Connections between cadherin-catenin proteins, spindle misorientation, and cancer

Abstract

Cadherin-catenin mediated adhesion is an important determinant of tissue architecture in multicellular organisms. Cancer progression and maintenance is frequently associated with loss of their expression or functional activity, which not only leads to decreased cell-cell adhesion, but also to enhanced tumor cell proliferation and loss of differentiated characteristics. This review is focused on the emerging implications of cadherin-catenin proteins in the regulation of polarized divisions through their connections with the centrosomes, cytoskeleton, tissue tension and signaling pathways; and illustrates how alterations in cadherin-catenin levels or functional activity may render cells susceptible to transformation through the loss of their proliferation-differentiation balance.

Keywords:

• Adherens Junctions
• cadherin
• cell polarity
• cancer
• catenin
• signaling
• spindle orientation

Introduction

Since the formulation of the Cell Theory in the 19th century, giant steps in diverse scientific areas have contributed to our current understanding of the function and organization of cells within tissues and how they lead to cancer when gone awry. One major breakthrough that advanced the understanding of the architectural organization of tissues was published in 1963, where Farquar and Palade using electron microscopy analyses defined 3 major intercellular structures as contact points between cells in epithelial tissues.¹ These 3 structures defined as zonula occludens (tight junctions), zonula adherens (adherens junctions) and desmosomes exhibit a paradigmatic organization as complexes formed by transmembrane proteins. Among them, the role of cadherin-catenin complexes at adherens junctions (AJs) in maintaining epithelial homeostasis is the best understood. Traditionally, cadherin-catenin complexes were thought to function as static complexes that anchor to the cytoskeleton via cytoplasmic scaffolding proteins, thereby sustaining tissue architecture. However it is now well acknowledged that cadherin-catenin complexes present a dynamic organization and the fundamental paradigms about their mechanical and signaling functions have been constantly subjected to active remodeling. Remarkably, their constitutive and associated proteins can also function outside of the realm of intercellular adhesion. These seemingly jack-of-all-trade proteins are involved in the control of cell polarity, transcription, signaling, proliferation, cell fate and migration.^2-4

E-cadherin is the prototypical member of the family of type I classical cadherins, which are transmembrane proteins that through their extracellular domain promote adhesion between adjacent epithelial cells in a calcium dependent manner.⁵ The intracellular domain associates to β-catenin and p120-catenin, while α-catenin binds to the complex via β-catenin.³ Cadherin-catenin complexes present a dynamic stability, which allows the modulation of adhesion strength to preserve adult tissue homeostasis and the reshaping of epithelial tissues to drive morphogenetic movements during development. The mechanism underlying this key function is the association of catenins with the actin and microtubule networks via actin- and microtubule- binding proteins.^6,7 Several studies have provided evidence that alterations in the expression and/or functional activity of the cadherin-catenin complex lead to tumor initiation and progression. This is a consequence not only of decreased cell-cell adhesion, but also of alterations in signaling cascades, and loss of cell contact inhibition of growth.⁸ Moreover, loss of E-cadherin expression, together with upregulation of N-cadherin, is associated with loss of epithelial characteristics and increased cell migration: a process known as epithelial to mesenchymal transition (EMT). All these processes have been thoroughly delineated in excellent reviews elsewhere.^9-12 Less appreciated are the implications of cadherin-catenin complex malfunctioning for the polarized orientation of cell divisions. In this review, we focus on the emerging roles of cadherin-catenin complex proteins in the regulation of oriented cell divisions in mammalian epithelia. First, we briefly describe the polarized features of epithelia to later highlight the novel connections of AJ proteins with centrosomes and spindle positioning during cell division. This is followed by a discussion on the potential role of AJ/actin complexes as mechanosensors that promote polarized cell divisions. Finally, we discuss how when these connections go awry may render epithelial cells more susceptible to transformation, cancer progression and tumor maintenance.

Cell polarity and polarized cell divisions

The correct establishment and maintenance of cell polarity is crucial for cell physiology and tissue homeostasis, and occurs in response to cell intrinsic and extrinsic cues.^13,14 At the cellular level, cell polarity involves the acquisition and maintenance of a high level of cellular organization, including a differential distribution of proteins and lipids along the plasma membrane, and the spatial positioning of organelles in the cytoplasmic space, such as the centrosome, the Golgi apparatus and the nuclei.¹⁵

A robust polarized distribution of proteins and organelles is also maintained during cell division. Polarized cell divisions can be symmetric, when one cell gives raise to 2 identical daughter cells, or asymmetric, when a dividing cell gives rise to 2 daughter cells that are different in molecular composition, cell size, developmental potential and/or cell fate.^14,16,17 Although it still debated for some mammalian tissues, asymmetric cell divisions are particularly relevant in certain adult epithelial stem cell compartments such as the intestine,^18,19 mammary epithelia,^20,21 lung^22,23 and skin,^24-28 where the orientation of cell divisions or the final positioning of the daughter cells have major implications in the self-renewal or differentiation properties of progenitor cells, thereby regulating their homeostatic regeneration (Table 1). Given the implications of neoplastic alterations in some stem cell compartments to promote tumor initiation and cancer maintenance, or the acquisition of self-renewal characteristics of differentiated cells, it has been proposed that alterations that favor an increase in symmetric cell divisions could serve as a mechanism to expand the cancer stem cell pool.^14,17,29 Indeed, the connection between increased symmetric cell divisions and the expansion of cancer stem cells has been observed in some tumors such as breast,^20,30 intestine,^31,32 lung,³³ and skin carcinomas.³⁴

Table 1. Spindle position and cellular fates in different mammalian epithelial tissues

Epithelial tissue	Position of the spindle	Cell fate	References
Epidermis (stratified)	1. Symmetric (parallel to the basement membrane)	1. Both daughter cells remain as progenitors in the basal layer	^24-28
	2. Asymmetric (perpendicular to the basement membrane)	2. One daughter cell is positioned suprabasally and initiates terminal differentiation via Notch activation
Intestine (simple)	Oriented cell divisions can be planar or perpendicular to the apical surface of the cell	Cellular fates are determined by the final position of the daughter cells in the intestinal crypt	^18,19
Mammary gland (simple)	1. Parallel to the basement membrane	1. Basal lineage	^20,21
	2. Perpendicular to the basement membrane	2. Luminal lineage
Embryonic lung epithelium (simple)	Parallel and perpendicular to the basement membrane	? Perpendicular divisions lead to asymmetric inheritance of Numb	^22,23

The cellular events that occur during polarized cell divisions involve the organization of intracellular components in relation to the axis of cell division and the basal membrane. This includes an asymmetry of duplicated centrosomes between the mother and daughter cells, the orientation of the mitotic spindle, and the distribution of cell polarity and cell fate determinants. The centrosome functions as the major microtubule-organizing center (MTOC) both in interphase cells and during mitosis.^35,36 During mitosis, 3 types of microtubules nucleate from the centrosomes: kinetochore microtubules that attach to the chromosomes, polar microtubules, and astral microtubules that extend to the cell cortex.³⁷ Centrosomes are composed of 2 centrioles surrounded by a pericentriolar material, which contains numerous proteins necessary for microtubule nucleation (γ-tubulin) and anchoring (e.g. Ninein or the motor protein Dynein).^35,36 During cell division the centrosome duplicates at the G1/S phase of the cell cycle, and as cell cycle progresses one centrosome migrates to the opposite pole of the dividing cell to form the polar mitotic spindle.^17,38,39 The oriented separation/positioning of centrosomes is controlled by several kinases of the cell cycle including members of the NEK, Aurora, and Polo kinase families.⁴⁰ This is also regulated by a crosstalk with cell polarity determinants promoting the polarized inheritance of the mother or the daughter centrosome by the mother and daughter cell.⁴¹ Interestingly, as discussed below, catenin proteins have also been recently associated to the centrosome and the regulation of some of the kinases involved in centrosome duplication.

The orientation of the mitotic spindle requires the interaction of the spindle with cortical sites, and a coordinated crosstalk with the polarity machinery.⁴² Many of the polarity factors involved in mitotic spindle orientation have been described in Drosophila and are generally conserved in mammalian cells.^14,16,17,43 Their role in spindle positioning and asymmetric cell division has been examined in reviews to which we refer the readers for more comprehensive information.^{14,16,17,42-44} Briefly, before an asymmetric cell division occurs, the apical Par polarity proteins Par3/Par6/atypical protein kinase C (aPKC) localize to the apical cell cortex, along with the Gαi subunit of heteromeric G proteins. aPKC requires the small G-protein Cdc42 for its apical localization and activation. During mitosis, Gαi interacts with proteins associated with astral microtubules at the spindle pole, LGN and NuMA, which are physically linked by the adaptor protein Inscuteable in a mutually exclusive manner.^45,46 This complex is also associated with the motor complex Dynein/Dynactin, which generates the force to pull astral microtubules and the centrosome toward the apical cell cortex, ensuring that the mitotic cleavage plane is perpendicular to the apical-basal axis. The cleavage plane then influences the identity and fate adopted by the 2 daughter cells since it is coupled with the asymmetric distribution of cell fate determinants. The Gαi complex also partakes in planar epithelial divisions of epithelial monolayers.^47-49 In this case, the Gαi complex recruits Dynein-dynactin to the lateral cortex, which pull spindle poles toward the lateral side of the dividing cells. In certain cell types aPKC plays an active role excluding LGN from the apical domain and restricting it to the lateral cortex.^{47,50 48}

How cells choose their axis of division has been a matter of intense investigation. Recently cadherins are emerging as components of the polarizing machinery during cell division in some cells and tissues. Hence, it is tantalizing to speculate that cadherins and their connections with the cytoskeleton may regulate the position of the mitotic spindles.

Links between cadherin-catenins and positioning of mitotic spindles

The direct functional involvement of AJs in the maintenance of tissue integrity makes it difficult to distinguish the contributions of AJs to organelle positioning from a general disruption of epithelial architecture when AJ proteins are lost or dysfunctional. However, the direct contributions of cadherin-mediated contacts in promoting intracellular asymmetry have been recently substantiated in various mammalian cell types in culture.^51-53 In these studies, it was observed that cadherins control the positioning of the nucleus and centrosomes of cells in interphase,^51,52 and the spindle orientation of dividing cells.⁵³ In the context of organisms, the best examples of the contributions of cadherin-mediated adhesion to intracellular asymmetry and oriented cell divisions have been obtained from studies in Drosophila and Caenorhabditis elegans. For example, in the Drosophila ovary⁵⁴ and in the male germ stem cell niche,⁵⁵ germ stem cells differentiate precociously when the levels of E-cadherin are reduced or absent and stem cells are no longer maintained within their niche. Interestingly, in the male germline stem cell niche, E-cadherin contributes to centrosome and spindle positioning.⁵⁵ In addition, the development of the Drosophila neuroepithelium and the sensory organ depends on the AJ-mediated regulation of the distribution of polarity determinants and the orientation of asymmetric cell divisions.⁵⁶ As a final example, it has also been observed that the ortholog of β-catenin in C.elegans controls cell division orientation in early embryos.⁵⁷

In mammals, a connection between AJ proteins and intracellular asymmetry during cell division and cell fate has been observed in certain tissues, but mostly characterized in stratified epithelia. For example, in embryonic neural stem cells, it has been documented that AJs are organized into different microdomains that are split unequally during asymmetric cell divisions by the cleavage plane.⁵⁸ The inheritance of cell fate determinants together with reduced levels of AJs may explain the posterior detachment of the cells that undergo differentiation. Moreover, robust levels of N-cadherin in progenitor cells support their maintenance in their niche by the activation of β-catenin signaling.⁵⁹ In simple epithelia, it has been proposed that mutations in E-cadherin correlate with an increase in symmetric cell divisions and the expansion of the cancer stem cell pool.⁶⁰ In stratified epithelia such as the skin, the absence of α-catenin in the basal progenitor cells of the epidermis leads to reductions of AJs, loss of the cortical distribution of polarity determinants and randomized orientation of mitotic spindles.²⁴ In the epicardium, absence of β-catenin leads to a disruption of AJs and a randomization of mitotic spindle orientation.⁶¹ These results suggest that AJs may play an active role in the regulation of oriented cell divisions promoting the occurrence of asymmetric cell divisions in certain tissue types. However, as opposed to Drosophila male germ cells, neuroblasts, and sensory organ cells, in Drosophila follicle cells mitotic spindles are not aligned with AJs and reductions on cadherins do not result in spindle misorientation.⁶² A similar scenario was described in Drosophila imaginal discs and in Xenopus embryonic epithelia.^63,64 In mammals, absence of E-cadherin in mouse skin and mammary progenitor epithelial cells does not lead to an expansion of the stem cell compartment,^65-67 suggesting the involvement of additional regulatory mechanisms. Clearly, more insights about the role of cadherin-catenin proteins are needed to understand the extent to which AJs contribute as spatial cues in the regulation of centrosome and mitotic spindle positioning in different tissues and species.

Potential insights into cadherin-catenin mediated positioning of mitotic spindles: known interactions between AJs, centrosomes and microtubules

During cell division, AJs undergo a dynamic remodeling thereby allowing the regenerative process of epithelial tissues.^68,69 This has been better defined during cytokinesis and abscission of the 2 nascent daughter cells. In Drosophila, it has been observed that the levels of AJs are reduced at the cleavage furrow of dividing cells.^70-72 However, in mammalian epithelia such as MDCK cells, intestinal crypt cells, or basal progenitor keratinocytes AJs are maintained during this process.^73,74 It will be interesting to explore how this process is regulated in different cell types and tissues, and the implications of adhesion disengagement and the formation of new junctional contacts for the regenerative properties of progenitor cells. While the dynamic organization of AJs is modulated during cytokinesis, recent reports have unveiled a role for cadherin-catenin proteins in the position of the mitotic spindle and the organization of centrosomes. The potential mechanisms at hand may function through their interactions with the microtubules and centrosomes. Typically, microtubules nucleate from the centrosome and their dynamic growing plus-ends explore the periphery of the cell. In 1986, Kirschner and Mitchison proposed a “search and capture” model for the attachment of microtubules to cortical sites.⁷⁵ This model proposes that dynamic microtubules can be captured at specific sites, such as the membrane, the kinetochores or cell adhesion sites. In interphase cells, microtubules minus- and plus-ends interact with AJs through microtubule binding proteins,^76-78 including ACF7,⁷⁹ APC,^80,81 CLIP170,⁸¹ Dynein,^82,83 CLASP2,^84,85 and the microtubule minus-end binding protein Nezha.⁸⁶ Whether all these connections between AJs and microtubule binding proteins take place during cell division awaits further investigation.

An interesting candidate is the microtubule motor Dynein. During mitosis it localizes to spindle poles and astral microtubules and generates the pulling force to orient the mitotic spindle; whereas in interphase it binds to β-catenin at AJs.⁸² This has raised the possibility that β-catenin promotes the proper positioning of centrosomes and mitotic spindles by anchoring astral microtubules to the cell cortex. It has also been recently shown that β-catenin localizes to interphase centrosomes and spindle poles, where it promotes centrosome separation and spindle formation.^87-91 The underlying mechanism involves the interaction of β-catenin with NIMA-related protein kinase 2 (Nek2), a protein with critical roles in centrosome separation.⁹¹ The mitotic protein polo-like kinase acts upstream of Nek2, which in turn phosphorylates β-catenin promoting its stabilization and preventing its degradation.⁸⁹

Another candidate to link AJs to centrosome and spindle positioning is the adenomatous polyposis coli protein (APC). APC binds to β-catenin as well as to the microtubule binding protein EB1.^92,93 APC localizes to kinetochores, centrosomes and cell adhesion sites.^94-96 Interestingly, in the budding yeast, Kar9 (the proposed functional homolog of APC) is required for the attachment of astral microtubules to the cortex to orient mitotic spindles.⁹⁷ In Drosophila neuroepithelial cells and germline stem cells, the orientation of the mitotic spindles is regulated in an APC dependent manner.^98,99 In addition, APC and β-catenin function in the canonical Wnt signaling pathway.¹⁰⁰ A link between Wnt signaling and spindle orientation was recently uncovered in mouse embryonic stem cells. Habib et al. (2013) observed that a localized Wnt signal leads to asymmetric spindle orientation, partitioning of cell fate determinants, asymmetric inheritance of β-catenin and differential fates of the 2 daughter cells.¹⁰¹ It is tempting to speculate that in this context spindle orientation is controlled through the APC-β-catenin interaction, but the precise molecular mechanism remains to be explored.

It is well accepted that inactivating mutations of APC and stabilization of β-catenin lead to cancer.^102,103 Their emergent roles in spindle orientation and cell fate regulation add additional ways by which these proteins contribute to cancer progression. APC has also been linked to genetic instability.^94,95 The heterozygous loss of APC is sufficient to cause spindle misorientation in gastric tissues leading to a pre-tumorigenic state,^18,104 while inactivation of both APC alleles is required for carcinogenesis.¹⁰⁵

p120-catenin has also been found at centrosomes, associated to kinesin motors, the microtubule network and mitotic spindles.^106,107 At centrosomes, p120-catenin colocalizes and associates with cyclin E and cyclin-dependent kinase 2, key proteins involved in centrosome duplication during mitosis.¹⁰⁸ Interestingly, the transcription factor Kaiso, originally described as a binding partner of p120-catenin,¹⁰⁹ also localizes to centrosomes, mitotic spindles and the midbody,^110,111 and colocalizes with p120-catenin at spindle poles in HeLa cells.¹¹⁰ Through these novel connections, and the roles of p120-catenin in the regulation of RhoGTPases and the cytoskeleton,^112,113 p120-catenin may potentially regulate centrosomal and mitotic spindle functions and integrity. Indeed decreased levels of p120-catenin in epidermal cells lead to alterations in centrosomes and spindles, mitotic defects and aneuplody in a cell autonomous manner.¹¹⁴ In this context, it has been recently found that p120-catenin interacts with the microtubule binding protein CLASP2 at AJs in epidermal basal cells.⁸⁴ Whether p120-catenin interacts with CLASP2 at interphase or mitotic centrosomes awaits further investigation. It is tempting to speculate that given the pivotal roles of CLASP2 in mitosis,^115,116 a potential CLASP2-p120 interaction at centrosomes could have an impact on centrosome function and mitotic spindle position and integrity. Interestingly, it has been documented that CLASP1, another member of the family, cooperates to align the spindle along the long axis of division in mammalian cells.^117,118

Overall, these recent results point to another interesting facet of AJs, in which their proper levels of expression may also regulate cell division through centrosome and spindle organization.

Cadherin-catenins, actin cytoskeleton and spindle orientation

The dynamic anchoring of the mitotic spindle to the cell cortex by astral microtubules underlies most of the mechanisms that orient cell division relative to the shape of the cell or to cortical regulators.^42,119 To allow the alignment and separation of the mitotic spindles, 3 major elements partake: 1) cell cortical tension, 2) microtubule dynamics and 3) a restoring force preventing the collapse of the mitotic spindle on the cortex.^42,120 It was originally described in the budding yeast that disruption of either astral microtubules or actin function results in improper spindle orientation. These results sparked the notion that proper spindle orientation is maintained by directly or indirectly tethering astral microtubules to cortical actin to generate a pulling force on the spindle.¹²¹ The involvement of cortical actin in the orientation of spindles has also been recently observed in mammalian tissues. For example, in mouse epidermal cells, the absence of the serum response transcription factor SRF, which induces actin polymerization and regulates the expression of several actin and actin binding proteins, leads to reductions in cortical actin, alterations in the distribution of LGN, NuMA and randomized positioning of mitotic spindles.²⁵

It is well established that cadherin-catenin proteins integrate cell-cell adhesion with cytoskeletal dynamics to establish and maintain tissue architecture. Reductions or alterations in the functional activity of cadherin-catenins lead to defects in the organization of both actin and microtubule cytoskeletal networks. Studies in mammalian cell lines indicate that the cadherin-mediated regulation of centrosome and spindle positioning requires both actin and microtubule cytoskeletons.^51,52 This suggests that cytoskeletal dynamics and forces are involved in centrosome and spindle positioning in a manner directly associated to AJs. α-catenin plays a central role in recruiting a number of proteins that link cadherin-catenin complexes to cortical actin, including vinculin, and the Arp2/3 complex actin nucleator.^122-125 The consequences of reduced α-catenin for the orientation of mitotic spindles are exemplified in the mouse epidermis, in which absence of α-catenin results in the loss of LGN localization to the cell cortex, random NuMA crescents and misoriented spindles.²⁴ This is also accompanied by reductions in intercellular adhesion, loss of epithelial architecture and hyperproliferation. Another interesting candidate that may be involved in spindle orientation is the Arp2/3 complex. It acts at nascent contacts upon homophilic cadherin binding and marks sites for actin assembly at the cell surface.¹²⁶ In cooperation with the nucleator factor WAVE,¹²⁷ it is also necessary for junctional integrity and contractile tension at AJs in a process coupled to myosin driven contractility.¹²⁸ Recently, it has been shown that loss of Arp2/3 activity in the epidermis leads to an increased proliferation of basal progenitor cells, alterations in terminal differentiation, disorganization of epidermal architecture and impaired Tight Junction assembly and function.^129,130 Future studies will determine whether spindle positioning is altered in the absence of the Arp2/3 complex.

Another interesting link between AJs, the actin cytoskeleton and oriented cell divisions in the epidermis is the interaction of α-catenin with the neurofibromatosis-2 (NF2) tumor suppressor gene Merlin, a member of the FERM (Four-point-one/Ezrin/Radixin/Moesin) domain family of proteins. It has been observed that Merlin directly links α-catenin with Par3, allowing the formation of a cortical actin ring and regulating cell polarity and the orientation of cell divisions within the basal epidermis.¹³¹ Whether the observed defects triggered by the loss of α-catenin, Arp2/3 or other α-catenin associated proteins are a consequence of spindle misorientation awaits further investigation. However, given their roles in cytoskeletal organization, an interesting possibility is that they control spindle orientation by regulating tissue tension.

Emerging roles of cadherin mediated adhesion in growth control and differentiation: Possible links to cortical tension and mitotic spindle orientation

Cells have the ability to sense their physical environment and translate mechanical forces into intracellular signals that regulate cell behavior and tissue homeostasis.¹³² These mechanical signals are able to influence a plethora of cellular responses that range from changes in cell shape, proliferation, migration to the acquisition of a different cell fate.¹³³ Alterations that perturb mechanosensing (the ability of a cell to sense its physical environment) may result in evasion of growth inhibitory signals and alterations in differentiation, potentially endowing cells with neoplastic characteristics.¹³⁴ Mechanical inputs are sensed by cells and translated by mechanosensors in an equivalent way to the ligand-receptor binding events that initiate signal transduction cascades and influence cell behavior.¹³⁵ The best studied mechanosensors are integrins at focal adhesions, which transduce physical cues via the actin cytoskeleton through a number of adaptor proteins, including vinculin.¹³⁶ Over past years, it is becoming increasingly clear that cadherin-catenin proteins are also able to respond to the mechanical environment through their connections with the actin cytoskeleton, allowing groups of cells to behave as a coordinated tissue.^2,137 Recently, several elegant studies have demonstrated that cadherins are bona-fide mechanosensors that transduce physical forces to the actin cytoskeleton.^138-140 These studies have unveiled how mechanical stresses trigger a conformational change in α-catenin at cadherin complexes allowing its direct interaction with the actin cytoskeleton¹⁴¹ or with vinculin, which in turn associates directly with the actin cytoskeleton,^138,140 leading to the remodeling of AJs.¹⁴²

Is AJ mechanosensing activity a contributor to spindle orientation and what are the potential molecular players? Although this field is still in its infancy, an exciting possibility relates to the finding that AJs act as upstream regulators of the transcription factor Yap1. Yap1 is a major effector of the Hippo signaling pathway, which regulates growth, organ size and tumorigenesis.^143,144 Briefly, the activation of the Hippo pathway leads to Yap1 phosphorylation and its cytoplasmic retention. Conversely, the inactivation of Hippo signaling allows Yap1 to enter into the nucleus to activate transcription. Yap1 can also be activated by Hippo-independent mechanisms such as mechanical cues of the physical environment, including changes in cell shape, and tissue tension.^145-147 It is important to mention that it is not yet well understood if Hippo/Yap1 are directly involved in the regulation of spindle orientation. However, it has been observed that cell polarity regulators,^148–150 and the organization of the actin cytoskeleton influence the localization of Yap1.^151,152 Interestingly, loss of Yap1 has been associated to centrosomal and mitotic spindle defects, by a mechanism that involves cyclin-dependent kinase CDK1 mediated phosphorylation,¹⁵³ which is also important to regulate mitotic spindle bi-orientation.¹⁵⁴ Moreover, members of the Hippo pathway are important to orient the plane of cell division in certain tissues such as in the developing kidney.^155,156 In addition, upstream regulators of the Hippo/Yap pathway such as the tumor suppressor LKB1,^157,158 AMPK¹⁵⁹ and TAO1 kinase^160,161 are involved in spindle orientation.

The connections between AJs and Hippo/Yap1 have been observed in several contexts. In Drosophila, the AJ associated protein Echinoid regulates the activation of the Hippo pathway and inhibits cell growth.¹⁶² Additionally, the LIM protein Ajuba that associates to α-catenin and the actin cytoskeleton¹⁶³ has been recently identified as an inhibitor of the Hippo pathway.¹⁶⁴ In the context of cultured mammalian cells, it has been recently shown that E-cadherin ligation can directly activate the Hippo pathway independently of other types of interactions,¹⁶⁵ which strongly positions AJs as important regulators of the pathway. In terms of cell fate in mammalian organisms, one of the most interesting examples is the formation of the first 2 cell lineages of the mammalian embryo, the inner cell mass and the outer epithelial trophectoderm. During this process, the TJ protein Angiomontin is found localized at AJs in nonpolar inner cells of preimplantation embryos, where it activates the Hippo pathway preventing differentiation. In contrast, in outer epithelial cells Angiomontin localizes to the apical domain, suppressing Hippo signaling and activating the trophectoderm differentiation program.^166-168 In the context of expansion of stem cell pools, a link between Yap1 and α-catenin in the control of the proliferation of multipotent epidermal progenitor basal cells has been recently observed. Specifically, α-catenin inactivates Yap1 in the epidermis and controls epidermal proliferation in a cadherin independent manner.^169,170 Remarkably, the absence of α-catenin and an increase in nuclear Yap1 levels were found to be a feature of squamous cell carcinomas.¹⁶⁹ Thus, this positions α-catenin as an upstream regulator of the Hippo pathway effector Yap1 in skin.

Overall, the connection of AJs and cortical tension to the Hippo/Yap1 signaling pathway provides a mechanistic link to the regulation of cell fate decisions in response to architectural and mechanical cues. Whether these connections converge to potentially regulate oriented cell divisions in some tissues is still ill understood. Future research will shed light into this tantalizing hypothesis, and its implications for tumor formation and growth.

Conclusions and Future Directions

The molecular mechanisms regulated by cadherin-catenin molecules in epithelial tissues are not yet completely understood. Much remains to be done to understand how, through a plethora of different connections, AJs integrate and transmit signals to sustain cell physiology and tissue homeostasis. It is well established that cadherin dysfunction is a major contributor to cancer progression and transformation to malignancy.^9-12 However, the molecular mechanisms by which cadherin-catenin proteins regulate the cell proliferation and differentiation balance, beyond their canonical adhesive roles, continue expanding over the years. In this review, we aimed to summarize the emerging implications of cadherin-catenins in the regulation of polarized divisions through their connections with the centrosome, microtubule and actin cytoskeletons; and their implications in cell physiology, tissue function and neoplastic transformation.

Although the precise mechanistic events are still unclear, the emerging picture suggests that AJ levels are important regulators of cell fate in certain tissues and organisms, by maintaining cellular asymmetries and regulating stem cell spindle orientation (Fig. 1). Through their novel connections to the centrosome, β- and p120-catenin are emerging as regulators of the centrosomal cycle, fundamental for the proper inherence of centrosomes between daughter cells. These observations could explain how their loss leads to apoptosis (β-catenin deficiency) or genetic instability and aneuploidy (p120 deficiency), and increases the susceptibility to oncogenic transformation. In the context of polarized cell divisions, AJs can potentially regulate the orientation of cell divisions by anchoring astral microtubules to the cell cortex through their documented associations with microtubule binding proteins. Moreover, through their bona-fide roles as mechanosensors, AJs via α-catenin and its interaction with the actin cytoskeleton and actin binding proteins, integrate mechanical signals to possibly control cell fate decisions. This also includes the emerging role of AJs in the regulation of cell proliferation via the Hippo/Yap1 pathway. It will be interesting to learn in the future whether these events converge to coordinate the orientation of mitotic spindles and regulate polarized cell divisions in stem cell compartments. Future research will shed light on how the complex repertoire of molecular functions in which cadherin-catenins partake may synergistically contribute to the expansion of cancer stem cells pools in different ways in specific cells and tissues.

Figure 1. Links between Adherens Junctions, centrosomes and the cytoskeleton during interphase and cell division. (A) In interphase, classical cadherins mediate homophilic cell-cell adhesion through their extracellular domains. The cadherin cytoplasmic domain binds to β- and p120-catenin, while α-catenin binds to the complex via β-catenin. Cadherin-catenin complexes are linked to the actin and microtubule networks via actin- and microtubule- binding proteins. (B) During cell division, centrosomes duplicate and one of the centrosomes migrates to the other pole of the dividing cell forming the mitotic spindle (not shown), which is anchored to the cell cortex through the astral microtubules. Both β- and p120-catenin are found in centrosomes and at the cell cortex potentially promoting the proper positioning of centrosomes and mitotic spindles by anchoring astral microtubules to the cell cortex through their interaction with microtubule binding proteins. Moreover, through their mechanosensing functions, cadherin-catenin complexes may transmit the forces exerted on astral microtubules pulling the spindles toward their final position, leading to either symmetric or asymmetric divisions through their connections with the actin cytoskeleton and/or the Hippo effector Yap1. (C) A proper balance between symmetric and asymmetric cell divisions needs to be maintained in order to preserve tissue homeostasis. Loss of the expression or functional activity of cadherin-catenin proteins may perturb centrosome organization and the oriented positioning of mitotic spindles rendering cells susceptible to transformation. Microtubules and microtubule binding proteins are shown in blue; actin filaments and actin binding proteins are shown in green, and Adherens Junction proteins are highlighted in orange. Actin BP: actin binding proteins.

In summary, the connection of AJ proteins to centrosomes and mitotic spindles, their roles as mechanosensors and their connection with the Hippo/Yap1 signaling, are unraveling additional ways by which AJs preserve tissue homeostasis, and novel mechanisms by which their loss fosters tumor growth.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Funding

M.N.S. is a recipient of a Fundación Ramon Areces postdoctoral fellowship. M.P-M laboratory is funded by a grant from the Spanish Secretary of Science and Innovation (BFU2012–33910).