Connexins and the gap in context

Abstract

Gap junctions (GJ) can no longer be thought of as simple channel forming structures that mediate intercellular communication. Hemi-channel and channel-independent functions of connexins (Cxs) have been described and numerous Cx interacting partners have been uncovered ranging from enzymes to structural and scaffolding molecules to transcription factors. With the growing number of Cx partners and functions, including well-documented roles for Cxs as conditional tumor suppressors, it has become essential to understand how Cxs are regulated in a context-dependent manner to mediate distinct functions. In this review we will shed light on the tissue and context-dependent regulation and function of Cxs and on the importance of Cx-interactions in modulating tissue-specific function. We will emphasize how the context-dependent functions of Cxs can help in understanding the impact of Cx mis-expression on cancer development and, ultimately, explore whether Cxs can be used as potential therapeutic targets in cancer treatment. In the end, we will address the need for developing relevant assays for studying Cx and GJ functions and will highlight how advances in bioengineering tools and the design of 3D biological platforms can help studying gap junction function in real time in a non-intrusive manner.

---

Insight, innovation, integration

Gap junctional communication is the least appreciated contributor to tissue specific function. This is partly due to the prevalent notion that gap junctions are simple pore forming channels between cells. Recently, the discovery of channel-independent functions of gap junctions, gap junction-independent roles of connexins, a multitude of connexin-interacting partners, and connexin conflicting roles in diseases including cancer, re-ignited interest in this field. Here, we will review these findings and highlight the importance of studying gap junctions and connexins in a context-dependent manner. We will also discuss the importance of systems biology approaches to studying connexin roles and how the advent of bioengineering tools will help in developing relevant experimental platforms for deciphering the complexity of connexin function.

Introduction

Integration of function in the tissue is achieved by the intricate interplay of signals between cells and their surrounding microenvironment in response to systemic cues.¹ Among micro-environmental factors that orchestrate tissue function, cell–cell interactions and communication are paramount for coupling the components of the tissue as a system. Gap junctions (GJ) are specialized intercellular channels that permit functional coupling of cells electrically and biochemically, allowing the selective diffusion of ions and metabolites between the cytoplasms of two adjacent cells. GJs in vertebrates are formed largely by the connexin (Cx) family of proteins. Cxs oligomerize into a hexameric hemichannel termed connexon, and two connexons dock in apposing membranes to form the GJ channel.^2,3 In mammals, Cxs are encoded by at least 20 different genes, which exhibit tissue and cell-type specific, but overlapping patterns of expression.^4–6 All Cx subtypes share a common structure, having four trans-membrane domains with two extracellular loops, an intracellular loop and the cytoplasmic C- and N-termini. Identical or different Cxs can oligomerize to form homomeric or heteromeric connexons respectively. When matching connexons from apposing membranes dock, the GJ channel is termed homotypic while non-identical connexons make heterotypic channels. Why would such a large number of Cxs be needed to generate a variety of GJ channel assemblies if they are to provide a universal type of intercellular communication? One reason is that different Cxs display unique functional properties within GJs; the type and stoichiometry of Cx distribution within GJs control the gating properties and the permeability of the channel, thus, regulating selectivity and rate of molecules passing through GJs.⁷ Another reason is that Cxs exhibit diverse roles in cells which are channel-independent and often mediated by a myriad of Cx binding partners that range from enzymes and signaling molecules to structural proteins and intercellular adhesion components.^8–12 The term “Cx proteome” was coined by Laird's laboratory to designate the heterogeneous pool of Cx-associated proteins which are still being discovered.^10,13 (refer to Scheme 1 for an illustration of a Cx protein, Connexon and a GJ channel and complex).

Scheme 1 From connexin to gap junction complex. The structure of the connexin protein with its 4 transmembrane (TM) domains, two extracellular loops (E1, E2), cytoplasmic loop (CL) and NH₂ and COOH termini is shown in (A). Six connexins (Cx is represented as a cylinder) cluster into a connexon (B), and two connexons dock together to form a gap junction channel (C). Gap junctions clustered in plaques and the gap junction complex made up of connexins and their associated proteins is shown in (D).

Cx gene mutations have been linked to at least nine different syndromes, such as deafness, skin diseases, oculodentodigital dysplasia, Vohwinkel syndrome, X-linked Charcot-Marie tooth among others^14,15 In addition, since Loewenstein and Kano¹⁶ reported that gap junctional intercellular communication (GJIC) is compromised in cancer cells, several studies investigated the role of GJIC in cancer progression and the abnormal expression of Cx in cancer. GJIC was shown to be either reduced in some cancer models or enhanced in others, depending on the type and stage of the tumor, and Cxs have been proposed to possess either tumor suppressing or tumor promoting abilities, depending on their cellular context.^12,15,17

How do Cxs contribute to cell-, tissue- and/or stage-dependent functions? And how does the context in turn dictate Cxs' versatile roles? Currently, there is no unifying model to describe the complexity ascribed to Cx and GJ function in development and diseases, nor should there be. However, there is a true deficit in the integration of information gathered about Cxs. The deficit lies mainly in that most culture-based systems used to study Cxs and GJs have addressed questions about Cxs in limited and sometimes artificial settings, such as forcing exogenous Cx expression in Cx-deficient cells to study their mechanisms of action. While such studies yielded important information about Cx dynamics within a cell, the need for model systems that can probe for endogenous Cx function under physiologically relevant conditions remains a necessity. The use of 3D-culture assays and co-culture systems that provide a more in vivo—like platform for experiments, in addition to transgenic and knock-in mouse models, have greatly advanced our knowledge of Cx and gap junction function in physiologically relevant contexts.^15,18–20

In this review we will touch upon the current and new findings in the Cx and GJ field. However, the primary focus of this article is to integrate the information on the multiple Cx functions and interactions and illustrate the implications of this functional heterogeneity on tissue- and cell-type specific function. We postulate that tissue context determines Cx expression, interactions and role and that Cxs in turn modulate and maintain tissue function, in a dynamic and reciprocal fashion as has been described by Bissell for cell-extracellular matrix (ECM) interactions.²¹ Ultimately, we aim to explore the potential use of Cx and GJIC as therapeutic targets in cancer, especially given their possible contradictory roles as either tumor suppressors, or facilitators of tumor metastasis dependent on the tumor context.

We will also address how the advent of systems biology approaches can aid in deciphering the complexity of Cx regulation and function in real time and under conditions that recapitulate tissue and organogenesisde novo. We will touch upon the importance of novel experimental tools in advancing research in the field such as bioprinting and microfluidics-based tools.

Cxs in the normal context

Channel-dependent functions

At the most basic level, GJs perform a highly specialized and unique role in the organism: they provide direct, yet selective, intercellular communication routes that allow the propagation of coordinated responses across cells, such that tissues are linked as a unit, electrically and metabolically.²² Electrical coupling via GJs is essential for synchronous contraction of cardiomyocytes and the rhythmic pumping of heart, uterine smooth muscle contraction during labor and the unidirectional beating of tracheal cilia to expel fluids from the airway passages.^23–28 In non-excitable tissues, GJIC has been shown to promote embryonic growth, bone modeling, alveolar differentiation and mammary epithelial differentiation among others.²⁹ Many studies highlighted the role of GJIC in promoting differentiation. For example, it was found that restoration of GJIC in colon cancer cells leads to re-establishment of a differentiation phenotype.³⁰ In another study using mouse mammary epithelial cells it was shown that enhanced GJIC induces partial differentiation of mammary epithelial cells in the absence of an exogenously provided basement membrane.³¹ GJIC occurs between similar and different cell types within a tissue and many reports have characterized hetero-cellular GJs in vivo and in culture. For example, heterocellular GJIC between Type I and Type II epithelial alveolar cells of the lung promotes lung epithelial differentiation in rat.³² Hetero-cellular GJIC has also been described between cardiac myocytes and surrounding fibroblasts, germ cells and Sertoli cells in the testis and neuronal and glial cells in the nervous system where it regulates information processing.^33–35 In addition to the above, hetero-cellular GJIC was recently found to be important for interaction between grafted and host cells in tissue grafts. In one study, the interaction via GJs of grafted neural stem cells and host cells was found to facilitate the integration of the grafted cells into the neural circuitry of the host.³⁶ In other studies the importance of gap junctional coupling was highlighted in cell transplantation after myocardial injury, where engrafted skeletal muscle cells over-expressing Cx43 were found to establish better communication with host cardiomyoctyes and this interaction favored their integration and differentiation in the host tissue.³⁷

Hemichannel-dependent functions

Unapposed hemi-channels, termed connexons have been described to occur independently in multiple cell types and in cultured cells including embryonic stem cells, where they act as paracrine conduits that spread signals to surrounding cells. Cx hemi-channels have been shown to mediate the release of ATP, glutamate, NAD+ and prostaglandin E2, from cells.³⁸ATP acts on purinergic receptors on adjacent cells and activates intercellular Ca²⁺ release possibly to complement the Ca²⁺ release signals occurring more directly via GJs. In addition, hemichannels are involved in the movement of NAD⁺ into and out of cells, reversibly, which may regulate Ca²⁺ concentrations through the CD38trans-membrane glycoprotein.³⁹ Hemichannels have also been shown to exist in heart ventricular myocytes, where they have an osmo-regulatory role, with both negative and positive potential impacts with respect to mycordial infarcts and cardiac physiology.^40–42 Furthermore, hemichannels, similarly to GJs, play a role in spreading cell survival and cell death signals such as during ischemic injury. For example, it was reported that Cx43 hemichannels enhanced and accelerated cell death following sautrosporine drug treatment.⁴³ In contrast, Cx43 hemichannels have been shown to play a role in the transduction of survival signals. Osteocytes and osteoblasts treated with biphosphonates activated Src Kinase and ERK by the hemichannel pore activity and the C-terminal domain of Cx43 led to ERK activation and attenuation of osteoblast cell death.⁴⁴

A novel family of proteins called pannexins has been indentified and described. Pannexins have been shown to have moderate sequence homology to innexins, the GJ forming proteins of invertebrates, and some structural/topological similarities to Cxs.⁴⁵ Using statistical, topological and conserved sequence motif analyses pannexins and innexins have recently been shown to belong to a single super-family of proteins,⁴⁶ implying that pannexins are capable of forming channels similar to their innexin homologues. Pannexins have also been shown to form hemi-channels termed pannexons in a similar fashion to connexons. The ability of pannexins to form hemi-channels that mediate similar roles to connexons suggest that some of the functions previously ascribed to connexons could be mediated by pannexons.⁴⁵ The dye permeability and the sensitivities to inhibitors are similar between connexons and pannexons and make it difficult to distinguish whether these hemi-channel functions are accomplished by pannexons or connexons. Pannexin 1 (Panx 1) channels were recently shown to play also an important role in apoptosis by allowing the release of “find-me” signals by apoptotic cells in their early stages of death to recruit phagocytes.⁴⁷ Connexons have been shown to play roles in the propagation of death signals and further studies will require elucidating how these two types of channels (pannexons and connexons) fulfill non-redundant functions in signal release. Interestingly, a recent study has identified Panx1 as the protein responsible for forming the large conductance channel of previously unknown molecular identity in heart myocytes,⁴⁸ thus, confirming the hypothesis that these proteins are capable of forming gap-junction like structures similar to their innexin homologues and adding to the known repertoire of functions ascribed to Panx channels.

Channel-independent roles of Cxs

It is now uncontested that Cxs have channel-independent roles.^10,13 Cxs have been shown to regulate gene expression partly via Cx-responsive elements (CxRE), where Cxs and GJIC induce differential recruitment of sp1 and sp3 transcription factors to the CxRE through the ERK/PI3K pathway and regulates CxRE-mediated gene expression. Cxs can also regulate cell differentiation in a GJIC-independent manner.^49,50 It has been shown that the C-terminal domain of Cx45.6 stimulates lens cell differentiation and that Cx43, through its C-terminal tail is crucial for preventing premature neuronal differentiation during embryonic brain development.^51,52 Cxs have been also localized to the nucleus and Cx43 has been shown to contain a putative nuclear targeting sequence in its C-terminal domain. In support of this observation, both full length Cx43 or its C-terminus by itself have been localized to the nucleus, where they inhibit cell growth.⁵³ The precise function and mechanism of action of Cxs in the nucleus require further study.

Cxs can also regulate signaling pathways, notably via the engagement of Cx-associated proteins en route to, or at, the membrane as part of the GJ complex. We don't know as yet whether these interactions are important for GJ channel function or whether they are just involved in regulating Cx targeting to the GJ, however, much evidence points to the assembly of Cx-interacting proteins at the GJ interface such that a GJ signaling complex is formed.

Most strategies to study the mechanisms of Cx-interactions have made use of culture systems due to the difficulty of probing interactions in real time in vivo. This is further complicated by the transient nature of these interactions which accommodate the highly dynamic life cycle of Cxs. Using real time imaging and fluorescently labeled Cx-fusion proteins, many studies were able to gage the interaction of Cxs with their trafficking modulators such as the interaction between Cx and microtubules.⁸ Other studies showed that Cx43 governs directional neural crest cell migration,⁵⁴ while siRNA knockdown of Cx43 reduced the cell-surface distribution of N-cadherin, suggesting that Cx43 regulates the status of N-cadherin and possibly cell adhesion in a cadherin-dependent mechanism.⁵⁵ Interestingly, binding of β-catenin to cadherins (E- or N-)^56,57 is essential for down regulation of TCF/LEF-gene expression and inhibition of cell proliferation; thus, Cx43, along with cadherins seems to be important for down regulation of β-catenin-dependent signaling and growth inhibition. This was also shown in a study using HEK293 cells.⁵⁸ Moreover, binding of β-catenin to Cx43 at cell–cell contact areas in cardiac myocytes was also shown to be important for down-regulation of β-catenin-dependent gene transactivation.⁵⁹

We have previously shown in our laboratory that the formation of a GJ complex in heterocellular co-cultures of mouse mammary epithelial cell line SCp2, and myoepithelial-like cell line SCg6, is important for functional differentiation of the SCp2 cells into milk protein expressing cells in the absence of exogenous basement membrane signaling. This functional differentiation was correlated with increased GJIC and increased association between Cxs and β-catenin, α-catenin, and ZO-2. From the latter observation we proposed¹⁹ that the recruitment of the adhesion and tight junction proteins to the GJ complex could serve in part to regulate differentiation by enhancing functional coupling between cells and also by inhibiting β-catenin's nuclear translocation where it interacts with the TCF complex to induce proliferation. Perhaps the significance of these interactions is more evident given the hierarchical regulation of Cx expression in the differentiated versus undifferentiated conditions, especially Cx30.⁶⁰ In fact, mammary epithelial Cx30 seems to be only expressed in differentiated co-cultures, but not in differentiated SCp2 cultures on exogenous ECM. This implies that context regulates Cx30 expression and in turn once expressed, Cx30 contributes to maintaining the differentiated state by assembling into heteromeric connexons⁶¹ recruiting transcription regulators (in this case β-catenin) away from the nucleus. More Cx-interacting partners are being identified, increasing the number of Cx functions.^10,13. Reference is made to such functions in the section addressing GJ-independent roles of Cxs as tumor suppressors. Why Cx have evolved to fulfill multiple roles in the cell remains unclear. Perhaps having the same protein responsible for several functions within the cell allows for a concerted and timely coordination of context-dependent function to promote a unique phenotype.

Context-dependent functions of Cxs: Tissue- and cell type-specific expression and regulation of Cxs and Cx interactions. Tissue and cell-specific expression of Cx isoforms coupled with Cx-pairing specificity and interaction with associated proteins has multiple consequences on cell–cell interactions within a tissue. Channel assembly by compatible connexon docking dictates which cell types are capable of interacting via GJs.

In addition and despite the fact that many tissues express two or more members of the Cx family and that two Cxs can be co-expressed in the same cell, there is little functional overlap between Cx subtypes. This is not surprising given the differences in the functional domains on the Cx protein, which regulate Cx–Cx interactions, Cx interactions with other proteins and trafficking and localization of Cxs within the cell and on the membrane. One example of the stark difference in the structural domains and function of two Cx isoforms is the distinction between Cx26 and Cx43; both isoforms are co-expressed in many tissues and in cultured cells, but to date they have not been shown to assemble into heteromeric connexons.^62,63 The reason for this apparent incompatibility is not fully understood, however many studies suggested that the differences in Cx26 and Cx43 structural domains and post translational trafficking and modifications account for their inability to interact. More precisely, Cx26 has a very short carboxy-tail compared to other Cx isoforms and in contrast to Cx43 is neither phosphorylated nor glycosylated and hence might be trafficked to the membranevia an alternate pathway independent of the common secretory pathway used by modified proteins targeted to the membrane.^64,65 One hypothesis suggests that Cx26 is spatially segregated from Cx43 and cannot oligomerize with it en route to the membrane.^66,67 However separate studies using different cell culture systems found conflicting results about Cx26 and Cx43 trafficking. A study using Cos and HeLa cells showed that targeting of Cx43-GFP to GJs was inhibited in cells treated with Brefeldin A, a drug that disassembles the Golgi network. However, GJs constructed of Cx26-GFP were only minimally affected by Brefeldin A, but were sensitive instead to Nocodazole treatment, suggesting that Cx26 shuttles to the membraneviamicrotubules rather than through the secretory pathway and confirming the proposed hypothesis of alternate routing.⁶⁸ However, another study using HEK and HeLa cells found that pharmacological treatment with Brefeldin A or nocodazole affected both Cx26 and Cx43 similarly suggesting that both isoforms followed similar routes of cellular trafficking and assembly into GJs.⁶² In contrast yet another study using NRK cells revealed that disruption of microtubules with nocodazole inhibited the recruitment of Cx43-GFP into GJs but had limited effect on the transport and clustering of Cx26-YFP into GJs within the photobleached regions of cell–cell contact suggesting that Cx43 shuttles using the MT network, which is plausible given that Cx43 has a tubulin binding site.^8,69 The conflicting findings from these studies point to a more complex regulation of these two proteins in the different cell systems and culture conditions examined, and suggest that there are other factors involved in the trafficking control of these two proteins. Another striking feature of these two Cx isoforms is that their co-expression in the same cell was found to reduce the total intercellular junctional conductance to a little more than 10% of that in cells expressing only a single Cx (either Cx26 alone or Cx43 alone).⁶² The presence of incompatible connexons might be behind the observed reduction of GJIC by preventing functional docking of apposing hemi-channels and causing the accumulation of uncoupled hemi-channels at the intermembrane junction. Alternatively, the lowered conductance might be caused by negative interference between the adjacent connexons which recruit gating regulators away from the nearby channels resulting in uncooperativity between the distinct channels at the GJ plaque region.

Insight from new transgenic and knock-in mouse models support the findings that Cx26 and Cx43 functions do not overlap. This was elegantly shown in a knock-in mouse model where the coding region of Cx43 is replaced by that of Cx26 (Cx43KICx26), such that Cx26 is either solely expressed (homozygous mice) or co-expressed (heterozygous mice) in cells expressing Cx43 endogenously. Cx26 in this case exerted a dominant negative effect on Cx43 (perhaps due to what was mentioned earlier about negative cooperativity) in many tissues and altered proper tissue-function. For example, Cx43KICx26 mice had dysfunctional reproductive organs and slowed ventricular conduction in the heart, abnormalities that were similar to the ones found in Cx43-deficient mice.⁷⁰ In addition, Cx43KICx26 female heterozygous mice had impaired mammary gland growth and differentiation suggesting that Cx26 could not functionally substitute for Cx43 in the mammary gland. Cx26 and Cx43 have distinct expression patterns in the gland, whereby Cx43 is localized mainly to the myoepithelial contractile cells, while Cx26 is exclusively luminal epithelial. In another model where Cx32 substituted for Cx43 expression, the mammary gland developed normally, but milk ejection was impaired suggesting that Cx32 could substitute in part for Cx43 function but that their roles do not overlap, at least, in the mammary gland.²⁰ The mammary phenotypes associated with knock-in of Cx26 and Cx32 in place of Cx43 were not similar to those observed in heterozygous Cx43 (GJa1±) female mice,²⁰ suggesting that the effect is not only mediated by decreased Cx43 levels, but by a dominant-negative effect of the substituting Cxs. However, a mouse model of oculodentodigital dysplasia expressing a mutant Cx43, GJa1 (Jrt/+) has a comparable mammary gland defect to the knock-in models described for Cx32 and Cx26 whereby alveolar development and milk delivery to the pups are impaired.^71,72 Surprisingly, Cx32 and Cx26 have been shown to co-localize in connexons and at GJs, and the stoichiometry of their association in the mammary gland is regulated temporally such that Cx32–Cx32 GJs replace Cx32–Cx26 GJs at lactation since Cx26 is sensitive to taurine, which is synthesized at lactation, and can induce Cx26 channel closure.⁷³ This differential oligomerization is yet another testament to the tight regulation of Cx expression and its profound effect on tissue function. It is obvious then that dynamic changes in Cx expression alter the physiology within a tissue. This was further highlighted in a study where the conditional knockout of Cx26 at different developmental time points in the mouse mammary gland, altered the normal differentiation status of the mammary gland and lead to early apoptosis of epithelial cells.⁷⁴Cx26 loss during late pregnancy did not have significant effects on the differentiation program of mammary cells, perhaps because at this exact time point in the gland, Cx32 or Cx30, a newly identified isoform in the mammary tissue⁶⁰ can functionally compensate for Cx26 loss. Refer to Table 1 for an outline of Cx roles in the normal mammary gland and for available mouse models of Cxs.

Table 1 Connexin expression, distribution and roles in the mammary gland. Cx26, Cx30, Cx32 and Cx43 expression, localization and roles in the mammary gland are outlined in the table and the types of Cx/Cx interactions that occur at the pertinent developmental stages are shown. In addition, reference is made to the relevant mouse models of Cx mis-expression in the mammary gland and the mammary phenotypes associated with these mice. The schematic diagram illustrates the dynamic interplay between the three noted axes; a cell in its right tissue context optimally expresses its connexins and faithfully assembles them into heteromeric (green and pink) or homomeric (blue) connexons at a precise stage of development. Ectopic or disrupted expression of connexins will impair normal mammary gland development and differentiation.

---

For a description of the several knock-in mouse models generated and what they have revealed about the functional overlap and tissue- and cell-type- specific properties of Cx isoforms refer to ref. 20, 70 and 75. The significance of these mouse models is that they provide a physiological environment to study Cx function.

Knock-in strategies have paved the way for a better understanding of tissue and stage specific functions of Cxs but we still need to understand what regulates this cell-specificity of Cx expression and function. Ultimately, these strategies should not only study the interchangeability of different Cxs in a tissue, but also identify the mechanisms behind observed differences. Why can some Cxs replace others but others cannot? More importantly, how is it that interchangeability between Cxs had different consequences in distinct tissues and across development within the same tissue? For example, why can Cx40 replace Cx43 without causing mammary gland functional impairment, but cause ventricular dysfunctions and arrhythmias?²⁰ Is it due to channel-dependent or channel independent functions? And what is the role of Cx-interacting proteins in regulating cell and tissue-specific function, especially in light of recent data pointing to a role of Cx-interacting proteins in regulating tissue and cell selective Cx expression and Cx–Cx pairing? To address these questions it is possible to take the knock-in strategy to a different level, such as temporally controlling the substitution between Cxs, or using Cx fragments in place of a full Cx protein to reveal what parts of the protein are required for different functions. Engineering Cxs with interchanged peptide sequences such as fusing the C-terminus of Cx43 to Cx26 might yield important information about the functional domains in the proteins and possible interacting partners. Studying normal context dependent expression of Cxs and Cx functional interactions is essential for understanding how mis-expression of the GJ proteins can lead to diseases including cancer. In the next section, we will discuss the conflicting roles of Cxs in cancer progression and shed some light on how these different functions are mediated.

Cxs in the cancer context

One of the hallmarks of cancerous cells is their loss of interaction and communication with surrounding and neighboring cells such that they are able to escape the confines of the normal tissue microenvironment. As such, many human and rodent epithelial cell-derived tumors lose GJIC, and it has been shown that re-expression of Cxs in cell lines derived from tumors can lead to decreased growth and partial cell re-differentiation to a less aggressive state.^18,76–78 These findings suggest that GJs and Cxs could act as tumor suppressors. However, details of the mechanisms by which Cxs act as tumor suppressors have generated conflicting reports depending on the cell type studied and the stage of the tumors and on whether GJIC or GJ-independent mechanisms were investigated. For instance, whereas Cx26 and Cx43 are reported to be down-regulated in primary mammary tumors, these same Cxs are shown to be up-regulated during later stages of breast carcinomas and correlate with increased metastatic and invasive potential.⁷⁹ This is not surprising given that at later stages of tumor progression, these same cells need to interact with endothelial and/or lymphatic cells in order to penetrate the circulatory and lymphatic systems and metastasize to target tissues/organs. In fact, enhanced GJ formation has been reported between invasive rat mammary adeno-carcinoma cells and endothelial cells.⁸⁰ Moreover, Cx expression was shown to be upregulated in lymph node metastases of breast cancer even when the primary tumor does not express Cxs.⁸¹ Transfection of Cx43 or Cx32 into cervical and hepatocellular carcinoma cell lines respectively was shown to enhance their invasive properties in culture, as well as in vivo.⁸² These findings lead to the hypothesis that during early stages of tumor formation, GJs are down-regulated so as to allow tumor cell detachment and intravasation, whereas during later stages, metastatic cells increase their GJIC, possibly to facilitate extravasation.^83,84

With the emerging GJ-independent functions of Cxs it is no longer possible to adhere only to the communication hypothesis and channel-independent roles of Cxs and Cx-interacting proteins need to be assessed in the cancer context. It should also be noted that few studies have systematically evaluated the positive or negative roles of Cxs in a physiologically relevant tissue environment where all stages of cancer onset, progression and metastasis can be explored. As such a comprehensive view of Cxs and GJ involvement in cancer progression in a tissue-dependent manner is still lacking.

Tumor suppressive roles of Cxs and GJs

GJIC-dependent mechanisms. Since tumor cells interact with the surrounding normal milieu, it should not be surprising that the exchange of molecules between healthy and cancerous cells is a mechanism that modulates tumor growth, although there is little information regarding the identity of such molecules that need to be exchanged in order to modulate cellular function. One example of such a molecule is glutathione, a tripeptide that has high permeability through gap junction channels, and that has antioxidant properties which might protect cells from reactive oxygen species, thus protecting against DNA damage and detoxifying carcinogens.^85,86 The intercellular exchange of small molecules through GJs has also been explored as a route for enhancing the therapeutic potential through the 'bystander effect',⁸⁷ the process in which cells are subject to the effects of neighbor cells that are connected by GJs. In addition to channel-dependent function, hemi-channels might mediate signals between cells and their surrounding environment, which can have regulatory effects on growth and cell cycle progression and proliferation. However, there is little information about connexon's involvement in cancer mostly due to the lack of systems where hemichannel function can be probed independent of GJ presence.

GJIC-independent mechanisms. Tumor suppressing effects of GJs could be mediated by channel-independent roles. The mechanisms at the core of these findings may be rooted in the expanding GJ proteome. For instance, recent studies have shown that the tumour suppressive properties of Cx43 in keratinocytes might be linked to its interaction with another reported tumour-suppressing molecule, caveolin 1.⁸⁸ Although the role of caveolin 1 in tumour suppression is controversial, it has been identified as a tumour suppressor in several tissues, including the skin.^89,90Cx43 has also been shown to interact with nephroblastoma overexpressed protein (NOV; also known as CCN3), the tumour susceptibility gene 101 protein (TSG101) and Src; all of these proteins have known relationships to cell cycle control and tumorigenesis.^91–93Cx26 was shown to suppress the growth of HeLa cells both in culture and in vivo in the rodent mammary tumor cell line BICR-M1R_k , as well as in the human tumor cell line MDA-MB-453via a GJIC-independent mechanism, by interacting with cell cycle regulating molecules and angiogenesis factors.^94,95 Exogenous expression of either wild-type Cx43 or the C-terminal domain of Cx43 alone, which has no intrinsic channel-forming properties, reduces proliferation rate of neuroblastoma Neuro2a cell line in a GJIC-independent manner.⁹⁶ Similarly, Cx43 expression was shown to revert the transformed phenotype of human glioblastoma tumor cells,^97,98 in a GJIC-independent manner. In addition, differential gene profiling and gene knockdown studies have identified Cxs as potent tumor suppressors, where mice expressing reduced or dominant-negative Cxs are more susceptible to forming tumors when challenged with carcinogens. For example, Cx32 knockout mice are more prone to developing liver cancers when compared to wild-type mice.^99,100 Likewise, knocking down Cx43 in breast cancer cell lines increases their malignancy by increasing proliferation and migratory potential, possibly by down-regulating the anti-angiogenic molecule thrombospondin-1.¹⁰¹ Interestingly, retroviral delivery of Cx43 and Cx26 to MDA-MB-231 cellsin culture did not reduce growth substantially, whereas growth was significantly retarded compared to wild-type MDA-MB-231 cells when implanted in mammary fat pads of nude mice, in the absence of a significant increase in GJIC.¹⁰² In another report, over-expression of Cx26 or Cx43 in three dimensional (3D) cultures of MDA-MB-231 cells reduced growth rate and cell migration and led to a partial re-differentiation of the organoids in a GJIC-independent manner.¹⁸ A recent study in our laboratory (data not published), showed that Cx43 exerts a tumor suppressive effects in culture in a context-dependent manner, by forming a GJ complex that includes and might not be restricted to α-catenin, β-catenin and ZO-2. This complex may be implicated in reducing the growth rate, invasiveness, and hence, malignant phenotype of 2D and 3D cultures of moderately invasive mammary MCF-7 tumor cells and 3D cultures of the highly invasive mammary tumor MDA-MB-231 cells, by sequestering β-catenin away from the nucleus. These studies highlight the importance of using culture models that recreate the in vivo microenvironment and conditions of cells as closely as possible and hence generate results that are more representative of and relevant to the in vivo situation.

With expanding knowledge of the Cx proteome it is possible that Cx-mediated tumor suppression will not be found to correlate consistently with GJIC. Collectively the molecular signatures of distinct tumors might ultimately be required to determine if Cxs could potentially act as supplementary therapeutic targets for cancer treatment. In one recent study, Cx37 dramatically reduced cell cycle progression in a rat insulinoma cell line with a major delay in progression through the G1/S checkpoint¹⁰³ and Cx43 caused an increase in cells found in S phase.¹⁰³ In the case of Cx43, cell cycle progression is tightly linked to specific phosphorylation events.¹⁰⁴ Furthermore, Cx43 has been shown to modulate the expression of several genes involved in the cell cycle, including cyclin A, cyclin D1 and cyclin D2, cyclin-dependent kinases, p21 and p27.^105–107 Mechanistic studies demonstrated that the exogenous expression of Cx43 might exert its inhibitory effect on cell proliferationvia the inhibition of the expression of S phase kinase-associated protein 2 (skp2), the protein that promotes the ubiquitination of cyclin-dependent kinase inhibitor p27kip. A down-regulation of cyclin D1 in osteosarcoma cells was shown upon Cx43 transfection, which correlated with up-regulation of p27kip. The deletion mutation analyses revealed that the C-terminal domain of Cx43 was still sufficient to inhibit skp2 expression, while coupling inhibitors did not affect the inhibitory effect of Cx43 on skp2 expression, indicating that channel independent mechanisms of Cx43 regulation interfere with the cell cycle.¹⁰⁸ As Cxs have a role in cell cycle dynamics, combined approaches with chemotherapeutic agents could have beneficial effects, but only in a context-dependent fashion.

In addition to signaling downstream from Cxs and Cx-associated proteins, GJIC-independent tumor suppression might be regulated by the adhesive properties of GJs, especially given their interaction with cytoskeletal elements and for being found in physical proximity or association with other adhesion proteins. The strong non-covalent links between docked connexons affords adhesive characteristics and therefore, Cxs can also influence cell adhesion and migration. Both of these processes have been shown to have key roles in cancer, as factors that influence adhesion and migration can facilitate or suppress tumour progression and metastasis.^109,110 Many studies have linked Cxs to migration especially via their association with members of the cadherin family of proteins. For example, Cx43 and N-cadherin interact to control cell motility.⁵⁵ The role for Cxs in migration was also noted during the initial characterization of the Cx43-knockout mice that harbor cardiac defects, which result from the disturbed migration of neural crest cells.⁵⁴ Furthermore, Cx43 has been identified in an RNA interference knockdown screen as having a role in migration.¹¹¹ Although the mechanistic link to cell migration is not fully understood, Cx43 has been shown to modulate the cytoskeleton, presumably through interactions with actin through the zonula occludens and by mediating intercellular adhesion, especially via ZO-1's interaction with actin-binding molecules such as drebrin. Collectively these results suggest that Cxs interactions with signaling modulators and adhesion proteins have tumor suppressing effects and that a complete understanding of these interactions in a tissue and stage specific fashion is essential for customizing treatments for tumors with molecularly defined signature.

Cxs in metastasis

In order for tumor cells to metastasize they first need to detach from neighboring cells and penetrate natural barriers such as the basement membrane and endothelial layer. Then neoplastic cells need to invade blood vessels and/or lymphatic vessels in the process known as intravasation, which enables them to access the circulatory system. After reaching the bloodstream, tumor cells enter the microcirculation and, finally exit the bloodstream by extravasation, and form mestastases by embedding themselves in the target tissue.¹¹² Thus, the ability to migrate through the blood stream is a prerequisite of invasion. As mentioned previously, several studies have revealed relatively high levels of Cxs and gap junctional coupling in populations of invasive tumor cells. Kamibayashi et al. demonstrated that Cx26 and Cx43 were expressed on the plasma membranes of skin cells invading the lymph nodes.¹¹³ Similar observations were made for breast cancer¹¹⁴ and prostate cancer.¹¹⁵ Furthermore, GJ formation was observed between melanoma cells and mammary adenocarcinoma cells and the endothelium. In one study using F10 and BL6 sublines of B16 mouse melanoma cells, it was found that both lines are metastatic after intravenous injection, but only BL6 cells are metastatic after subcutaneous injection. Cx26 was found to be upregulated in BL6 cells, and these cells could transfer dye into endothelial cells, but F10 cells could not. However, transfection of wild-type Cx26 rendered F10 cells competent for coupling with endothelial cells and as metastatic as BL6 cells.¹¹⁶ Conversely, transfection with a dominant-negative form of Cx26 rendered BL6 cells deficient in coupling and less metastatic. Another study using a model of Human T-cell lymphotropic virus type I (HTLV-I)-associated adult T-cell leukemia/lymphoma (ATL, which is an aggressive disease characterized by visceral invasion), found that ATL and HTLV-I-associated myelopathy patients exhibit enhanced gap-junction-mediated heterocellular communication between FTLV-1 and endothelial cells. The interaction of HTLV-I-transformed cells with endothelial cells induces the gelatinase activity of matrix metalloproteinase (MMP)-2 and MMP-9 in endothelial cells and down-regulates the tissue inhibitor of MMP. This leads to sub-endothelial basement membrane degradation followed by endothelial cell retraction, allowing neoplastic lymphocyte extravasation. After specific adhesion to endothelia of target organs, tumor cells induce a local and transient angiogenesis-like mechanism through paracrine stimulation and direct cell–cell communication with endothelial cells. This culminates in a breach of the endothelial barrier function, allowing cancer cell invasion. It is evident then that GJIC and possibly GJ mediated adhesion are important steps in the extravasation of tumor cells and their subsequent attachment to the cells in the metastasis site.¹¹⁷

Given the conflicting roles of Cxs in cancer growth and the different mechanisms employed by Cxs to control these processes, from GJIC to adhesion and signaling downstream of Cxs, it is essential to understand how Cxs are regulated at each step in the progression to cancer. In addition to the channel activity, it is plausible to assume that GJ-independent roles of Cxs are important in tumor suppression. On the other hand, GJIC seems to play a dominant role in heterocellular coupling during tumor cell intra- and extravasation.¹⁷ Could a therapeutic strategy then be designed to enable one aspect of Cx function but not the other? For example, is it possible to use channel-incompetent Cxs as tumor suppressors? More importantly, given Cxs interactions with other molecules, could the use of specific domains in the Cx protein as therapeutic molecules promote one function versus the other? This will only be possible once we understand the tight regulation of Cx expression and once we can determine a Cx-proteome signature for different stages of tumor growth.

Perspectives and future directions

Although noted advances have significantly enhanced our understanding of gap junction function, two major obstacles still hinder experimentation on GJs: the scarcity of model systems that can faithfully recapitulate the whole tissue context for Cx expression and the lack of functional assays that can capture GJIC in real time without altering function. To date, we have been limited in our thinking about Cxs and some of our limitations are due to lack of appropriate technologies as mentioned above while some are due to “dogma” and advancement in this field requires overcoming both of these bounds. A new era of biological sciences is emerging marked by interdisciplinary collaborations which have enabled the conception of complex biological model systems that permit high throughput and automated study of tissue function in real time.¹¹⁸ More so, there is a definite shift in dogmatic thinking from studying isolated mechanistic events in 2D reductionist settings to a new “systems biology” approach of experimentation, which focuses on understanding the integration of the dynamic components that lead to tissue and organ function beyond the sum of the individual parts.^119,120 Systems biology approaches take into account the increasingly acknowledged concept that successful modeling of tissues recognizes the importance of assembly and organizing principles of cells, and the need to customize model systems that permit the timely delivery of the component parts (cells, hormones, growth factors and matrices) so that tissues and organs are reconstructed de novo.^121–123 This, in a sense, takes us from 3D cultures into 4D cultures with the dictated timely and spatial delivery of multiple components of the microenvironment being the 4th dimension. How can we make use of these tools to put the pieces of the Cx puzzle together? Of the numerous approaches that are being pursued toward that goal, microfluidic devices and bioprinting offer adequate solutions for addressing questions of interest in the Cx and GJ field.^124,125 Recently, a microfluidic assay was designed and described by Chen & Lee ¹²⁶ that can probe for GJIC in real time and without injuring cells. The basic premise of the assay makes use of microfluidics for dye loading, which is usually accomplished by microinjection, scrape loading, or electroporation, but in this case dye loading is done by hydrodynamic focusing of the dye onto one row of cells. This technique circumvents the need for cellular injury. Once loaded into the “row” of cells, the dye, which in this case is calcein AM, is cleaved by esterases and unable to leak out of cells into the medium; thus, it spreads to the neighboring cellsvia GJs and the diffusion of the dye is used as a measure of GJIC. The advantage of this method lies not only in the dye loading technique, but also in the fact that dye transfer dynamics from loaded cells into the neighboring unexposed cells can be followed viatime-lapse microscopy in a quantifiable manner, enabling in situ monitoring of GJIC.¹²⁶ This assay has thus far been customized for cells in monolayers, but it can be rendered more successful if it is modified to accommodate cells in a 3D, or a 4D bioprint, platform such that a more native environment than culture plastic is provided.

Bioprinting, is in its early stages of use,¹²⁷ but has the potential to surpass the traditional solid-scaffold-based approaches of tissue engineering in addressing questions of tissue dynamics. The basic premise of this technology is that it allows directed tissue assembly of cells based on a prototyped model by precise automated robotic placement of cells and tissue pieces according to computer-aided design. Precise robotic placement of two tissue slices or cells in close contact creates permissive, but not instructive, conditions for tissue genesis, thus, the tissue assembles based on the intrinsic capacity of cellular fusion.¹²¹ A fundamental feature of this process is its capability of simultaneous delivery of scaffolding materials, living cells, nutrients, therapeutic drugs, growth factors and or other important chemical components at the right time, right position, right amount and within the right microenvironment to form living cells/extracellular matrix (or scaffold) for in vitro or in vivogrowth.¹²⁸ Bioprinting has been successfully used to study the effect of spatial patterning on the differentiation of osteoblast progenitor cells both in vitro and in vivo using bone morphogenetic protein-2 (BMP-2) as a growth factor to stimulate osteoblast cell differentiation. The study showed that 3D bioprinting of BMP-2 or its inhibitor in the “print” construct directed progenitor cell differentiation and even allowed bone formation when the constructs were implanted into mouse skull region that had a defect. Despite its novelty, the technique is rapidly evolving and we can envision many uses for bioprinting to study Cx-regulation and function within tissue.¹²⁸ One application for example, would be printing cells transfected with different Cxs or Cxs with specific mutations and comparing their assembly and integration into a particular tissue. This permits the study of different Cx mutations on differentiation of cell-types within a tissue. Another way to make use of bioprinting would be to print specific Cxs on a matrix and then add primary cells on top of them and study the effect of Cx–Cx interactions on the function of the printed cells. On the other hand, printing two different cell types containing fluorescently labeled Cxs might be important for deciphering the complexity of Cx–Cx interactions and other functions of heterocellular GJs. What these assays should provide us with is an experimental system where Cx and GJ expression can be manipulated in a physiologically relevant milieu such that we can probe for and understand the requirements for tissue and cell specific expression of GJ proteins and their functional interactions. The more pertinent question will be one that attempts to achieve a better understanding of how deregulation of Cx function and functional interactions contribute to cancer progression? How can we integrate all our knowledge and use these available techniques to generate a functional map that can outline potential therapeutic strategies depending on the tumor context and on the status of Cx expression and regulation?

As we reconstruct the tissue context and outline the intricate signaling that leads to tissue specific function using bioprinting approaches for example, we will better understand the important role of intercellular connections and communication in mediating context-dependent signaling. Cxs with their interacting partners and in GJ complexes present yet another level of dynamic and reciprocal regulation of tissue function. However, our understanding of Cx and GJ channel dependent and independent functions will definitely change as we continue to integrate the different aspects that make up the tissue context, and as we increase our knowledge of the organizing principles governing cell, tissue, organ and ultimately organism function.

Acknowledgements

The authors are grateful for Dr Hidetoshi Mori and Dr Mike Osta for critical reading of the manuscript and Dr Cyrus Ghajar for helpful advice. Ms Elia El-Habre is acknowledged for her help in manuscript and figure preparation. This work was supported by the University Research Board and Lebanese National Council for Scientific Research (RST and MES), and Medical Practice Plan (MES). Rana Mroue is supported by a predoctoral fellowship from the DOD Breast Cancer Research Program.

References

1. C. M. Nelson and M. J. Bissell, Of extracellular matrix, scaffolds, and signaling: tissue architecture regulates development, homeostasis, and cancer, Annu. Rev. Cell Dev. Biol., 2006, 22, 287–309.
2. N. M. Kumar and N. B. Gilula, The gap junction communication channel, Cell, 1996, 84(3), 381–8.
3. C. J. Wei, X. Xu and C. W. Lo, Connexins and cell signaling in development and disease, Annu. Rev. Cell Dev. Biol., 2004, 20, 811–38.
4. D. A. Goodenough, J. A. Goliger and D. L. Paul, Connexins, connexons, and intercellular communication, Annu. Rev. Biochem., 1996, 65, 475–502.
5. K. Willecke, et al., Structural and functional diversity of connexin genes in the mouse and human genome, Biol. Chem., 2002, 383(5), 725–37.
6. D. W. Laird, Life cycle of connexins in health and disease, Biochem. J., 2006, 394(Pt 3), 527–43.
7. C. G. Bevans, et al., Isoform composition of connexin channels determines selectivity among second messengers and uncharged molecules, J. Biol. Chem., 1998, 273(5), 2808–16.
8. B. N. Giepmans, et al., Gap junction protein connexin-43 interacts directly with microtubules, Curr. Biol., 2001, 11(17), 1364–8.
9. B. N. Giepmans, Role of connexin 43-interacting proteins at gap junctions, Adv. Cardiol., 2006, 42, 41–56.
10. D. W. Laird, The gap junction proteome and its relationship to disease, Trends Cell Biol., 2010, 20(2), 92–101.
11. D. Singh, et al., Connexin 43 interacts with zona occludens-1 and -2 proteins in a cell cycle stage-specific manner, J. Biol. Chem., 2005, 280(34), 30416–21.
12. D. Singh and P. D. Lampe, Identification of connexin-43 interacting proteins, Cell Commun. Adhes., 2003, 10(4–6), 215–20.
13. H. A. Dbouk, et al., Connexins: a myriad of functions extending beyond assembly of gap junction channels, Cell Commun. Signaling, 2009, 7, 4.
14. K. Willecke, et al., Biological functions of connexin genes revealed by human genetic defects, dominant negative approaches and targeted deletions in the mouse, Novartis Found Symp., 1999, 219, 76–88 , discussion 88-96.
15. R. Dobrowolski and K. Willecke, Connexin-caused genetic diseases and corresponding mouse models, Antioxid. Redox Signaling, 2009, 11(2), 283–95.
16. W. R. Loewenstein and Y. Kano, Intercellular communication and the control of tissue growth: lack of communication between cancer cells, Nature, 1966, 209, 1248–49.
17. C. C. Naus and D. W. Laird, Implications and challenges of connexin connections to cancer, Nat. Rev. Cancer, 2010, 10(6), 435–41.
18. E. McLachlan, et al., Connexins act as tumor suppressors in three-dimensional mammary cell organoids by regulating differentiation and angiogenesis, Cancer Res., 2006, 66(20), 9886–94.
19. R. S. Talhouk, et al., Heterocellular interaction enhances recruitment of alpha and beta-catenins and ZO-2 into functional gap-junction complexes and induces gap junction-dependant differentiation of mammary epithelial cells, Exp. Cell Res., 2008, 314(18), 3275–91.
20. A. Plum, et al., Unique and shared functions of different connexins in mice, Curr. Biol., 2000, 10(18), 1083–91.
21. M. J. Bissell, H. G. Hall and G. Parry, How does the extracellular matrix direct gene expression?, J. Theor. Biol., 1982, 99(1), 31–68.
22. W. H. Evans and P. E. Martin, Gap junctions: structure and function (Review), Mol. Membr. Biol., 2002, 19(2), 121–36.
23. Y. Haraguchi, et al., Electrical coupling of cardiomyocyte sheets occurs rapidly via functional gap junction formation, Biomaterials, 2006, 27(27), 4765–74.
24. J. E. Saffitz, J. G. Laing and K. A. Yamada, Connexin expression and turnover: implications for cardiac excitability, Circ. Res., 2000, 86(7), 723–8.
25. M. J. Sanderson, I. Chow and E. R. Dirksen, Intercellular communication between ciliated cells in culture, Am. J Physiol., 1988, 254(1 Pt 1), C63–74.
26. E. R. Dirksen and M. J. Sanderson, Regulation of ciliary activity in the mammalian respiratory tract, Biorheology, 1990, 27(3–4), 533–45.
27. D. Tong, et al., A dominant loss-of-function GJA1 (Cx43) mutant impairs parturition in the mouse, Biol. Reprod., 2009, 80(6), 1099–106.
28. B. Doring, et al., Ablation of connexin43 in uterine smooth muscle cells of the mouse causes delayed parturition, J. Cell Sci., 2006, 119(9), 1715–22.
29. J. C. Herve, et al., Connexins, innexins and pannexins: bridging the communication gap, Biochim. Biophys. Acta, Biomembr., 2005, 1719(1–2), 3–5.
30. Y. Nakamura, et al., Augmentation of differentiation and gap junction function by kaempferol in partially differentiated colon cancer cells, Carcinogenesis, 2005, 26(3), 665–71.
31. M. E. El-Sabban, et al., ECM-induced gap junctional communication enhances mammary epithelial cell differentiation, J. Cell Sci., 2003, 116(17), 3531–41.
32. V. Abraham, et al., Heterocellular gap junctional communication between alveolar epithelial cells, Am. J. Physiol. Lung. Cell Mol. Physiol., 2001, 280(6), L1085–93.
33. V. Alvarez-Maubecin, et al., Functional coupling between neurons and glia, J. Neurosci., 2000, 20(11), 4091–8.
34. N. P. Lee, W. S. Yeung and J. M. Luk, Junction interaction in the seminiferous epithelium: regulatory roles of connexin-based gap junction, Front. Biosci., 2007, 12, 1552–62.
35. P. Camelliti, C. R. Green and P. Kohl, Structural and functional coupling of cardiac myocytes and fibroblasts, Adv. Cardiol., 2006, 42, 132–49.
36. J. Jaderstad, et al., Communication via gap junctions underlies early functional and beneficial interactions between grafted neural stem cells and the host, Proc. Natl. Acad. Sci. U. S. A., 2010, 107(11), 5184–9.
37. K. Suzuki, et al., Overexpression of connexin 43 in skeletal myoblasts: Relevance to cell transplantation to the heart, J. Thorac. Cardiovasc. Surg., 2001, 122(4), 759–66.
38. W. H. Evans, E. De Vuyst and L. Leybaert, The gap junction cellular internet: connexin hemichannels enter the signalling limelight, Biochem. J., 2006, 397(1), 1–14.
39. S. Bruzzone, et al., A self-restricted CD38-connexin 43 cross-talk affects NAD+ and cyclic ADP-ribose metabolism and regulates intracellular calcium in 3T3 fibroblasts, J. Biol. Chem., 2001, 276(51), 48300–8.
40. K. Shintani-Ishida, K. Uemura and K. Yoshida, Hemichannels in cardiomyocytes open transiently during ischemia and contribute to reperfusion injury following brief ischemia, Am. J. Physiol.: Heart Circ. Physiol., 2007, 293(3), H1714–20.
41. S. John, D. Cesario and J. N. Weiss, Gap junctional hemichannels in the heart, Acta Physiol. Scand., 2003, 179(1), 23–31.
42. C. Rucker-Martin, et al., Chronic hemodynamic overload of the atria is an important factor for gap junction remodeling in human and rat hearts, Cardiovasc. Res., 2006, 72(1), 69–79.
43. K. C. Hur, J. E. Shim and R. G. Johnson, A potential role for cx43-hemichannels in staurosporin-induced apoptosis, Cell Commun. Adhes., 2003, 10(4–6), 271–7.
44. L. I. Plotkin and T. Bellido, Bisphosphonate-induced, hemichannel-mediated, anti-apoptosis through the Src/ERK pathway: a gap junction-independent action of connexin43, Cell Commun. Adhes., 2001, 8(4–6), 377–82.
45. G. Dahl and S. Locovei, Pannexin: to gap or not to gap, is that a question?, IUBMB Life, 2006, 58(7), 409–19.
46. M. R. Yen and M. H. Saier, Jr., Gap junctional proteins of animals: the innexin/pannexin superfamily, Prog. Biophys. Mol. Biol., 2007, 94(1–2), 5–14.
47. F. B. Chekeni, et al., Pannexin 1 channels mediate ‘find-me’ signal release and membrane permeability during apoptosis, Nature, 2010, 467(7317), 863–7.
48. M. C. Kienitz, et al., Pannexin 1 constitutes the large conductance cation channel of cardiac myocytes, J. Biol. Chem., 2011, 286(1), 290–8.
49. J. P. Stains and R. Civitelli, Gap junctions regulate extracellular signal-regulated kinase signaling to affect gene transcription, Mol. Biol. Cell, 2005, 16(1), 64–72.
50. J. P. Stains, et al., Gap junctional communication modulates gene transcription by altering the recruitment of Sp1 and Sp3 to connexin-response elements in osteoblast promoters, J. Biol. Chem., 2003, 278(27), 24377–87.
51. E. A. Banks, et al., Promotion of lens epithelial-fiber differentiation by the C-terminus of connexin 45.6 a role independent of gap junction communication, J. Cell Sci., 2007, 120(20), 3602–12.
52. C. Cina, et al., Involvement of the cytoplasmic C-terminal domain of connexin 43 in neuronal migration, J. Neurosci., 2009, 29(7), 2009–21.
53. X. Dang, B. W. Doble and E. Kardami, The carboxy-tail of connexin-43 localizes to the nucleus and inhibits cell growth, Mol. Cell. Biochem., 2003, 242(1/2), 35–8.
54. X. Xu, et al., Connexin 43-mediated modulation of polarized cell movement and the directional migration of cardiac neural crest cells, Development, 2006, 133(18), 3629–39.
55. C. J. Wei, et al., Connexin43 associated with an N-cadherin-containing multiprotein complex is required for gap junction formation in NIH3T3 cells, J. Biol. Chem., 2005, 280(20), 19925–36.
56. C. J. Gottardi and B. M. Gumbiner, Adhesion signaling: how beta-catenin interacts with its partners, Curr. Biol., 2001, 11(19), R792–4.
57. C. J. Gottardi, E. Wong and B. M. Gumbiner, E-cadherin suppresses cellular transformation by inhibiting beta-catenin signaling in an adhesion-independent manner, J. Cell Biol., 2001, 153(5), 1049–60.
58. J. Kamei, T. Toyofuku and M. Hori, Negative regulation of p21 by beta-catenin/TCF signaling: a novel mechanism by which cell adhesion molecules regulate cell proliferation, Biochem. Biophys. Res. Commun., 2003, 312(2), 380–7.
59. Z. Ai, et al., Wnt-1 regulation of connexin43 in cardiac myocytes, J. Clin. Invest., 2000, 105(2), 161–71.
60. R. S. Talhouk, et al., Developmental expression patterns and regulation of connexins in the mouse mammary gland: expression of connexin30 in lactogenesis, Cell Tissue Res., 2005, 319(1), 49–59.
61. D. Locke, et al., Nature of Cx30-containing channels in the adult mouse mammary gland, Cell Tissue Res., 2007, 328(1), 97–107.
62. J. Gemel, et al., Connexin43 and connexin26 form gap junctions, but not heteromeric channels in co-expressing cells, J. Cell Sci., 2004, 117(Pt 12), 2469–80.
63. E. C. Beyer, et al., Heteromeric mixing of connexins: compatibility of partners and functional consequences, Cell Commun. Adhes., 2001, 8(4–6), 199–204.
64. S. Maeda, et al., Structure of the connexin 26 gap junction channel at 3.5 Å resolution, Nature, 2009, 458(7238), 597–602.
65. M. Delmar, et al., Structural bases for the chemical regulation of Connexin43 channels, Cardiovasc. Res., 2004, 62(2), 268–75.
66. C. H. George, J. M. Kendall and W. H. Evans, Intracellular trafficking pathways in the assembly of connexins into gap junctions, J. Biol. Chem., 1999, 274(13), 8678–85.
67. W. H. Evans, et al., Trafficking pathways leading to the formation of gap junctions, Novartis Found Symp., 1999, 219, 44–54 ; discussion 54-9.
68. P. E. Martin, et al., Multiple pathways in the trafficking and assembly of connexin 26, 32 and 43 into gap junction intercellular communication channels, J. Cell Sci., 2001, 114(Pt 21), 3845–55.
69. T. Thomas, K. Jordan and D. W. Laird, Role of cytoskeletal elements in the recruitment of Cx43-GFP and Cx26-YFP into gap junctions, Cell Commun. Adhes., 2001, 8(4–6), 231–6.
70. E. Winterhager, et al., Replacement of connexin43 by connexin26 in transgenic mice leads to dysfunctional reproductive organs and slowed ventricular conduction in the heart, BMC Dev. Biol., 2007, 7, 26.
71. I. Plante and D. W. Laird, Decreased levels of connexin43 result in impaired development of the mammary gland in a mouse model of oculodentodigital dysplasia, Dev. Biol., 2008, 318(2), 312–22.
72. I. Plante, et al., Milk secretion and ejection are impaired in the mammary gland of mice harboring a Cx43 mutant while expression and localization of tight and adherens junction proteins remain unchanged, Biol. Reprod., 2010, 82(5), 837–47.
73. D. Locke, et al., Altered permeability and modulatory character of connexin channels during mammary gland development, Exp. Cell Res., 2004, 298(2), 643–60.
74. C. Bry, et al., Loss of connexin 26 in mammary epithelium during early but not during late pregnancy results in unscheduled apoptosis and impaired development, Dev. Biol., 2004, 267(2), 418–29.
75. Q. Zheng-Fischhofer, et al., Connexin31 cannot functionally replace connexin43 during cardiac morphogenesis in mice, J. Cell Sci., 2006, 119(4), 693–701.
76. K. K. Hirschi, et al., Gap junction genes Cx26 and Cx43 individually suppress the cancer phenotype of human mammary carcinoma cells and restore differentiation potential, Cell Growth Differ., 1996, 7(7), 861–70.
77. P. Hellmann, et al., Transfection with different connexin genes alters growth and differentiation of human choriocarcinoma cells, Exp. Cell Res., 1999, 246(2), 480–90.
78. D. Zhu, et al., Transfection of C6 glioma cells with connexin 43 cDNA: analysis of expression, intercellular coupling, and cell proliferation, Proc. Natl. Acad. Sci. U. S. A., 1991, 88(5), 1883–7.
79. S. Jamieson, et al., Expression of gap junction proteins connexin 26 and connexin 43 in normal human breast and in breast tumours, J. Pathol., 1998, 184(1), 37–43.
80. M. E. el-Sabban and B. U. Pauli, Cytoplasmic dye transfer between metastatic tumor cells and vascular endothelium, J. Cell Biol., 1991, 115(5), 1375–82.
81. L. Kanczuga-Koda, et al., Increased expression of gap junction protein-connexin 32 in lymph node metastases of human ductal breast cancer, Folia Histochem. Cytobiol., 2007, 45(Suppl 1), S175–80.
82. S. H. Graeber and D. F. Hulser, Connexin transfection induces invasive properties in HeLa cells, Exp. Cell Res., 1998, 243(1), 142–9.
83. M. E. el-Sabban and B. U. Pauli, Adhesion-mediated gap junctional communication between lung-metastatatic cancer cells and endothelium, Invasion Metastasis, 1994, 14(1–6), 164–76.
84. E. McLachlan, Q. Shao and D. W. Laird, Connexins and gap junctions in mammary gland development and breast cancer progression, J. Membr. Biol., 2007, 218(1–3), 107–21.
85. G. S. Goldberg, P. D. Lampe and B. J. Nicholson, Selective transfer of endogenous metabolites through gap junctions composed of different connexins, Nat. Cell Biol., 1999, 1(7), 457–9.
86. G. K. Balendiran, R. Dabur and D. Fraser, The role of glutathione in cancer, Cell Biochem. Funct., 2004, 22(6), 343–52.
87. A. F. Andrade-Rozental, et al., Gap junctions: the “kiss of death” and the “kiss of life”, Brain Res. Rev., 2000, 32(1), 308–15.
88. S. Langlois, et al., The tumor-suppressive function of Connexin43 in keratinocytes is mediated in part via interaction with caveolin-1, Cancer Res., 2010, 70(10), 4222–32.
89. F. Capozza, et al., Absence of caveolin-1 sensitizes mouse skin to carcinogen-induced epidermal hyperplasia and tumor formation, Am. J. Pathol., 2003, 162(6), 2029–39.
90. C. Trimmer, et al., CAV1 inhibits metastatic potential in melanomas through suppression of the Integrin/Src/FAK signaling pathway, Cancer Res., 2010, 70(19), 7489–99.
91. A. Gellhaus, et al., Connexin43 interacts with NOV: a possible mechanism for negative regulation of cell growth in choriocarcinoma cells, J. Biol. Chem., 2004, 279(35), 36931–42.
92. E. Leithe, et al., Ubiquitylation of the gap junction protein connexin-43 signals its trafficking from early endosomes to lysosomes in a process mediated by Hrs and Tsg101, J. Cell Sci., 2009, 122(21), 3883–93.
93. E. Fujimoto, et al., Inhibition of Src activity enhances the tumor-suppressive effect of the connexin 32 gene in Caki-1 renal cancer cells, Oncol. Rep., 2006, 15(5), 1359–65.
94. Y. Omori and H. Yamasaki, Gap junction proteins connexin32 and connexin43 partially acquire growth-suppressive function in HeLa cells by deletion of their C-terminal tails, Carcinogenesis, 1999, 20(10), 1913–8.
95. J. Kalra, et al., Cx26 inhibits breast MDA-MB-435 cell tumorigenic properties by a gap junctional intercellular communication-independent mechanism, Carcinogenesis, 2006, 27(12), 2528–37.
96. C. Moorby and M. Patel, Dual functions for connexins: Cx43 regulates growth independently of gap junction formation, Exp. Cell Res., 2001, 271(2), 238–48.
97. R. P. Huang, et al., Reversion of the neoplastic phenotype of human glioblastoma cells by connexin 43 (cx43), Cancer Res., 1998, 58(22), 5089–96.
98. R. Huang, et al., Connexin 43 suppresses human glioblastoma cell growth by down-regulation of monocyte chemotactic protein 1, as discovered using protein array technology, Cancer Res., 2002, 62(10), 2806–12.
99. A. Temme, et al., High incidence of spontaneous and chemically induced liver tumors in mice deficient for connexin32, Curr. Biol., 1997, 7(9), 713–6.
100. E. G. Luebeck, et al., Modulation of liver tumorigenesis in Connexin32-deficient mouse, Mutat. Res., Fundam. Mol. Mech. Mutagen., 2005, 570(1), 33–47.
101. Q. Shao, et al., Down-regulation of Cx43 by retroviral delivery of small interfering RNA promotes an aggressive breast cancer cell phenotype, Cancer Res., 2005, 65(7), 2705–11.
102. H. Qin, et al., Retroviral delivery of connexin genes to human breast tumor cells inhibits in vivo tumor growth by a mechanism that is independent of significant gap junctional intercellular communication, J. Biol. Chem., 2002, 277(32), 29132–8.
103. J. M. Burt, et al., Connexin 37 profoundly slows cell cycle progression in rat insulinoma cells, Am. J. Physiol.: Cell Physiol., 2008, 295(5), C1103–12.
104. J. L. Solan, et al., Connexin43 phosphorylation at S368 is acute during S and G2/M and in response to protein kinase C activation, J. Cell Sci., 2003, 116(11), 2203–11.
105. S. C. Chen, et al., Connexin43 reverses the phenotype of transformed cells and alters their expression of cyclin/cyclin-dependent kinases, Cell Growth Differ., 1995, 6(6), 681–90.
106. R. Sanchez-Alvarez, et al., Tolbutamide reduces glioma cell proliferation by increasing connexin43, which promotes the up-regulation of p21 and p27 and subsequent changes in retinoblastoma phosphorylation, Glia, 2006, 54(2), 125–34.
107. Y. W. Zhang, et al., Connexin43 suppresses proliferation of osteosarcoma U2OS cells through post-transcriptional regulation of p27, Oncogene, 2001, 20(31), 4138–49.
108. Y. W. Zhang, K. Nakayama and I. Morita, A novel route for connexin 43 to inhibit cell proliferation: negative regulation of S-phase kinase-associated protein (Skp 2), Cancer Res., 2003, 63(7), 1623–30.
109. J. H. Lin, et al., Connexin 43 enhances the adhesivity and mediates the invasion of malignant glioma cells, J. Neurosci., 2002, 22(11), 4302–11.
110. M. K. Elzarrad, et al., Connexin-43 upregulation in micrometastases and tumor vasculature and its role in tumor cell attachment to pulmonary endothelium, BMC Med., 2008, 6, 20.
111. K. J. Simpson, et al., Identification of genes that regulate epithelial cell migration using an siRNA screening approach, Nat. Cell Biol., 2008, 10(9), 1027–38.
112. G. P. Gupta and J. Massague, Cancer metastasis: building a framework, Cell, 2006, 127(4), 679–95.
113. Y. Kamibayashi, et al., Aberrant expression of gap junction proteins (connexins) is associated with tumor progression during multistage mouse skin carcinogenesis in vivo, Carcinogenesis, 1995, 16(6), 1287–97.
114. L. Kanczuga-Koda, et al., Increased expression of connexins 26 and 43 in lymph node metastases of breast cancer, J. Clin. Pathol., 2006, 59(4), 429–33.
115. K. Miekus, et al., Contact stimulation of prostate cancer cell migration: the role of gap junctional coupling and migration stimulated by heterotypic cell-to-cell contacts in determination of the metastatic phenotype of Dunning rat prostate cancer cells, Biol. Cell, 2005, 97(12), 893–903.
116. A. Ito, et al., A role for heterologous gap junctions between melanoma and endothelial cells in metastasis, J. Clin. Invest., 2000, 105(9), 1189–97.
117. A. Bazarbachi, et al., Human T-cell lymphotropic virus type I-infected cells extravasate through the endothelial barrier by a local angiogenesis-like mechanism, Cancer Res., 2004, 64(6), 2039–46.
118. D. Wlodkowic and J. M. Cooper, Tumors on chips: oncology meets microfluidics, Curr. Opin. Chem. Biol., 2010, 14(5), 556–67.
119. P. Kohl, et al., Systems biology: an approach, Clin. Pharmacol. Ther., 2010, 88(1), 25–33.
120. A. Aderem, Systems biology: its practice and challenges, Cell, 2005, 121(4), 511–3.
121. K. Jakab, et al., Tissue engineering by self-assembly and bio-printing of living cells, Biofabrication, 2010, 2(2), 022001.
122. R. P. Visconti, et al., Towards organ printing: engineering an intra-organ branched vascular tree, Expert Opin. Biol. Ther., 2010, 10(3), 409–20.
123. V. Mironov, et al., Organ printing: tissue spheroids as building blocks, Biomaterials, 2009, 30(12), 2164–74.
124. C. Zhang, et al., Towards a human-on-chip: culturing multiple cell types on a chip with compartmentalized microenvironments, Lab Chip, 2009, 9(22), 3185–92.
125. X. Feng, et al., Microfluidic chip: next-generation platform for systems biology, Anal. Chim. Acta, 2009, 650(1), 83–97.
126. S. Chen and L. P. Lee, Non-invasive microfluidic gap junction assay, Integr. Biol., 2010, 2(2–3), 130–8.
127. V. Mironov, N. Reis and B. Derby, Review: bioprinting: a beginning, Tissue Eng., 2006, 12(4), 631–4.
128. G. M. Cooper, et al., Inkjet-based biopatterning of bone morphogenetic protein-2 to spatially control calvarial bone formation, Tissue Eng. A, 2010, 16(5), 1749–59.

---

Footnote

† Published as part of an Integrative Biology themed issue in honour of Mina J. Bissell: Guest Editor Mary Helen Barcellos-Hoff.

---