or therapeutic targets. Key molecules, including LILRB3, EPLIN, Rab5, and S1PR2, are emerging as candidates for early detection or chemoprevention [34, 35,39, 40, 46, 48, 66, 67, 75, 76]. This section explores EDAC’s clinical potential, challenges from evasion mechanisms, and strategies to integrate EDAC-based approaches into next-generation diagnostics and therapies.

EDAC in pre-malignant fields and clonal expansion

EDAC acts as an early barrier in pre-malignant fields at risk of malignancy, where oncogenic mutations drive field cancerization without morphological changes, as seen in breast, lung, and colorectal cancers [68–70]. By integrating cell-intrinsic mechanisms like apical extrusion and anoikis, EDAC prevents clonal expansion of transformed cells in high-risk epithelial tissues [3, 71]. During dysplastic transitions, a context that can eventually evolve to the establishment of cancer, EDAC should modulate the competition among clones with varying mutational loads, potentially determining which lesions in tissues like the esophagus, cervix, or oral mucosa progress to malignancy. Failure of EDAC, accompanied by disrupted TGF-β signaling, chronic inflammation (e.g., IL-6, TNF-α), or immune evasion) could foster a permissive microenvironment for clonal dominance, which would increase malignancy risk [72, 73]. Biomarkers such as EPLIN, Rab5, S1PR2, and therapies like S1PR2 agonists [35] or COX-2 inhibitors [66], could enhance risk stratification and in theory, restore EDAC to prevent cancer progression.

Recognition mechanisms and biomarker potential

EDAC relies on healthy cells recognizing transformed neighbors, yet no clinical tools detect these events in intact tissues. LILRB3 and S1PR2, critical for intercellular communication, show promise as biomarkers in PDAC and CRC [34, 35]. Pharmacological restoration of EDAC, such as through COX-2 inhibition or stiffness modulation, improves transformed cell clearance by up to 60% in preclinical CRC models [21, 66]. These findings suggest EDAC enhancement is a viable therapeutic strategy.

Challenges of evasion and inverse EDAC

EDAC is not foolproof. Oncogenic cells may evade elimination by gaining super-competitor traits via additional mutations. In advanced stages, EDAC may shift to basal extrusion, promoting dissemination—a phenomenon termed inverse EDAC [41, 72, 75]. Some extruded cells survive, seeding distant sites. High-throughput screens in MDCK models have identified compounds boosting RasV12 extrusion by 40%, hinting at chemopreventive potential [75]. Targeting EDAC pathways could strengthen early cancer control.

Advanced models and clinical validation

Monolayer epithelial cultures, common in EDAC research, lack the architectural and immunological complexity of human tissues. Mouse models with stratified or genetically mosaic epithelia better capture EDAC dynamics in dysplastic and premalignant states. Human-derived organoids, incorporating chronic inflammation and co-occurring mutations, provide robust platforms for validating EDAC-targeted therapies.

Diagnostic and therapeutic horizons

Liquid biopsies could enable non-invasive EDAC monitoring, detecting markers like EPLIN, S1PR2, Rab5, or vesicle-associated proteins in plasma or saliva. Soluble mediators, such as ADAMDEC1, secreted during cell competition, are candidates for PCR-based assays before histological changes appear [48]. As outlined in Table 3, profiling these markers may predict neoplastic progression in epithelial fields. EDAC’s interplay with immunity merits further study: chronic inflammation impairs EDAC, but immune components may enhance it by clearing aberrant clones [47]. This synergy could support immunoprevention in high-risk tissues. While EDAC’s primary translational value lies in early intervention, it could also guide long-term surveillance for patients with high-risk mutations, such as TP53 or APC, or precancerous lesions.

Table 3.

Functional biomarkers of epithelial defense against cancer with diagnostic and mechanistic relevance

Phase	Marker	Cell expression	Mechanism	Knockdown
Recognition	EphA2	Transformed cells	Promotes repulsion	Halves the efficiency of cell competition (45)
Recognition	LILRB3	Normal cell	Recognizes MHC-I	Reduces the zelimination of transformed cells by 70% (34)
Recognition & defense signaling	S1PR2	Normal cell	Activates actinomyosin contraction (extrusion)	Promotes survival and reverses EDAC (35)
Defense signaling	Rab5	Transformed cells	Endocytosis of E-cadherin activates myosin II extrusion	Reduces myosin II activation and contraction (42)
Cytoesketal changes	Filamin A	Normal cell	Accumulates during extrusion	—
Defense signaling	EPLIN	Normal cell	Stabilizes actomyosin ring (extrusion)	Reduces extrusion up to 60% (24)
Defense signaling & paracrine regulation	ADAMDEC1	Normal cell	Promotes the accumulation of filamin and EPLIN; increases apical extrusion	Increases apical extrusion by up to 50% (48)
Anoikis	Annexin A2 / S100A10	Expelled transformed cells	Resists anoikis post-extrusion	—

Summary of molecular biomarkers associated with epithelial defense against cancer (EDAC), categorized by functional phase: recognition, signaling, execution, and post-extrusion fate. Each entry includes the biomarker name, primary cell type in which itis expressed (transformed vs. normal), associated mechanism within the EDAC cascade, and the phenotypic impact of knockdown or disruption, where applicable. This biomarker panel informs potential diagnostic strategies—such as liquid biopsy or early immunohistochemical screening—and therapeutic approaches aimed at restoring EDAC activity in preneoplastic tissues. Functional readouts from these markers may indicate EDAC competence or evasion, aiding in the prediction of progression risk in early-stage epithelial lesions

A key challenge in translating EDAC into therapeutic strategies lies in selectively enhancing apical extrusion without inadvertently promoting basal extrusion, which could favor invasion and metastasis [4, 55]. Pharmacological agents that target actomyosin contractility, such as ROCK or PKC modulators, must be carefully dosed to avoid polarity shifts or excessive cytoskeletal tension that predispose to basal delamination. Moreover, delivery hurdles remain: EDAC-enhancing compounds, including S1P–S1PR2 agonists or EPLIN stabilizers, require localized administration to minimize systemic off-target effects and immune disruption [66]. Addressing these pharmacokinetic challenges through controlled-release formulations, organoid-based preclinical testing, and imaging biomarkers could pave the way for safe clinical translation of EDAC-based therapies.

Recent evidence from human tissues underscores the translational relevance of EDAC-associated proteins. Immunohistochemical studies have revealed that Annexin A2, a key regulator of post-extrusion survival, is downregulated in basal layers of oral dysplasia and head and neck squamous cell carcinoma, correlating with the loss of normal epithelial architecture [76]. Similarly, EPLIN, which stabilizes the actomyosin ring during apical extrusion, shows differential expression between normal and cervical cancer tissues, suggesting that its reduction accompanies early transformation events [77]. Rab5, an important mediator of E-cadherin endocytosis and extrusion dynamics, is markedly activated in oral dysplasia, where it contributes to nuclear β-catenin accumulation and dysplastic progression [78]. Furthermore, S1PR2, essential for sphingosine-1-phosphate signaling in EDAC, displays variable expression in human cancers, with low S1PR2 levels predicting poor prognosis in extrahepatic cholangiocarcinoma [79]. Collectively, these findings highlight that EDAC biomarkers are not only detectable in human precancerous lesions but may also stratify risk and inform early intervention strategies.

The clinical translation of EDAC is still in its infancy, but these studies provide promising proof-of-concept evidence that EPLIN, Rab5, Annexin A2, and S1PR2 can serve as early indicators of epithelial transformation. Their integration into diagnostic panels or liquid biopsy platforms could represent a major step toward EDAC-based cancer interception.

Perspectives

EDAC represents a dynamic epithelial surveillance system. To unlock its clinical potential, research must advance biomarker discovery, develop sophisticated models, and map immune interactions. Understanding EDAC’s failures and reactivation strategies could yield stage-specific prevention and non-invasive diagnostics. EDAC stands as a promising frontier for intercepting cancer before clinical onset.

Building on recent human tissue findings, future perspectives should include the systematic validation of EDAC biomarkers, such as EPLIN, Rab5, Annexin A2, and S1PR2, in precancerous and dysplastic lesions. Incorporating these proteins into prospective clinical studies, including liquid biopsy platforms, could help bridge the current translational gap and facilitate the development of early cancer interception strategies.

Conclusion

Epithelial defense against cancer (EDAC) integrates fitness sensing, mechanical extrusion, and cytoskeletal coordination into a dynamic and multifaceted barrier against early carcinogenesis. Yet, this safeguard is not infallible. In advanced or hostile tissue contexts, transformed cells may evade, resist, or even invert EDAC, converting a protective mechanism into a route for invasion and progression (see Fig. 3). As mechanistic insights into EDAC deepen, its translational potential becomes increasingly apparent—ranging from the identification of early biomarkers in field cancerization to the development of therapies aimed at restoring epithelial surveillance in precancerous lesions. Spanning diverse tissues, oncogenic mutations, and cancer stages, EDAC is emerging not only as a cellular curiosity but as a foundational paradigm for intercepting cancer at its inception. Validating EDAC as a clinically relevant defense mechanism could shift the landscape of early cancer detection and redefine therapeutic strategies during the earliest phases of tumor development.

Fig. 3.

Hallmarks and outcomes of epithelial defense against cancer. This conceptual diagram illustrates the core features and possible trajectories of EDAC, a protective mechanism by which epithelial tissues detect and promote the elimination of transformed cells to preserve tissue integrity. The process involves three coordinated stages: [1] recognition of aberrant neighbors via fitness cues [2], activation of defense signaling in surrounding normal cells, and [3] extrusion of the transformed cell, typically through an apical route. The outcome of EDAC is context-dependent: it may succeed in clearing precancerous cells, fail due to oncogenic adaptation or microenvironmental suppression, or invert—where transformed cells actively displace normal neighbors. These divergent outcomes are shaped by tissue type, epithelial architecture, oncogenic load, and inflammatory or mechanical stress. By integrating molecular triggers with biomechanical execution, EDAC emerges as a dynamic surveillance system with critical relevance to early tumor interception and field cancerization

Abbreviations

EDAC

epithelial defense against cancer

MDCK

Madin–Darby Canine Kidney

MCF-10A

human mammary epithelial cell line 10 A

Caco-2

human colorectal adenocarcinoma cell line

MC38

murine colon adenocarcinoma cell line

PDAC

pancreatic ductal adenocarcinoma

OSCC

oral squamous cell carcinoma

YAP

Yes-associated protein

EPLIN

epithelial protein lost in neoplasm

FBP17

formin-binding protein 17

FAK

focal adhesion kinase

PKA

protein kinase A

PKCι

protein kinase C iota

Cav-1

caveolin-1

SHP2

Src homology region 2-containing protein tyrosine phosphatase 2

ROCK

Rho-associated coiled-coil-containing protein kinase

MAPK

mitogen-activated protein kinase

JNK

c-Jun N-terminal kinase

S1P

sphingosine-1-phosphate

S1PR2

sphingosine-1-phosphate receptor 2

Rab5

Ras-related protein Rab-5

EphA2

Ephrin type-A receptor 2

MHC-I

major histocompatibility complex class I

LILRB3

leukocyte immunoglobulin-like receptor B3

ANXA2

Annexin A2

S100A10

member of the S100 family of proteins, forms complex with ANXA2

APC

adenomatous polyposis coli

ADAMDEC1

a disintegrin and metalloproteinase domain-like decysin-1

N-WASP

neural Wiskott–Aldrich syndrome protein

EMT

epithelial–mesenchymal transition

COX-2

cyclooxygenase-2

PGE2

prostaglandin E2

Scr

scrambled control (or mutant allele, contextdependent)

Cdc42

cell division control protein 42 homolog

RhoA, Rac1

Rho family GTPases

Vimentin

an intermediate filament protein associated with the mesenchymal phenotype

Filamin A

actin-binding scaffold protein

Author contributions

G.R-C and C.R. contributed equally to the conceptualization, drafting, and critical revision of the manuscript.

Funding

This work did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.