Oncogene (2019) 38:4429–4451 https://doi.org/10.1038/s41388-019-0741-6 REVIEW ARTICLE Connexins in cancer: bridging the gap to the clinic Trond Aasen 1 Edward Leithe2 Sheila V. Graham3 Petra Kameritsch4 María D. Mayán5 Marc Mesnil6 ● ● ● ● ● ● Kristin Pogoda4 Arantxa Tabernero 7 ● Received: 22 November 2018 / Revised: 26 January 2019 / Accepted: 26 January 2019 / Published online: 27 February 2019 © The Author(s) 2019. This article is published with open access Abstract Gap junctions comprise arrays of intercellular channels formed by connexin proteins and provide for the direct communication between adjacent cells. This type of intercellular communication permits the coordination of cellular activities and plays key roles in the control of cell growth and differentiation and in the maintenance of tissue homoeostasis. After more than 50 years, deciphering the links among connexins, gap junctions and cancer, researchers are now beginning to translate this knowledge to the clinic. The emergence of new strategies for connexin targeting, combined with an improved understanding of the molecular bases underlying the dysregulation of connexins during cancer development, offers 1234567890();,: 1234567890();,: novel opportunities for clinical applications. However, different connexin isoforms have diverse channel-dependent and -independent functions that are tissue and stage specific. This can elicit both pro- and anti-tumorigenic effects that engender significant challenges in the path towards personalised medicine. Here, we review the current understanding of the role of connexins and gap junctions in cancer, with particular focus on the recent progress made in determining their prognostic and therapeutic potential. Introduction tissue homoeostasis and regulation of cell growth and dif- ferentiation [2]. In addition, connexins form functional Connexins are integral membrane proteins that form chan- channels at the non-junctional areas of the plasma mem- nels between adjacent cells and thereby permit the bidir- brane (Fig. 1) [3]. These so-called hemichannels provide a ectional cytosolic exchange of ions, metabolites and communication pathway between the intracellular and secondary messengers ( < ~ 1200 Da) (Fig. 1a) [1]. These extracellular milieus that is important in autocrine and channels assemble into distinct plasma membrane domains paracrine signalling [3]. Connexins also possess significant termed gap junctions (Fig. 1b) [1]. Gap junction inter- channel-independent roles, including the one as signalling cellular communication (GJIC) plays essential roles in hubs, which may occur at the plasma membrane, in the cytoplasm or even in the nucleus [2]. The human connexin protein family contains 21 mem- bers, of which the most widely studied is connexin 43 These authors contributed equally: Trond Aasen, Edward Leithe (Cx43) [2]. A large body of experimental evidence suggests * Trond Aasen Veterinary and Life Sciences, University of Glasgow, trond.aasen@vhir.org Glasgow, UK * Arantxa Tabernero 4 Walter Brendel Centre of Experimental Medicine, Ludwig- ataber@usal.es Maximilians-Universität München and Munich University Hospital, München, Germany 1 5 Translational Molecular Pathology, Vall d’Hebron Institute of CellCOM Research Group, Instituto de Investigación Biomédica Research (VHIR), Autonomous University of Barcelona, de A Coruña (INIBIC), Servizo Galego de Saúde (SERGAS), CIBERONC, Barcelona, Spain University of A Coruña, A Coruña, Spain 2 6 Department of Molecular Oncology, Institute for Cancer Research, STIM Laboratory, Faculté des Sciences Fondamentales et Oslo University Hospital and K.G. Jebsen Colorectal Cancer Appliquées, Université de Poitiers, Poitiers, France Research Centre, Oslo University Hospital, Oslo, Norway 7 Departamento de Bioquímica y Biología Molecular, Instituto de 3 MRC-University of Glasgow Centre for Virus Research, Institute Neurociencias de Castilla y León (INCYL), Universidad de of Infection, Immunity and Inflammation, College of Medical, Salamanca, Salamanca, Spain 4430 T. Aasen et al. A B Connexin Extracellular loops Membrane 5 Connexons/hemichannels Gap junction Connexons/hemichannels 4 H3N+ 12 COO– ? Cytoplasmic loop 6 10 ? Microvesicle Connexon Trans-Golgi network 8 Connexosome 3 Golgi apparatus 7 Early endosome 9 11 Gap junction 2 intercellular channel Phagophore Endoplasmic Proteasome Exosomes reticulum Multivesicular 1 endosome Connexin Connexon Autophagosome Lysosome Fig. 1 Connexins, connexons and gap junctions. a Connexins are junction endocytosis results in the formation of a connexosome (also tetraspanning integral membrane proteins with cytosolic C and N known as annular gap junction). (7) The connexosome may be termini. Six connexins oligomerize to form a connexon. At the plasma degraded by autophagy. (8) Alternatively, the connexosome may membrane, the connexon can dock head-to-head with a connexon in an change its morphology to that of a connexin-enriched multivesicular adjacent cell to form a gap junction intercellular channel. b Bio- endosome in a process that is associated with the fusion between the synthesis, intracellular trafficking and degradation of connexins. (1) connexosome and early endosomes. (9) Connexins are sorted from Connexins are cotranslationally inserted into the endoplasmic reticu- early endosomes to lysosomes via late endosomes. (10) Under certain lum. (2) A subpool of newly synthesised connexins undergo endo- conditions, endocytosed connexons may undergo recycling to the plasmic reticulum-associated degradation. (3) During their trafficking plasma membrane. Hemichannels also undergo endocytosis and to the plasma membrane, connexins oligomerise into connexons. (4) recycling, but their endocytic and recycling pathways are poorly After their arrival at the plasma membrane, the connexons can dock characterised (question marks). (11) Multivesicular endosomes can with connexons from adjacent cells to form gap junction channels. (5) fuse with the plasma membrane to secrete exosomes containing con- Connexons may also form functional channels at the non-junctional nexons. (12) Microvesicles containing connexons can be formed by areas of the plasma membrane (also known as hemichannels). (6) Gap the outward budding of the plasma membrane that connexins are causally involved in cancer pathogenesis, seminal ex vivo study, in which they demonstrated the loss possessing both pro- and anti-tumorigenic functions (see the of electrical coupling occurring in rat liver tumours [7]. below section and recent reviews [2, 4]). Significant Since then, several connexin knockout mouse models have advances have been made towards understanding how supported the notion that connexins have tumour- connexins can act as either tumour suppressors or tumour suppressive functions (reviewed in ref. [2]), notably Cx32 promoters depending on the isoform, cancer stage and tis- (multiple organs), Cx43 (breast and lung) and Cx26 sue. In addition to the emerging prognostic value of con- (breast). Cancer cells often display a loss of GJIC due to the nexins in cancer, several studies indicate that connexin dysregulation of connexin expression at multiple levels targeting might have considerable therapeutic implications (Fig. 2). At the transcriptional level, the reduced connexin [5, 6]. This review provides an overview of our current expression in tumours involves transcriptional and epige- understanding of the role of connexins in cancer, with netic mechanisms such as promoter methylation [8–11]. At emphasis on their prognostic and therapeutic potential. the post-transcriptional level, the suppression of connexin expression can be due to connexin-targeting microRNAs [12, 13]. At the protein-synthesis level, internal translation Complex roles of connexins in cancer of N-terminally truncated forms of Cx43 [14–16] has recently been described to regulate connexin membrane Connexins as tumour suppressors targeting [14], and Smad3/ERK-dependent repression of this process was reported to reduce Cx43 gap junctions The idea that GJIC regulates tumour growth was proposed during epithelial-to-mesenchymal transition [17]. At the by Loewenstein and Kanno more than 50 years ago in a post-translational level, many growth factors, oncogenes Connexins in cancer: bridging the gap to the clinic 4431 and tumour-promoting chemicals are potent inducers of depletion of connexins may result in more aggressive cancer Cx43 phosphorylation, which is often associated with the cell growth [29]. In addition to their role in modulating cell inhibition of GJIC [18] and increased Cx43 ubiquitination, proliferation [30], connexins can either promote or prevent endocytosis and degradation [19]. cell death by apoptosis [31]. Such effects may be due to the In accordance with the notion that connexins might act as gap junction-mediated intercellular passage of survival or tumour suppressors, the ectopic expression of connexins in death signals such as Ca2+, IP3 and cAMP [2, 32–34]. cancer cells often partly restores growth control (e.g. refs. Moreover, hemichannels may exchange proapoptotic and [20–25]) and differentiation potential (e.g. refs. [26–28], survival factors between extracellular and intracellular reviewed in ref. [2]). Conversely, the experimental environments [35]. Cell 2 Cell 1 3) Post-translational Post-translational Protein-protein interactions regulation modifications: Phosphorylation Acetylation Multiple Ubiquitination targets SUMOylation Others N-term C-term Internal translation e.g. PP242, Cx43 GJA1-20K mTOR rapamycin Mnk1/2 e.g. CGP 57380 2) mRNA and translational regulation miRNAs miRNA replacement miRNA antagonists 5’UTR 3’UTR CDS Cytoplasm DNMTs DNA methylation 1) Transcription CC C M Nucleus C MC C C M C C C C C M MM Histone acetylation β HATs Promoter Exon 1 Exon 2 HDACs AP-1 SP-1 GJA1 Runx1 Transcription factors β-catenin TCF/LEF 4432 T. Aasen et al. Fig. 2 Dysregulation of connexins in cancer: therapeutic opportunities. suppressors can vary considerably among tissue types and Multiple stages of connexin life cycle are subject to dysregulation cancer stages, as well as among connexin isoforms [2]. during cancer progression, as exemplified by GJA1 (Cx43). (1) Furthermore, there is growing evidence that some connexin Transcription: connexin expression is often reduced (but sometimes increased) in human tumours at the mRNA expression level, of which isoforms are protumorigenic under certain conditions. For multiple pathways are therapeutic targets (text highlighted in red for example, connexins can promote the migration and invasion key targets), including transcription factor activity and epigenetic of tumour cells [44, 45]. In addition, connexins can form silencing by histone acetylation and promoter methylation (promoter heterologous gap junctions between tumour cells and region in green, with C and M illustrating the non-methylated and methylated sites, respectively; blue, some important transcription endothelial cells to facilitate intravasation and extravasation factors regulating Cx43 expression). Histone acetylation can be [46–49]. Connexins can also nurture metastatic growth and modified by targeting histone acetyltransferase enzymes (HATs) or may promote resistance to cancer treatments [50–53]. In histone deacetylases (HDACs), typically promoting and repressing accordance with this notion, a recent study demonstrated transcription, respectively. Transcriptional silencing due to promoter hypermethylation by DNA methyltransferase enzymes (DNMTs) may that brain metastatic cancer cells establish gap junctions also be amenable to therapeutic intervention leading to the restoration with astrocytes to promote tumour growth and chemore- of GJIC. (2) mRNA regulation: mRNA stability and translation is sistance [50]. subject to regulation by multiple cancer-associated microRNAs. Moreover, alternative translation initiation, resulting in the synthesis of truncated forms of Cx43, might regulate Cx43 and have important Connexins and cancer stem cells implications for its dysregulation in cancer. This process is regulated by key cancer signalling pathways such as mTOR and Mnk1/2 and is Emerging experimental evidence indicates that connexins altered during pathological conditions such as hypoxia. Truncated play important roles in cancer stem cell (CSC) biology. forms of Cx43, notably the 20-kDa form named GJA1–20k, may be important for the efficient targeting of Cx43 to the membrane. Indeed, Malignant glioma stem cells (GSCs) express very low levels Smad3/ERK-dependent repression of GJA1–20k was recently shown of Cx43, and GSC stemness is reduced by its transfection to reduce Cx43 gap junctions during epithelial-to-mesenchymal tran- with Cx43 or treatment with peptides containing sequences sition (EMT). (3) Post-translational regulation: connexins frequently of Cx43 that interact with c-Src (Cx43 mimetic peptides) display an aberrant localisation in cancer cells. Phosphorylation and other multiple post-translational events, occurring mainly at their C [54, 55]. A recent work has found that the expression levels terminus, regulate connexin trafficking and stability at the plasma of Cx43 and Cx46 were increased and reduced, respec- membrane. Cx43 is regulated by several kinases that are frequently tively, during GSC differentiation and that targeting of overactivated or overexpressed during cancer development and sus- Cx46 compromises GSC maintenance [56]. While Cx43 is ceptible to pharmacological inhibition, such as mitogen-activated protein kinase (MAPK), protein kinase C (PKC), protein kinase A almost absent in liver CSCs [57], the accumulation of (PKA), cdc2/cyclin B and v-src/c-src. Cx43 is also regulated by cytoplasmic Cx32 has been associated with metastasis and acetylation, ubiquitination and SUMOylation enhanced self-renewal of CSCs in human HuH7 hepatoma cells [58]. Some subtypes of breast cancer cells positive for the putative CSC marker CD44 express Cx43 [59, 60]. There is increasing evidence that connexins can suppress Moreover, Cx43 clearly plays a subtype-dependent role in the growth of cancer cells through channel-independent breast cancer [61]. In triple-negative breast cancers, Cx26 mechanisms [22, 30, 36–39] (Fig. 3). For example, the was shown to be upregulated in CSC population, in which it ectopic expression of the intracellular C terminus (CT) of forms a complex with NANOG and focal adhesion kinase to Cx43 can in some cases inhibit cell proliferation to a similar drive self-renewal (Fig. 3) [62]. extent as full-length protein [24]. Connexins may also mod- Overall, connexins can be both anti- and pro-tumori- ulate the activity of some of their partners by affecting their genic, acting via processes that include the regulation of cellular location, as proposed by Skp2 for Cx50 [40], β- CSCs, and this depends on the connexin isoform and tissue catenin for Cx43 [38], discs large homologue 1 (Dlgh1) for type. Understanding this complexity is the key to the Cx32 [41] and Cx43 [42], or by other mechanisms, such as development of new and efficient therapies. the recruitment of Src together with its endogenous inhibitors CSK and PTEN resulting in a switch from the active to inactive conformation of c-Src [43] (Fig. 3). Because con- Prognostic value of connexins in cancer nexins present a low level of homology within their CT sequences, the channel-independent regulation of cell growth A number of studies suggest that connexins can be inde- is expected to vary considerably among different isoforms. pendent tumour markers predicting both better and worse prognoses. Further insights into this seeming contradiction Connexins as promoters of invasion and metastasis have been provided by recent analytical tools for public databases. The Pathology Atlas within the Human Protein When assessing the role of connexins in cancer pathogen- Atlas database (https://www.proteinatlas.org/pathology) esis, we must consider that their ability to act as tumour includes an open-access database (retrieved from The Connexins in cancer: bridging the gap to the clinic 4433 A B C D Cx43 Cx26 Cx32 Cx50 Dlgh1 Cell migration NANOG Tubulin Smad2/3 release FAK Nuclear Dlgh1 activity inhibition CSK Src Src inhibition Drives self-renewal PTE N in triple-negative breast cancer 4 dd Ne β-catenin Transcriptional SKP2 activity inhibition degradation brin Dre SKP2 ZO-1 n Ezri Fig. 3 Interactions between connexins and proteins that affect tumour have therapeutic potential. b Cx26 has been proposed to maintain a growth and migration. Examples of proteins that interact with specific cancer stem cell phenotype specifically in triple-negative breast cancer regions of connexins and may act as therapy targets. a The interaction cells through its interaction with NANOG and focal adhesion kinase between Cx43 and tubulin is involved in the regulation of cell (FAK). c Cx32 binds to the scaffold protein discs large homologue 1 migration. Similar mechanisms have been proposed for other proteins (Dlgh1) and may control cell proliferation in liver cells through its associated with the cytoskeleton, such as cadherins, catenins, vinculin, interaction with and maintenance of Dlgh1 at the plasma membrane. ZO-1 and drebrin. In addition, Cx43 may compete with the tubulin– The interaction of Dlgh with Cx43 has also been associated with Smad2/3 interaction causing Smad2/3 release. Cx43 binds to c-Src and cancer progression through a mechanism involving the oncoprotein E6 its endogenous inhibitors CSK and PTEN, promoting c-Src inhibition. (see section “Connexins and tumour viruses”). d Cx50 interacts with Cx43, by interacting with β-catenin, prevents the transcriptional and promotes auto-ubiquitination and the subsequent degradation of activity of β-catenin in the nucleus, where it regulates the expression of Skp2, a key negative regulator of the cyclin-dependent kinase (CDK) genes involved in promoting cell malignancy. A similar sequestration inhibitor p27. a–d To complicate this scenario, the phosphorylation of mechanism may occur with drebrin, ezrin or ZO-1. These proteins, and connexins modifies their binding affinities to various protein partners. many others such as Nedd4, also have important roles in regulating For instance, c-Src phosphorylation affects the binding of several Cx43 gap junction plaques, which influence GJIC and therefore may Cx43 partners. GJIC, gap junction intercellular communication Cancer Genome Atlas (TCGA)) containing correlation prediction results for the same type of tumour (e.g. Cx26 vs analyses based on mRNA expression levels with respect to Cx32 in renal cancer). The sub-classification of cancers is the clinical outcome for 17 major cancer types and almost also relevant. For example, the Kaplan–Meier Plotter tool 8000 cancer patients [63]. Those mRNAs encoding con- (http://kmplot.com/analysis) suggests that high levels of nexins identified to have a highly significant prognostic Cx43 mRNA are associated with good and poor prognoses value (P < 0.001) in any of these cancer types are sum- in oestrogen receptor-positive and -negative breast cancer marised in Table 1. Notably, GJA1 (encoding Cx43) is subtypes, respectively [64]. featured in the list of 20 genes most significantly associated Several specific studies have also reported a correlation with an unfavourable prognosis in stomach cancer. Indeed, between connexin mRNA levels and survival or phenotypic for most connexins and cancer types, high levels of con- features. For instance, high levels of Cx43 mRNA in glioma nexin mRNA expression are associated with a poor prog- tumours have been correlated with poor survival [65]. nosis. It is also notable that the mRNA expression of a Expression of connexin mRNA in the surrounding tissue given connexin-encoding gene can be both favourable and may also correlate with a specific tumour phenotype or unfavourable, depending on the type of tumour (Table 1). behaviour. For example, in melanoma, the increased Different connexins can also have differing prognostic expression of Cx26 and Cx30 at the mRNA level in the 4434 T. Aasen et al. Table 1 Significant associations between connexin mRNA expression and cancer prognosis Connexin (gene) Cancer Clinical relevance Survival,a % alive P value (best separation) Cx26 (GJB2) Renal U H = 52% (n = 188), L = 73% (n = 689) 5.98E-04 KIRC U H = 46% (n = 130), L = 70% (n = 398) 7.79E-05 Pancreatic U H = 15% (n = 86), L = 40% (n = 90) 2.90E-04 Glioma U H = 0% (n = 39), L = 11% (n = 114) 5.98E-04 Lungb U H = 43% (n = 795), L = 49% (n = 199) 6.25E-04 LUADb U H = 33% (n = 330), L = 53% (n = 170) 1.45E-05 Cx30.3 (GJB4) Pancreatic U H = 15% (n = 139), L = 66% (n = 37) 3.24E-05 Cervical F H = 74% (n = 181), L = 52% (n = 110) 5.48E-04 Cx31 (GJB3) Pancreatic U H = 15% (n = 70), L = 37% (n = 106) 7.56E-05 Lungb U H = 44% (n = 709), L = 45% (n = 285) 4.56E-04 LUAD b U H = 18% (n = 107), L = 45% (n = 393) 3.54E-08 Cx31.1 (GJB5) Pancreatic U H = 0% (n = 66), L = 42% (n = 110) 1.54E-04 Cx32 (GJB1) Renalb F H = 77% (n = 451), L = 61% (n = 426) 2.85E-06 KIRCb F H = 75% (n = 284), L = 50% (n = 244) 4.88E-08 Cx37 (GJA4) Renal b U H = 67% (n = 700), L = 80% (n = 177) 7.44E-04 KIRCb F H = 68% (n = 417), L = 48% (n = 111) 8.57E-04 KIRPb U H = 58% (n = 77), L = 82% (n = 208) 1.07E-05 Liver F H = 53% (n = 225), L = 39% (n = 140) 3.28E-04 Cx43 (GJA1) Stomach U H = 14% (n = 98), L = 45% (256) 4.99E-05 KIRCb F H = 68% (n = 409), L = 45% (119) 2.60E-04 Cx45 (GJC1) Renal U H = 63% (n = 562), L = 80% (n = 316) 3.55E-08 Urothelial U H = 34% (n = 286), L = 58% (n = 120) 7.83E-04 a Survival: percentage of patients alive after 3 years for glioma and 5 years for all other cancer types. The table and the P values are based on a best- fit model in which patient numbers (n) are stratified into high (H) and low (L) connexin expression groups b Cancer subtype-specific evidence (i.e. only some subtypes correlate). All data are taken from the Human Protein Atlas (https://www.proteinatlas. org/pathology) after analysis by Uhlen et al. [63]. Fragments per kilobase million (FPKM) cut-off (median separation) is set to 1 F connexin expression is favourable for outcome, U connexin expression is unfavourable for outcome, H high expression, L low expression, n number of patients, LUAD lung adenocarcinoma, KIRC kidney renal clear cell carcinoma, KIRP kidney renal papillary cell carcinoma surrounding skin keratinocytes is significantly correlated progression from ductal carcinoma in situ to invasive ductal with malignant features such as tumour thickness and, in the carcinoma [69]. At the protein level, Cx26 is detected in the case of Cx26, metastasis [66]. In lung cancer, the reduced invasive carcinoma foci [69]. In agreement with these expression of Cx43 mRNA in adjacent normal lung tissue findings, Cx26 was recently found to drive CSC self- (due to promoter methylation) is significantly correlated renewal in triple-negative breast cancer [62]. with nodal involvement, suggesting that Cx43 could be a At the DNA level, few studies have provided an insight marker for micrometastasis in non-small cell lung cancer into the role of connexins in cancer. Mutations in Cx43 [8]. Recently, Cx30.3 (GJB4) mRNA expression was have been described in advanced metastatic colon cancer shown to be increased in lung tumours, with the levels in lesions [70]. However, current large-scale genetic studies blood buffy coat samples serving as a biomarker for diag- and analytical tools (e.g. Intogen https://www.intogen.org nosis and poor prognosis [67]. Mechanistically, Cx30.3 was or COSMIC http://cancer.sanger.ac.uk/cosmic) suggest shown to promote tumour growth and induce chemoresis- relatively low mutation rates in connexin-encoding genes tance via the MET-induced activation of Src. and, based on genetic data, none of the 21 connexin genes As shown in Table 1, Cx26 mRNA expression is asso- are classified as tumour suppressors or oncogenes. Never- ciated with a poor prognosis in various cancers. For theless, individual genetic studies merit follow-up. For example, high levels of Cx26 mRNA are associated with example, the C1019T polymorphism in GJA4 (encoding reduced overall survival in pancreatic cancer (P < 0.001) Cx37) is associated with Helicobacter pylori infection in [68]. Microarray analysis (and confirmatory quantitative patients with gastric cancer [71]. In glioma, GJB6 (encoding real-time polymerase chain reaction analysis) of breast Cx30) was deleted in 25.8% of the 751 analysed tumours cancer samples show the upregulation of Cx26 during and was mutated in 15.8% of 158 tumours [51]. However, Connexins in cancer: bridging the gap to the clinic 4435 again, the overall patient survival was not correlated with cancer stage and type/subtype as well as connexin isoform, the presence of GJB6 deletions (517 tumours from TCGA complicates their clinical applicability. Standardisation is and 67 tumours from REMBRANDT) or GJB6 mRNA also difficult in terms of reliable antibodies and evaluation expression (in patients from TCGA). However, at the pro- parameters such as the quantification of subcellular con- tein level, high Cx30 expression adversely influenced the nexin localisation (e.g. membrane vs cytoplasm), which can survival (Table 2). significantly influence the prediction of the clinical A number of immunohistochemical studies of connexins outcome. in cancer—most on Cx26 and Cx43—have described spe- cific prognostic associations. As with mRNA-expression studies, connexin protein expression has been associated Targeting of connexins in cancer: with both good and poor prognoses. The studies specifically therapeutic opportunities and challenges reporting the significant findings related to patient survival are summarised in Table 2. The complex and multifunctional role of connexins in In accordance with mRNA-based studies, the expression cancer provides a wide spectrum of therapeutic opportu- of different connexins at the protein level can have an nities and challenges. Increasing numbers of drugs, peptides opposing prognostic value depending on cancer type, cancer and RNAi approaches are available to inhibit or enhance subtype and connexin isoform. For example, in the same GJIC, hemichannel activity or connexin protein-signalling breast cancer series, elevated Cx43 and Cx30 levels have activity. Below, we highlight a selection of in vivo studies been associated with improved and worse breast cancer demonstrating that the cancer phenotype is altered as a outcomes, respectively [64]. Another shared finding between direct consequence of such an experimental targeting of the mRNA and protein studies is the observation that high connexins. levels of Cx26 are associated with a poor prognosis in most studies/cancers (Tables 1 and 2). This observation is sup- Modulation of connexin expression and GJIC ported by additional studies evaluating tumour histology or aggressiveness, but not the overall survival, including those The restoration of GJIC and/or connexin expression has of thyroid cancer [72] and bladder cancer [73]. Connexin long been a potential therapeutic strategy. GJIC can be expression has also been proposed as a diagnostic aid in augmented either by enhancing the permeability of existing relation to specific tumour subtypes. In gastric cancer, Cx30 gap junctions or by increasing the expression of connexins was proposed as a potential marker for an intestinal-type and thereby elevating the number of open gap junction cancer phenotype, and the negative expression of Cx30 channels [81]. Many fungal and plant-based compounds, as correlates with a more advanced T grade ( < 0.0001), N well as an increasing number of synthetic chemical com- grade (0.0123) and tumour stage (0.0014) [74]. An asso- pounds, either increase GJIC or prevent the loss of GJIC in ciation between the expression level of connexins and response to cellular exposure to tumour promoters, which is metastasis can be found for several cancer types. For often associated with reduced tumour growth in vivo. instance, Cx43 expression is reduced in primary gastric Compounds shown to modify connexins or GJIC and affect cancer but increased in matched metastatic lymph nodes the tumour phenotype in vivo are summarised in Table 3. It [75]. It should also be noted that, in a number of studies, is worth noting that many chemotherapeutic agents also connexin proteins were overexpressed but mislocalised in increase connexin expression and/or GJIC [82]. For exam- the cytoplasm. Prominent examples include Cx26 in pan- ple, docetaxel increases Cx43 expression in murine salivary creatic [76] and colon [77] cancer, and Cx43 and Cx32 in gland carcinoma to reduce tumour growth [83]. This has prostate cancer [78]. Nuclear Cx43 has also been reported in implications for combinatorial effects in cancer therapy. a number of tumours (e.g. colon cancer [38] and gliomas However, because connexins under specific conditions [79]). It remains to be seen how this mislocalisation corre- can facilitate malignant features, in particular, metastasis or lates with a recent report showing that nuclear Cx43 (or the growth of metastases [2], there are situations in which the CT alone) can act as a direct nuclear transcription factor that blockade of GJIC or other functions of connexins may offer induces the expression of N-cadherin, an important regulator distinct therapeutic opportunities. The inhibition of GJIC can of processes including epithelial-to-mesenchymal transition be achieved through a number of different mechanisms and cell migration [80]. (Table 3 and recent comprehensive reviews [5, 84]). Despite In summary, connexins are potential prognostic markers their non-specificity, many chemical gap junction channel in cancer, but a number of limitations need to be addressed inhibitors have been extensively used, even in vivo. For before any clinical application. The fact that connexins have instance, the GJIC blocker oleamide (a fatty-acid derivative) both pro- and antitumourigenic properties and are asso- reduces the formation of pulmonary and hepatic metastatic ciated with both good and poor prognoses, depending on foci and increases the overall survival of mice injected Table 2 Direct correlations between connexin protein levels and patient survival 4436 Cancer Cx Association Clinical relevance P value Comments Reference Bladder Cx43 U High Cx43 = reduced PFS and RFS in stage pTa and P < 0.001 High Cx43 associated with a higher tumour grade, Poyet et al. [190] pT1 patients multiplicity and increased proliferation (all P < 0.05) Bone Cx43 F Reduced Cx43 = reduced PFS P = 0.024 Study examined giant-cell tumour of bone Balla et al. [191] Breast Cx26 U High Cx26 = worse prognosis P < 0.001 Cx26 was correlated with a large tumour size (P = 0.013) Naoi et al. [192] and a high histological grade (P = 0.043). Multivariate analysis: Cx26 (P < 0.05) expression an independent prognostic factor Cx26 U Decreased Cx26 expression ( < 5%) post-chemotherapy P = 0.011 Connexin expression may improve the assessment of the Teleki et al. [193] = better OS pathological response and refine intermediate prognostic Cx46 F Cx46 expression ( > 20%) pre- and post-chemotherapy P = 0.002–0.05 subgroups after neoadjuvant chemotherapy = better OS in the intermediate prognostic subgroups Cx26 F High Cx26 = better RFS P = 0.013 Contradicted earlier study findings at the mRNA level (see Teleki et al. [64] ref. [193]) Cx32 U High Cx32 = worse RFS P = 0.027 High Cx32 mRNA expression = improved RFS. Antibody staining not shown Cx30 U High Cx30 = reduced RFS in grade 3 patients P = 0.016 Independent prognostic markers, but mRNA analysis Cx43 F High Cx43 = better RFS P = 0.026 suggested high tumour subtype-specific differences Cx43 F High Cx43 = better OS P < 0.001 Independent predictor of survival and distant metastasis- Chasampalioti free survival et al. [194] Colorectal Cx26 U High Cx26 expression (cytoplasm) = shorter DFS P = 0.041 High expression = less-differentiated histology (P = Ezumi et al. [77] Shorter DFS (P = 0.041) and lung metastasis-free P = 0.028 0.0053) and venous invasion (P = 0.0084). Lung survival (P = 0.028) metastases, but not liver and lymph node metastases, expressed higher Cx26 than the CRC series or corresponding primary CRCs (P < 0.0001 and P = 0.0001, respectively) Cx26 F Cx26-positive tumours associated with significantly P = 0.0128 Cx26 = independent prognostic factor (P < 0.05). Nomura et al. longer survival than Cx26-negative tumours Correlation identified between Cx26 and recurrence, [195] histology and p53 expression (P < 0.05) Cx26 U High Cx26 = shorter patient survival P = 0.02 Cx26 associated with a higher tumour grade (P = 0.02) Knosel et al. [196] Cx43 F Loss of Cx43 expression = shorter relapse-free P = 0.017 Depends on subcellular Cx43 staining. Not significant if Sirnes et al. [38] survival and OS more advanced stage III and IV tumours are analysed Oesophageal Cx26 U High levels = reduced OS P = 0.007 Correlated with lymph node metastasis (P = 0.014) and the Inose et al. [197] SCC number of metastatic lymph nodes (P = 0.047) Cx43 U High levels = reduced OS P = 0.018 Independent prognostic indicator Tanaka et al. [198] T. Aasen et al. Table 2 (continued) Cancer Cx Association Clinical relevance P value Comments Reference Gastric Cx26 F High Cx26 = better OS P = 0.002 Negative associations between Cx26 expression and Liu et al. [199] clinicopathologic features (all P < 0.05). High Cx26 associated with histological intestinal-type (P = 0.017) and early-stage gastric carcinoma. Positive Cx26 expression an independent prognostic predictor of intestinal-type GC (P = 0.023, hazard ratio = 2.019) Glioma Cx30 U High Cx30 = worse OS P < 0.05 Correlated with patient survival after irradiation. No Artesi et al. [51] correlation at DNA or mRNA level HNSCC Cx43 F Cx43 expression = better disease-specific survival of P = 0.004 Cx43 positively correlated with p53 expression (P = Danos et al. [200] patients 0.036). No statistical association between Cx43 levels and T, N or M stage Connexins in cancer: bridging the gap to the clinic Liver (hepatitis Cx43 F High Cx43 = delayed tumour recurrence (P = 0.001), All P ⩽ 0.026 In patients with serum alpha-fetoprotein < 400 ng/ml, Cx43 Wang et al. [201] B-related) HCC longer DFS (P = 0.026) and OS (P = 0.000) in patients expression is an independent predictor of later recurrence with serum alpha-fetoprotein < 400 ng/ml and longer OS Lung Cx26 U Cx26-positive SCC tumours associated with increased P = 0.0391 Multivariate analysis demonstrated that Cx26 expression (P Ito et al. [202] survival (cancer-related deaths, not OS) = 0.0448) is an independent prognostic predictor Cx43 F Better mean OS and PFS in NSCLC patients with high P < 0.001 Cx43, a reliable surrogate marker for predicting the Du et al. [203] Cx43, including the chemotherapy-responder group chemotherapy response and prognosis of patients with advanced NSCLC Oral SCC Cx43 U High membrane expression of Cx43 = short OS P = 0.0088 No correlation for Cx26 or Cx45. Cx43 expression in Brockmeyer et al. dysplasia-free mucosa may indicate a very early stage of [204] tumour progression Pancreatic Cx43 F Reduced Cx43 associated with higher TNM stage and P < 0.001 Also associated with the degree of differentiation (P = Liang et al. [205] lymph-node metastasis 0.002) Prostate Cx26 F Low Cx26 expression in noncancerous tissue in P = 0.002 Cx26 is only predictive when assessed in the noncancerous Bijnsdorp et al. prostatectomy sections = reduced BRFS and risk of areas of the tissue [206] metastasis Cx43 F Reduced Cx43 = shorter postoperative BRFS P < 0.001 Patients with lower Cx43 have higher preoperative PSA Benko et al. [207] levels Cx43 F Reduced Cx43 = shorter postoperative BRFS P < 0.001 Reduced levels of Cx43 associated with high levels of Xu et al. [208] preoperative PSA, a high Gleason score, an advanced pT stage and seminal vesicle invasion (all P < 0.05) Sarcoma Cx43 U Cx43 score correlated with OS P = 0.025 Possible oncogenic and prognostic roles for Cx43 and Cx26 Bui et al. [209] (EWS/PNET) in EWS/PNET. The lack of membranous staining suggests a GJIC-independent mechanism F connexin expression is favourable for outcome, U connexin expression is unfavourable for outcome, BRFS biochemical recurrence-free survival, CRC colorectal cancer, DFS disease-free survival, EWS/PNET Ewing’s sarcoma/primitive neuroectodermal tumour, GC gastric cancer, GJIC gap junctional intercellular communication, OS overall survival, HCC hepatocellular carcinoma, HNSCC head and neck squamous cell carcinoma, PFS progression-free survival, PSA prostate-specific antigen, RFS relapse-free survival, NSCLC non-small cell lung cancer, SCC 4437 squamous cell carcinoma, TNM tumour node and metastases Table 3 Modulators of connexins and/or GJIC that alter tumour response in vivo 4438 Compound Derived from Mechanism of action/effect Model/cells used In vivo tumour effect Cx Reference GJIC enhancers Resveratrol Natural phenol produced by multiple Increased Cx43 expression and reduced Male C57BL/6 mice inoculated with Decreased tumour size and F Buhrmann et al. plants MAPK and NF-κB activation B16F10 cells prolonged survival time [210] Orthotopic implantation of MIA PaCa-2 Tumour growth inhibition, F Harikumar et al. cells in athymic nu/nu nude mice synergy with gemcitabine in MIA [211] PaCa-2 cells Docetaxel Chemotherapeutic drug (a semisynthetic Increased Cx43 expression inhibits an PC-3 and LNCaP cells injected into male Decreased tumour size F Fukushima et al. analogue of paclitaxel) IL-1α-induced reduction in GJIC BALB/c nu/nu mice [212] PQ1, PQ7 Quinolones (heterocyclic aromatic Increased Cx43 expression and GJIC PyVT spontaneous mammary tumour Significantly reduced tumour F Shishido et al. [213] organic compounds) (PQ7 also shown to downregulate Cx46) mouse model growth T47D breast cancer cells subcutaneously Reduced tumour growth in vivo F Shishido et al. [214] injected into the inguinal region of the alone and in combination with mammary fat pad cisplatin All-trans retinoic acid (ATRA) An active metabolite of vitamin A Increased connexin expression Daoy medulloblastoma cells, Decreased tumour size F Li et al. [215] implementation in mice Injection of HepG2 human hepatoma cells Increased apoptosis and connexin F Wu et al. [216] into athymic BALB/c nu/nu mice. expression in tumours Injection of ATRA Valproic acid A branched short-chain fatty acid Increased Cx26 and Cx43 expression Intracranial implantation of human glioma Enhanced antitumour effect of F Ryu et al. [217] derivative of valeric acid U87 cells into the brain of nude mice mesenchymal stem cell-mediated HSV-TK gene therapy in intracranial glioma Apigenin, lovastatin Apigenin: a flavonoid; lovastatin: an Increased Cx43 expression and GJIC Injection of MCA38 (mouse colon Decreased tumour growth due to a F Touraine et al. HMG-CoA reductase inhibitor adenocarcinoma) into C57BL/6 mice combination of HSV-TK/GCV [218] and apigenin/lovastatin Shikonin, notoginsenoside R1, aconitine Natural compounds In combination, restored GJIC upon Urethane-induced lung adenocarcinoma in Decreased incidence of mouse F Liu et al. [219] urethane exposure mice lung carcinogenesis and decreased tumour volume GJIC inhibitors Oleamide Naturally occurring amide derived from Structural inhibition of gap junction MDA-MB-231 breast cancer cells Antimetastatic activity U Zibara et al. [85] the fatty acid oleic acid channels and decreased GJIC BL6 melanoma cells Reduced spontaneous metastasis U Ito et al. [220] 18‑α‑Glycyrrhetinic acid (18‑α‑GA) A bioactive triterpenoid found in Altered connexon particle packing in gap Cx26- and Cx43-mediated heterotypic Reduced antitumour effect of F Ryu et al. [217] liquorice junction plaques and decreased GJIC ‘bystander’ effect between endothelial mesenchymal stem cell-mediated cells expressing HSV-TK suicide gene HSV-TK gene therapy in and tumour cells intracranial glioma MI-18, MI-22 Derived from oleamide Decreased GJIC, structurally inhibited BL6 melanoma cells Reduced spontaneous metastasis U Ito et al., Ohba et al. Cx26 and Cx32 channels but not Cx40, [220, 221] Cx43 and Cx45 (possibly also interferes with cytoplasmic connexin functions) Camellia oil Natural oil rich in oleamide-like oleic Inhibited Cx26 channels (probably BL6 melanoma cells Reduced spontaneous metastasis U Miura et al. [222] acid structural inhibition) and decreased GJIC Meclofenamate A fenamate (anthranilic acid)-derived, Decreased GJIC (possibly due to steric Breast and lung carcinoma cell lines Reduced brain metastasis growth, U Chen et al. [50] FDA-approved drug for joint and hindrance) enhanced effect of cytotoxic drugs muscular pain, arthritis and dysmenorrhoea Tonabersat Tonabersat (SB-220453), a novel Decreased GJIC (possibly due to steric Breast and lung carcinoma cell lines Reduced brain metastasis growth, U Chen et al. [50] benzopyran derivative used against hindrance) enhanced effect of cytotoxic drugs migraine in human trials Mefloquine Licensed anti-malaria drug GJIC and hemichannel inhibitor (likely MCF-7 human luminal breast cancer cells Inhibition of bone metastasis by U Wang et al. [86] structural inhibition) limiting calcium signalling Arsenic trioxide FDA-approved drug for leukaemia Downregulation of Cx43 (GJA1) gene MCF-7 human luminal breast cancer cells Inhibition of bone metastasis by U Wang et al. [86] expression (in co-culture systems) limiting calcium signalling T. Aasen et al. Connexins in cancer: bridging the gap to the clinic 4439 intravenously with MDA-MB-231 breast cancer cells [85]. Yulyana et al. [136] Stoletov et al. [158] F GJIC or connexin expression is favourable for outcome, GJIC gap junction intercellular communication, HSV-TK herpes simplex virus thymidine kinase, TRAIL TNF-related apoptosis-inducing Murphy et al. [65] Alonso et al. [87] Wang et al. [86] Mechanistically, oleamide-induced loss of heterologous GJIC between tumour cells and endothelial cells was sug- Reference gested to interfere with cancer-cell extravasation. Using a brain metastasis mouse model, Massagué and colleagues recently demonstrated the potent growth inhibition of brain metastases from breast and lung cancer cells in response to Cx U U U U U blockade of heterologous GJIC between the cancer cells and astrocytes [50]. Inhibition of heterologous GJIC was Fewer Cx40 channels in endothelial cells Subcutaneous injection of B16F10 mouse Reduced tumour angiogenesis and Enhanced effect of TRAIL therapy extravasation and blood vessel co- achieved either by the knockdown of Cx43 expression or by MCF-7 human luminal breast cancer cells Inhibition of bone metastasis by Inhibited brain colonisation by the use of the gap junction channel inhibitors tonabersat or limiting calcium signalling Sensitised chemoresistant blockade of tumour cell meclofenamate, which pass the blood–brain barrier [50]. In vivo tumour effect glioblastoma cells to Tonabersat has been used as a migraine prophylaxis drug, tumour growth whereas meclofenamate is an FDA-approved anti-inflam- temozolomide matory drug for oral administration. Mechanistically, these option GJIC inhibitors seem to inhibit tumour growth by blocking a cGAMP-mediated signalling cascade that orchestrates para- papillomavirus oncogene-expressing cells zebrafish and chicken embryo models of Intracranial glioma model using primary Breast cancer and melanoma cells using derived from C57/Bl6 lung epithelium) crine signalling between host and tumour cells [50]. Con- Inhibited Cx43-CT interaction with ZO-1 Subcutaneous injection of glioma cells melanoma cells or TC-1 cells (human sequently, as the first cancer clinical trial in which GJIC is specifically being targeted, meclofenamate is being tested in human patients with recurrent or progressive brain metas- human glioma isolates (LN229/GSC) in mice tasis (https://clinicaltrials.gov/ct2/show/NCT02429570?id= Model/cells used brain metastasis NCT02429570&rank=1&load=cart). Recently, a novel role of Cx43-based gap junctions was also identified in breast cancer bone metastasis [86]. Gap junctions were found to mediate calcium flow from osteogenic cells to the cancer cells. Blocking such heterocellular intercellular calcium (and other proteins?). Inhibited Cx43 GJIC inhibition and reduced Cx43 transfers from the osteogenic niche to cancer cells, by gap hemichannels (enhanced GJIC?) junction inhibitors (carbenoxelone and arsenite trioxide) Mechanism of action/effect or by the ablation of Cx43 in either the osteogenic cells or causing decreased GJIC the cancer cells, prevented bone metastasis progression in mice [86]. glycyrrhetinic acid used in human trials expression Thus, many of the studies reporting the inhibitory effects on tumour growth upon inhibition of GJIC mechanistically ligand, U GJIC or connexin expression is unfavourable for outcome relate this effect to a loss of heterologous GJIC between tumour cells and cells in their microenvironment (as reviewed in [50, 85] and discussed elsewhere in this A semisynthetic derivative of manuscript). Other successful strategies include the peptide- mediated inactivation of Cx40 in endothelial cells causing the inhibition of tumour angiogenesis in vivo and subse- Synthetic peptide Synthetic peptide Derived from quently decreased tumour growth [87] (see Table 3 and section “Emerging concepts”). αCT1, a peptide that blocks ZO-1 and potentially other proteins from interacting with the Cx43-CT (see section “Targeting of connexins in can- Potent pannexin channel inhibitor cer”), has been demonstrated to prevent temozolomide resistance in human glioblastoma cell lines [65]. The use of peptides arguably provides for a more specific targeting compared with the more general chemical GJIC inhibitors Table 3 (continued) and may reduce potential adverse effects. Cx43-blocking a Carbenoxolone antibodies also reduce tumour growth in murine models. Notably, the intravenous administration of a monoclonal Compound Gap40 αCT1 antibody targeting the second extracellular loop of Cx43 reduces glioma growth and survival of experimental a 4440 T. Aasen et al. animals when used alone [88] or in combination with penetrating sequences (TAT-Cx43266-283) can mimic the standard cancer therapy approaches [89]. The same anti- anti-oncogenic effect of Cx43 in glioma cells such as pri- body has been used as a guidance system to deliver diag- mary GSCs [55, 100]. nostic markers or therapeutic compounds, such as cisplatin, In contrast, some proteins may favour cell growth or to Cx43-positive high-grade gliomas [90, 91]. migration upon interaction with Cx43-CT. For instance, Cx43 may interact with p21-activated protein kinase 1, Targeting of hemichannels which activates the MAPK p38 to increase the migration of HeLa cells [101], or may compete with tubulin–Smad2/3 Connexin hemichannels are normally closed but, when interaction causing Smad2/3 release [102] (Fig. 3). Other activated, they can release autocrine and paracrine signals, Cx43-interaction partners regulate the Cx43 protein level or such as NAD+, glutamate or ATP, that can affect cell subcellular localisation. Disruption of these interactions proliferation and survival [92]. For instance, ATP release may have therapeutic potential. αCT1 is a Cx43-CT activates the AKT/AMPK/mTOR signalling pathway, thus mimetic peptide that comprises the ZO-1 PDZ-binding linking hemichannel activity with cell proliferation and domain of Cx43 fused to a cell-penetration sequence, which survival [93]. Several lines of evidence suggest that Cx43 was designed as a tool to disrupt the interaction between hemichannels are involved in promoting tumour growth. endogenous ZO-1 and Cx43 [103]. Consequently, αCT1 Antibodies against the second extracellular loop domain of can increase the size of Cx43 gap junction plaques and Cx43, which blocks hemichannels [94], reduce glioma increase GJIC to the detriment of hemichannel activity tumours generated with C6 cells in rats [89]. Similarly, the [104]. αCT1 increases the sensitivity of glioma cells to peptide αCT1 increases the sensitivity to chemotherapy in temozolomide [65] and of breast cancer cells to tamoxifen glioma cells, possibly through the inhibition of Cx43 and lapatinib [105]. In an ongoing clinical trial in dogs with hemichannel activity [65]. On the other hand, Cx43 hemi- naturally occurring high-grade glioma, a slow controlled channels in osteoblast bone cells may provide an intrinsic release of αCT1 encapsulated in a polymer is being tested in self-defence mechanism against breast cancer metastasis combination with chemotherapy [106]. [95]. Among the other molecules released by Cx43 hemi- channels in osteoblasts, ATP acts as a paracrine signal that Enhanced and synergistic therapeutic effects of triggers an inflammatory cascade to inhibit the migration, connexin targeting invasion and anchorage-independent growth of breast can- cer cells [96]. Consequently, the antibody blockade of The bystander effect: kiss-of-life or kiss-of-death? osteocyte Cx43 hemichannels increases bone metastasis in mice [95]. Collectively, these data suggest that Cx43 The bystander effect underpins a key therapeutic strategy in hemichannel activity in healthy tissue cells may have a anticancer approaches and significantly enhances suicide beneficial effect by preventing metastasis, whereas hemi- gene therapy strategies [107]. A well-established suicide channels in tumour cells may favour their growth. gene therapy approach involves the viral transduction of the herpes simplex virus (HSV) thymidine kinase (TK) gene Targeting of the connexin interactome into cancer cells [108]. The viral TK enzyme phosphor- ylates the nucleoside analogue ganciclovir (GCV), which Connexins interact with a wide variety of proteins that, causes chain termination during DNA replication and leads independent of channel activity, can affect cancer cell to tumour cell death. Many reports in the 1990s revealed phenotype, including cell growth, migration and differ- that neighbouring tumour cells that were not transduced entiation (Fig. 3; recently reviewed in refs. [97, 98]). Stra- with the TK/GCV suicide system also died. This was tegies that mimic, promote or disrupt some of these specific demonstrated in vitro and in vivo to be due to the GJIC- interactions could be used in cancer therapy. The fact that mediated transfer of activated GCV and/or other toxic Cx43-CT is a disordered region favours the use of mimetic metabolites [109, 110]. This led to the hypothesis that the peptides that can restore or interfere with its functions. In induction of connexin expression or re-establishment of tumour cells with low levels of Cx43, the restoration of GJIC might potentiate the bystander effect [111], a claim Cx43 function may reduce cell proliferation. In glioma substantiated early on, both in vitro [109, 112] and in vivo cells, this occurs via the inhibition of the oncogenic activity [113]. This theory is supported by many studies (e.g. refs. of c-Src [99]. The mechanism by which Cx43 reduces c-Src [114–117]), including a recent one in breast cancer [118] activity resides in a short region of Cx43-CT (266–283) that and liver cancer, in which the co-expression of Cx43 and acts as a docking platform for c-Src together with its SUMO1 in liver CSCs increased GJIC and their sensitivity endogenous inhibitors PTEN and CSK [43]. Notably, the to HSV-TK/GCV therapy in vitro and in vivo [57]. How- mimetic peptides of this region fused to the TAT- ever, other studies reported no evidence for a GJIC- Connexins in cancer: bridging the gap to the clinic 4441 mediated bystander effect [119]. Moreover, GJIC can pro- breast cancer cells [130]. Cx32 potentiates the cytotoxicity tect transduced cells from the toxic active GCV by allowing of vinblastine and Src inhibitors in renal cell carcinoma effective drug dilution throughout the cell, preventing cell cells [131], whereas Cx26 increases the effect of cisplatin in death in a so-called “Good Samaritan” effect [120, 121]. human bladder cancer cells [132] and of doxorubicin in The bystander effect is important in a wide range of prostate cancer cells [133]. Both channel-dependent and circumstances other than suicide gene therapy. Notably, -independent mechanisms have been suggested to underlie some chemotherapeutic drugs or radiation-induced metabo- this effect [130, 134]. lites such as DNA damage-induced sensing/signalling In other contexts, connexins can have an inverse effect molecules or cytoplasmic irradiation responders (reactive with respect to drug sensitivity. Reduced Cx43 expression oxygen species, reactive nitrogen species, Ca2+ and cyto- is associated with increased drug sensitivity in glioma cells kines) can also transmit through GJIC, inducing a bystander [52, 134–136], whereas the upregulation of Cx26 is asso- effect [120]. However, the strength and type of the signal ciated with gefitinib resistance in lung cancer cells [137]. may have distinct and opposing effects: (1) toxic signals kill The specific inhibition of Cx46-mediated GJIC in GSCs neighbouring cells (which may be tumour cells or healthy attenuates proliferation, self-renewal and tumour growth cells) or (2) toxic signals are diluted into neighbouring cells, and synergises with temozolomide to induce apoptosis favouring the survival of the targeted cells. Importantly, this [138]. There is also in vivo evidence to support this implicates GJIC in coordinating cellular and tissue responses dichotomy in relation to therapeutic efficacy. For instance, to carcinogens, radiation and chemotherapies and leads to shRNA-mediated knockdown of Cx43 or inhibition of GJIC the possible strategies (via both enhanced or blocked GJIC) by meclofenamate and tonabersat strongly potentiates the for either protecting undamaged non-targeted cells against effect of carboplatin-based chemotherapy on brain metas- such signals or potentiating the signals in order to kill tases from breast and lung carcinoma cells [50]. Moreover, tumour cells. Application of this knowledge to other aspects the efficacy of tumour necrosis factor-related apoptosis- of connexin cancer biology is important. inducing ligand therapy is enhanced when combined with the GJIC inhibitor carbenoxolone in an intracranial glioma Need for combinatorial therapy model [136]. The combined use of tumour necrosis factor- related apoptosis-inducing ligand and carbenoxolone could Within the field of cancer therapy, there is a general con- offer a favourable alternative for the treatment of glioma, sensus on the need for rational combinatorial targeted particularly considering the low cytotoxic nature of carbe- therapy. Synergistic effects on cancer cell growth as a result noxolone. Notably, carbenoxolone, as well as peptides tar- of enhanced connexin expression or GJIC in combination geting Cx43, attenuates cancer-induced bone pain [139], with drugs targeting other cellular processes have been highlighting another facet of cancer care in which connexins described in many studies [122–127] (Table 3). For can be further explored. instance, kanglaite, a natural plant seed compound that upregulates Cx43 expression, sensitises colorectal cancer cells to Taxol [124]. Simvastatin (a statin) induces GJIC and Emerging concepts enhances the effect of platinum-based chemotherapeutic drugs [126]. In addition, protection can be provided against Role of connexins in tumour microenvironment the non-desired cytotoxic effects of cisplatin on healthy cells, including reproductive testicular Sertoli cells [127]. That GJIC between normal cells and tumour cells (hetero- Such dual kiss-of-life/kiss-of-death bystander effects can logous GJIC) can inhibit the growth of tumour cells was contribute to the overall chemotherapeutic outcome on shown in the mid 1980s by Loewenstein and colleagues multiple levels. It also poses a therapeutic challenge. The [140]. This phenomenon has since been demonstrated in putative positive and negative effects such therapies can many cancer types and model systems, as recently reviewed provide in relation to tumour growth must therefore be [141]. Heterologous GJIC between tumour cells and the carefully assessed in order to avoid a possible worsening cells within their microenvironment (Fig. 4) has been linked rather than an improvement of the clinical outcome. to both positive and negative effects on tumour progression The therapeutic response of cancer cells can also be and therapy resistance. affected by the direct overexpression of connexins, in an isoform-specific manner. For instance, Cx43 overexpression Communication between cancer cells and immune cells can enhance the sensitivity to common chemotherapeutic drugs such as doxorubicin, fluorouracil and oxaliplatin in Tumour cells deliver antigenic peptides through gap junctions human gastric cancer cells [128], etoposide, paclitaxel and to dendritic cells, and this cross-presentation is associated doxorubicin in glioma cells [129], and artesunate in MCF-7 with enhanced immune-mediated tumour elimination. 4442 T. Aasen et al. Cancer- associated fibroblasts Extracellular Immune vesicles cells Hemichannels Gap junctions Tunneling GJIC-competent nanotubes GJIC-incompetent cancer cells cancer cells ? Epithelial cells Vascular (non-cancerous) and lymphatic endothelial cells Fig. 4 Connexins and the tumour stroma. GJIC can occur between tunnelling nanotubes (microtubes) or extracellular vesicle function cancer cells or in a heterocellular manner between cancer cells and (e.g. exosomes). Other tumours, or parts of tumours, are devoid of nearby cells such as noncancerous epithelial tissue cells and stromal GJIC and may or may not express connexins at high levels in the cells, including cancer-associated fibroblasts, immune cells and vas- cytoplasm or nucleus, thus escaping the direct GJIC with surrounding cular and lymphatic endothelial cells. In addition, there is crosstalk via cells. This may be associated with a reduced polarity and cell–cell the hemichannel release of autocrine and paracrine signals. These adhesion. The benefits and drawbacks of maintained GJIC are likely signals influence tumour growth both positively and negatively in a tissue and stage dependent. An understanding of this complex network context-dependent manner and help to regulate apoptosis, prolifera- of signals is essential to move forward with additional therapeutic tion, invasion, intravasation and extravasation. In addition, connexins strategies of targeting connexins in cancer. GJIC, gap junction inter- are thought to be implicated in other communication forms, as a part of cellular communication Presentation of antigenic peptides between melanoma cells [148]. Overexpression of Cx43 in melanoma and breast and dendritic cells increases the melanoma-specific T-cell cancer cells suppresses tumour angiogenesis [28, 149]. In response [142]. The infection of melanoma tumours with accordance with these findings, Cx43 knockdown in mel- Salmonella bacteria induces Cx43 expression, which enhan- anoma cells increases vessel density [149]. The silencing of ces the antigen presentation into infiltrating dendritic cells, Cx43 in breast cancer cells results in increased vascular activating an antitumour response [143]. Cx43-mediated endothelial growth factor expression and decreased throm- GJIC may also improve the immune surveillance of bospondin expression [29]. In contrast, the overexpression tumours via the activation of natural killer cells [144, 145]. of Cx26 or Cx43 in breast cancer cells is associated with the GJIC-mediated transfer of miRNAs from immune cells to upregulation and secretion of IL-6 and MCP-1, which hepatocarcinoma cells seems to inhibit tumour growth [146]. inhibit endothelial cell tube formation in vitro and tumour These studies (and others, as reviewed recently [141]) are of vascularisation in vivo [28]. particular interest in relation to immunotherapy and tumour Endothelial connexins also play important roles in vaccines [141]. Connexins also regulate many other functions tumour angiogenesis. The knockdown of endothelial Cx37, within the immune and lymphatic systems, which awaits an Cx40 or Cx43 or the pharmacologic inhibition of GJIC additional analysis [147]. diminishes the angiogenic sprouting of endothelial cells in in vitro studies [150]. A recent work using a combination of Angiogenesis and communication between cancer cells and different in vitro, ex vivo and in vivo models has shown that endothelial cells targeting of endothelial Cx40 decreases tumour growth by reducing angiogenesis and improving vessel perfusion [87]. Angiogenesis, the formation of new capillaries from pre- This was demonstrated both in mouse endothelial-specific existing blood vessels, is required for cancer progression Cx40 knockout models and by the injection of the specific Connexins in cancer: bridging the gap to the clinic 4443 peptide 40Gap27 that binds to the extracellular loop of Cx40 Glioma cells can also become more resistant to che- and blocks channel activity [151]. motherapy- or radiotherapy-induced cell death through both Because endothelial and myoendothelial gap junctions homocellular and heterocellular GJIC pathways [51, 53, are fundamentally important for a coordinated vessel 165], as well as via Cx43-specific GJIC-independent response over longer distances [152, 153], a therapeutic mechanisms [166]. GJIC between reactive astrocytes and approach focused on vascular connexins would be con- melanoma cells protects against chemotherapy through the ceivable through the specialised targeting of tumour vessels sequestration of calcium [34]. A recent study has described in combination with anti-angiogenic strategies. another signalling network between astrocytes and breast There is also evidence of direct communication between and lung carcinoma cells involving cGAMP and the sti- tumour cells and endothelial cells (Fig. 4). Recent studies mulator of interferon genes (STING) pathway (see section have demonstrated GJIC-mediated heterotypic exchange of “Modulation of connexin expression and GJIC”), and the microRNAs between cancer cells and endothelial cells. inhibition of GJIC or Cx43 expression was found to sub- Notably, endothelial cells deliver miR-145–5p to colon stantially reduce the brain metastatic burden and enhance cancer cells via gap junctions, causing the upregulation of chemotherapy efficacy [50]. In summary, GJIC between Cx43 expression and reduced angiogenesis [154]. More- tumour cells and astrocytes promotes colonisation, resis- over, the inhibition of GJIC by carbenoxolone blocks the tance to chemotherapy and survival of tumour cells in the interchange of specific cancer-associated microRNAs brain [34, 50, 164, 167]. Targeting of this axis may provide between human microvascular endothelial cells and a clinical benefits. human glioma cell line [155]. As mentioned, Cx26 may facilitate extravasation into the Connexins and tumour viruses endothelium during metastasis [47], and Cx43 enhances attachment and diapedesis into endothelial cells in models An estimated 10–12% of human cancers worldwide are of breast cancer [48, 156] and melanoma metastasis [157]. caused by viruses and 4.5% are caused by the so-called Heterocellular Cx43-mediated GJIC between gastric cancer “high-risk” human papillomaviruses (HPVs) [168]. Rous cells and mesothelial cells may facilitate diapedesis during sarcoma virus-infected fibroblasts display a loss of GJIC, peritoneal metastasis [75]. Using zebrafish and chicken and its oncogene pp60v-src causes Cx43 tyrosine phosphor- embryo models of brain metastasis, Stoletov et al. [158] ylation and reduced Cx43 levels [169]. Similarly, SV40 determined that Cx26 and Cx43 acted in the early initiation large T antigen-transformed normal invasive trophoblasts stages of metastatic lesion formation in association with lose GJIC and display reduced Cx43 levels [169]. Human vasculature. RNAi-mediated depletion of Cx26 and Cx43 in cytomegalovirus is not recognised as a tumour virus. melanoma and breast cancer cells, respectively, or phar- However, human cytomegalovirus proteins have been macological inhibition of GJIC using carbenoxolone, was detected at high levels in gliomas and can downregulate found to inhibit brain colonisation by blocking tumour cell Cx43 and GJIC [170]. These data suggest that a range of extravasation and blood vessel co-option [158]. Such viruses inhibit GJIC. crosstalk between cancer cells and blood vessels may also In 1969, McNutt and Weinstein recognised that gap occur via hemichannels, such as the one through ATP junction plaques are lost in cervical cancers [171]. Subse- release, which ultimately can stimulate angiogenesis [159]. quently, cervical cancers were shown to be caused by HPVs There is also evidence that GJIC facilitates cancer cell [172]. Reduced connexin expression can occur in pre- migration through the lymphatic endothelium [160]. Thus, neoplastic cervical lesions [173], perhaps in part due to the blockade of connexin-mediated heterocellular communica- loss of epithelial differentiation. On the other hand, loss of tion is emerging as a potential viable strategy for reducing Cx43-mediated GJIC occurs as a direct consequence of metastasis. Indeed, the Cx43 channel blocker oleamide has HPV-associated cancer progression [174], and HPV infec- antimetastatic properties in the MDA-MB-231 breast cancer tions alter multiple connexins at the transcript level [175]. cell line, presumably due to the inhibition of extravasation High-risk HPVs encode two oncoproteins, E6 and E7, into the endothelium [85]. which are highly expressed in cervical cancer. E6 alters Cx43 trafficking to the plasma membrane either through its Communication between cancer cells and astrocytes ability to alter cell signalling pathways or through its interaction with the Cx43 partner protein Dlgh1 [42, 176]. Connexins can also elicit pro-tumourigenic effects in brain HPV E5, a subsidiary oncoprotein, also downregulates tumours. Cx43 expression and heterologous GJIC between Cx43 and GJIC when overexpressed in keratinocytes [177]. malignant glioma cells and reactive astrocytes enhance cell Upon infection, viral DNA is detected by cyclic GMP- invasion into the brain parenchyma [161–163], which may AMP synthase, which synthesises the STING activator in part occur through microRNA-mediated signalling [164]. cGAMP to induce an antiviral state through intercellular 4444 T. Aasen et al. transmission. Connexins are responsible for the bystander exosome release upon traumatic brain injury [189]. Thus, it immunity to viruses [178], and so viruses may have evolved seems clear that connexins are involved in the formation means to inactivate this response by regulating connexins, a and function of EVs, but the therapeutic implications await feature that could be exploited therapeutically. further works. Role of connexins in long-distance cell–cell communication Concluding remarks and future perspectives Tunnelling nanotubes Substantial knowledge on how gap junctions contribute to cancer has accumulated since the seminal work by Loe- Tunnelling nanotubes (TNTs) are thin actin-based mem- wenstein and Kanno demonstrated a loss of electrical cou- brane bridges that connect cells over distances of up to pling in liver cancer more than 50 years ago [7]. Connexins several cell diameters [179]. These structures allow for the predict the prognosis of a number of cancers, although the intercellular transfer of microRNAs, proteins and cyto- lack of established protocols and confirmatory independent plasmic organelles, including mitochondria. The presence studies has limited their clinical utility. Several promising of connexin channels interposed in the nanotube connec- studies using an expanded set of tools to modulate connexin tions permits long-distance electrical coupling between cells or gap junction function have demonstrated potent anti- [179, 180]. TNTs are formed between various cancer cell tumoral effects. As we decipher their cancer type- and types in vitro [179]. TNTs between malignant cells and stage-specific roles, significant progress towards targeting stromal cells may be involved in chemoresistance [181] and of connexins and gap junctions in a patient-specific ther- tumour–stromal crosstalk [182, 183]. Recently, Osswald apeutic setting can be expected. An ongoing clinical trial et al. [184] demonstrated in vivo that glioma cells form a testing the GJIC inhibitor meclofenamate in patients with network of TNT-like structures called microtubes that carcinoma metastasis to the brain underpins this positive contribute to their invasion in the brain. Cx43-containing outlook (https://clinicaltrials.gov/ct2/show/NCT02429570? gap junctions within this network were suggested to turn the id=NCT02429570&rank=1&load=cart). Nevertheless, tumour into a syncytium of interconnected cells that is there are significant challenges that remain to be addressed. highly resistant to radiation therapy, presumably by dis- The fact that connexins can be both anti- and pro- tributing calcium between cells to prevent apoptosis upon tumorigenic is of particular concern. For example, would radiation-induced release of intracellular calcium [184]. enhancing GJIC or connexin expression in a primary tumour run a risk of more efficient metastatic spread or Extracellular vesicles growth? Does inhibition of GJIC in patients with metastasis increase the risk of further tumour dissemination or re- Extracellular vesicles (EVs), which constitute micro- activated tumour growth in sites harbouring dormant vesicles, apoptotic bodies and exosomes, are membrane- tumour cells? Currently, this clinical problem can only be based structures that can carry and deliver bioactive mole- addressed by hypothetical risk assessment and thus further cules, including proteins and nucleic acids, from one cell to research, including an extensive use of in vivo models and a another. This form of cell-to-cell communication can occur careful follow-up of ongoing clinical trials, is required. over very long distances and efficiently cross the blood– Currently, only a few of the 21 connexin isoforms have brain barrier. EVs released by tumour cells and cancer- been characterised in terms of their role in cancer. Addi- associated fibroblasts affect cancer progression by trans- tional studies are necessary to elucidate the GJIC-dependent ferring molecules influencing tumour initiation, angiogen- and -independent mechanisms by which the various con- esis, metastasis and drug resistance [185]. Functional Cx43 nexins positively or negatively affect cell growth, differ- channels were recently identified in the membrane of EVs entiation, invasion and other important cancer-associated [186]. In melanoma, Cx32, Cx43 and Cx45 were detected features. In this context, more specific tools will be required in these structures [187]. Connexins facilitate the transfer or to target the different functions of connexins. Further efforts exchange of EV contents with target cells, possibly by need to be devoted to the identification of more specific improving fusion events with cells [186]. This may be connexin inhibitors. The use of specific peptides and pep- exploited therapeutically to improve drug delivery. How- tidomimetics have shown a great promise towards this. ever, in one mouse model, delivery of doxorubicin as a However, their side effects must be carefully addressed chemotherapeutic agent was not improved in the presence (particularly if applied systemically) due to the critical of Cx43 in EVs, although it strikingly reduced cardiotoxi- functions of several connexins in various organs. Another city [188]. In addition, Cx43 phosphorylation through important challenge will be to dissect the molecular basis of extracellular signal-regulated kinase signalling induces the regulation of the various connexin isoforms at the Connexins in cancer: bridging the gap to the clinic 4445 transcriptional, translational and post-translational levels 3. Saez JC, Leybaert L. Hunting for connexin hemichannels. FEBS and to define how the dysregulation of these processes Lett. 2014;588:1205–11. 4. Mesnil M, Aasen T, Boucher J, Chepied A, Cronier L, Defamie contributes to aberrant levels or subcellular localisation of N, et al. An update on minding the gap in cancer. Biochimica et connexins during various stages of cancer progression. The Biophysica Acta. 2018;1860:237–43. identification of solutions to these research challenges will 5. Laird DW, Lampe PD. Therapeutic strategies targeting con- set the stage for new diagnostic and therapeutic advances. nexins. Nat Rev Drug Discov. 2018;17:905–21. 6. Delvaeye T, Vandenabeele P, Bultynck G, Leybaert L, Krysko DV. Therapeutic targeting of connexin channels: new views and Acknowledgements The authors are members of The European Con- challenges. Trends Mol Med. 2018;24:1036–53. sortium for Research on Connexins and Cancer. We apologise to 7. Loewenstein WR, Kanno Y. Intercellular communication and the authors whose work could not be cited due to space limitations. Trond control of tissue growth: lack of communication between cancer Aasen acknowledges the support from Instituto de Salud Carlos III cells. Nature. 1966;209:1248–9. (grants PI16/00772 and CPII16/00042), co-financed by the European 8. Chen JT, Cheng YW, Chou MC, Sen-Lin T, Lai WW, Ho WL, Regional Development Fund (ERDF). Edward Leithe acknowledges et al. The correlation between aberrant connexin 43 mRNA the support from the Norwegian Cancer Society (grant number expression induced by promoter methylation and nodal micro- 709125), the South-Eastern Norway Regional Health Authority (grant metastasis in non-small cell lung cancer. Clin Cancer Res. number 2016013) and the Kristian Gerhard Jebsen Foundation. Arantxa 2003;9:4200–4. Tabernero acknowledges the support from the Ministerio de Economía 9. Chen Y, Huhn D, Knosel T, Pacyna-Gengelbach M, Deutsch- y Competitividad, Spain; FEDER BFU2015–70040-R, Consejería de mann N, Petersen I. Downregulation of connexin 26 in human Educación, Junta de Castilla y León, Spain; FEDER SA026U16 and lung cancer is related to promoter methylation. Int J Cancer. Fundación Ramón Areces. María D Mayán acknowledges the support 2005;113:14–21. from the Spanish Foundation for Research on Bone and Mineral 10. Sirnes S, Honne H, Ahmed D, Danielsen SA, Rognum TO, Meling Metabolism (FEIOMM), grant PRECIPITA-2015–000139 from the GI, et al. DNA methylation analyses of the connexin gene family FECYT-Ministerio de Economía y Competitividad, grant IN607B reveal silencing of GJC1 (Connexin45) by promoter hypermethy- 2017/21 from Xunta de Galicia and grant PI16/00035 from the Instituto lation in colorectal cancer. Epigenetics. 2011;6:602–9. de Salud Carlos III (ISCIII, Spain), and the European Regional 11. Tan LW, Bianco T, Dobrovic A. Variable promoter region CpG Development Fund, ‘A way of making Europe’ from the European island methylation of the putative tumor suppressor gene Con- Union. Sheila Graham acknowledges the support from Worldwide nexin 26 in breast cancer. Carcinogenesis. 2002;23:231–6. Cancer research grant 08–0159 and from the Medical Research Council 12. Jin Z, Xu S, Yu H, Yang B, Zhao H, Zhao G. miR-125b inhibits as core funding for the MRC University of Glasgow Centre for Virus Connexin43 and promotes glioma growth. Cell Mol Neurobiol. Research. Marc Mesnil acknowledges the support from Ligue contre le 2013;33:1143–8. Cancer (Comités de la Charente-Maritime et de la Vienne). 13. Li X, Pan JH, Song B, Xiong EQ, Chen ZW, Zhou ZS, et al. Suppression of CX43 expression by miR-20a in the progression Compliance with ethical standards of human prostate cancer. Cancer Biol Ther. 2012;13:890–8. 14. Smyth JW, Shaw RM. Autoregulation of connexin43 gap junc- Conflict of interest The authors declare that they have no conflict of tion formation by internally translated isoforms. Cell Rep. interest. 2013;5:611–8. 15. Salat-Canela C, Sese M, Peula C, Ramon y Cajal S, Aasen T. Internal translation of the connexin 43 transcript. Cell Commun Publisher’s note: Springer Nature remains neutral with regard to Signal. 2014;12:31. jurisdictional claims in published maps and institutional affiliations. 16. Ul-Hussain M, Olk S, Schoenebeck B, Wasielewski B, Meier C, Prochnow N, et al. Internal ribosomal entry site (IRES) activity Open Access This article is licensed under a Creative Commons generates endogenous carboxyl-terminal domains of Cx43 and is Attribution 4.0 International License, which permits use, sharing, responsive to hypoxic conditions. J Biol Chem. adaptation, distribution and reproduction in any medium or format, as 2014;289:20979–90. long as you give appropriate credit to the original author(s) and the 17. James CC, Zeitz MJ, Calhoun PJ, Lamouille S, Smyth JW. source, provide a link to the Creative Commons license, and indicate if Altered translation initiation of Gja1 limits gap junction forma- changes were made. The images or other third party material in this tion during epithelial-mesenchymal transition. Mol Biol Cell. article are included in the article’s Creative Commons license, unless 2018;29:773–880. indicated otherwise in a credit line to the material. If material is not 18. Lampe PD, Lau AF. The effects of connexin phosphorylation on included in the article’s Creative Commons license and your intended gap junctional communication. Int J Biochem Cell Biol. use is not permitted by statutory regulation or exceeds the permitted 2004;36:1171–86. use, you will need to obtain permission directly from the copyright 19. Leithe E. Regulation of connexins by the ubiquitin system: holder. To view a copy of this license, visit http://creativecommons. Implications for intercellular communication and cancer. Bio- org/licenses/by/4.0/. chim Biophys Acta. 2016;1865:133–46. 20. Eghbali B, Kessler JA, Reid LM, Roy C, Spray DC. Involvement of gap junctions in tumorigenesis: transfection of tumor cells with connexin 32 cDNA retards growth in vivo. Proc Natl Acad References Sci USA. 1991;88:10701–5. 21. Mesnil M, Krutovskikh V, Piccoli C, Elfgang C, Traub O, 1. Laird DW. Life cycle of connexins in health and disease. Bio- Willecke K, et al. Negative growth control of HeLa cells by chem J. 2006;394:527–43. connexin genes: connexin species specificity. Cancer Res. 2. Aasen T, Mesnil M, Naus CC, Lampe PD, Laird DW. Gap 1995;55:629–39. junctions and cancer: communicating for 50 years. Nat Rev 22. Qin H, Shao Q, Curtis H, Galipeau J, Belliveau DJ, Wang T, Cancer. 2016;16:775–88. et al. Retroviral delivery of connexin genes to human breast 4446 T. Aasen et al. tumor cells inhibits in vivo tumor growth by a mechanism that is 40. Shi Q, Gu S, Yu XS, White TW, Banks EA, Jiang JX. Connexin independent of significant gap junctional intercellular commu- controls cell-cycle exit and cell differentiation by directly pro- nication. J Biol Chem. 2002;277:29132–8. moting cytosolic localization and degradation of E3 ligase Skp2. 23. Zhang YW, Morita I, Ikeda M, Ma KW, Murota S. Dev Cell. 2015;35:483–96. Connexin43 suppresses proliferation of osteosarcoma U2OS 41. Duffy HS, Iacobas I, Hotchkiss K, Hirst-Jensen BJ, Bosco A, cells through post-transcriptional regulation of p27. Oncogene. Dandachi N, et al. The gap junction protein connexin32 interacts 2001;20:4138–49. with the Src homology 3/hook domain of discs large homolog 1. 24. Zhang YW, Kaneda M, Morita I. The gap junction-independent J Biol Chem. 2007;282:9789–96. tumor-suppressing effect of connexin 43. J Biol Chem. 42. Macdonald AI, Sun P, Hernandez-Lopez H, Aasen T, Hodgins 2003;278:44852–6. MB, Edward M, et al. A functional interaction between the 25. Zhu D, Kidder GM, Caveney S, Naus CC. Growth retardation in MAGUK protein hDlg and the gap junction protein connexin 43 glioma cells cocultured with cells overexpressing a gap junction in cervical tumour cells. Biochem J. 2012;446:9–21. protein. Proc Natl Acad Sci USA. 1992;89:10218–21. 43. Gonzalez-Sanchez A, Jaraiz-Rodriguez M, Dominguez-Prieto M, 26. Hellmann P, Grummer R, Schirrmacher K, Rook M, Traub O, Herrero-Gonzalez S, Medina JM, Tabernero A. Connexin43 Winterhager E. Transfection with different connexin genes alters recruits PTEN and Csk to inhibit c-Src activity in glioma cells growth and differentiation of human choriocarcinoma cells. Exp and astrocytes. Oncotarget. 2016;7:49819–33. Cell Res. 1999;246:480–90. 44. Kotini M, Mayor R. Connexins in migration during development 27. Hirschi KK, Xu CE, Tsukamoto T, Sager R. Gap junction genes and cancer. Dev Biol. 2015;401:143–51. Cx26 and Cx43 individually suppress the cancer phenotype of 45. Defamie N, Chepied A. Mesnil M. Connexins, gap junctions and human mammary carcinoma cells and restore differentiation tissue invasion. FEBS Lett. 2014;588:1331–8. potential. Cell growth & differentiation: the molecular biology 46. el-Sabban ME, Pauli BU. Cytoplasmic dye transfer between journal of the American Association for. Cancer Res. metastatic tumor cells and vascular endothelium. J Cell Biol. 1996;7:861–70. 1991;115:1375–82. 28. McLachlan E, Shao Q, Wang HL, Langlois S, Laird DW. 47. Ito A, Katoh F, Kataoka TR, Okada M, Tsubota N, Asada H, Connexins act as tumor suppressors in three-dimensional mam- et al. A role for heterologous gap junctions between melanoma mary cell organoids by regulating differentiation and angiogen- and endothelial cells in metastasis. J Clin Invest. esis. Cancer Res. 2006;66:9886–94. 2000;105:1189–97. 29. Shao Q, Wang H, McLachlan E, Veitch GI, Laird DW. Down- 48. Pollmann MA, Shao Q, Laird DW, Sandig M. Connexin 43 regulation of Cx43 by retroviral delivery of small interfering mediated gap junctional communication enhances breast tumor RNA promotes an aggressive breast cancer cell phenotype. cell diapedesis in culture. Breast Cancer Res. 2005;7:R522–534. Cancer Res. 2005;65:2705–11. 49. Saito-Katsuragi M, Asada H, Niizeki H, Katoh F, Masuzawa M, 30. Aasen T. Connexins: junctional and non-junctional modulators Tsutsumi M, et al. Role for connexin 26 in metastasis of human of proliferation. Cell Tissue Res. 2015;360:685–99. malignant melanoma: communication between melanoma and 31. Gilleron J, Carette D, Segretain D, Pointis G. Multiple and endothelial cells via connexin 26. Cancer. 2007;110:1162–72. complex influences of connexins and pannexins on cell death. 50. Chen Q, Boire A, Jin X, Valiente M, Er EE, Lopez-Soto A, et al. Biochim Biophys Acta. 2018;1860:182–91. Carcinoma-astrocyte gap junctions promote brain metastasis by 32. Kameritsch P, Khandoga N, Pohl U, Pogoda K. Gap junctional cGAMP transfer. Nature. 2016;533:493–8. communication promotes apoptosis in a connexin-type- 51. Artesi M, Kroonen J, Bredel M, Nguyen-Khac M, Deprez M, dependent manner. Cell Death Dis. 2013;4:e584. Schoysman L, et al. Connexin 30 expression inhibits growth of 33. Krutovskikh VA, Piccoli C, Yamasaki H. Gap junction inter- human malignant gliomas but protects them against radiation cellular communication propagates cell death in cancerous cells. therapy. Neuro Oncol. 2015;17:392–406. Oncogene. 2002;21:1989–99. 52. Munoz JL, Rodriguez-Cruz V, Greco SJ, Ramkissoon SH, Ligon 34. Lin Q, Balasubramanian K, Fan D, Kim SJ, Guo L, Wang H, KL, Rameshwar P. Temozolomide resistance in glioblastoma et al. Reactive astrocytes protect melanoma cells from che- cells occurs partly through epidermal growth factor receptor- motherapy by sequestering intracellular calcium through gap mediated induction of connexin 43. Cell Death Dis. 2014;5: junction communication channels. Neoplasia. 2010;12:748–54. e1145. 35. Wang N, De Bock M, Decrock E, Bol M, Gadicherla A, Vinken 53. Osswald M, Jung E, Sahm F, Solecki G, Venkataramani V, Blaes M, et al. Paracrine signaling through plasma membrane hemi- J, et al. Brain tumour cells interconnect to a functional and channels. Biochim Biophys Acta. 2013;1828:35–50. resistant network. Nature. 2015;528:93–98. 36. Krutovskikh VA, Troyanovsky SM, Piccoli C, Tsuda H, Asa- 54. Yu SC, Xiao HL, Jiang XF, Wang QL, Li Y, Yang XJ, et al. moto M, Yamasaki H. Differential effect of subcellular locali- Connexin 43 reverses malignant phenotypes of glioma stem cells zation of communication impairing gap junction protein by modulating E-cadherin. Stem Cells. 2012;30:108–20. connexin43 on tumor cell growth in vivo. Oncogene. 55. Gangoso E, Thirant C, Chneiweiss H, Medina JM, Tabernero A. 2000;19:505–13. A cell-penetrating peptide based on the interaction between c-Src 37. Olbina G, Eckhart W. Mutations in the second extracellular and connexin43 reverses glioma stem cell phenotype. Cell Death region of connexin 43 prevent localization to the plasma mem- Dis. 2014;5:e1023. brane, but do not affect its ability to suppress cell growth. Mol 56. Hitomi M, Deleyrolle LP, Mulkearns-Hubert EE, Jarrar A, Li M, Cancer Res. 2003;1:690–700. Sinyuk M, et al. Differential connexin function enhances self- 38. Sirnes S, Bruun J, Kolberg M, Kjenseth A, Lind GE, Svindland renewal in glioblastoma. Cell Rep. 2015;11:1031–42. A, et al. Connexin43 acts as a colorectal cancer tumor suppressor 57. Shen Y, Li Y, Ma X, Wan Q, Jiang Z, Liu Y, et al. Connexin 43 and predicts disease outcome. Int J Cancer. 2012;131:570–81. SUMOylation improves gap junction functions between liver 39. Zhang YW, Nakayama K, Nakayama K, Morita I. A novel route cancer stem cells and enhances their sensitivity to HSVtk/GCV. for connexin 43 to inhibit cell proliferation: negative regulation Int J Oncol. 2018;52:872–80. of S-phase kinase-associated protein (Skp 2). Cancer Res. 58. Kawasaki Y, Omori Y, Li Q, Nishikawa Y, Yoshioka T, Yoshida 2003;63:1623–30. M, et al. Cytoplasmic accumulation of connexin32 expands Connexins in cancer: bridging the gap to the clinic 4447 cancer stem cell population in human HuH7 hepatoma cells by enhances gastric cancer cell diapedesis from peritoneal meso- enhancing its self-renewal. Int J Cancer. 2011;128:51–62. thelium. PLoS One. 2013;8:e74527. 59. Park SY, Lee HE, Li H, Shipitsin M, Gelman R, Polyak K. 76. Kyo N, Yamamoto H, Takeda Y, Ezumi K, Ngan CY, Terayama Heterogeneity for stem cell-related markers according to tumor M, et al. Overexpression of connexin 26 in carcinoma of the subtype and histologic stage in breast cancer. Clin Cancer Res. pancreas. Oncol Rep. 2008;19:627–31. 2010;16:876–87. 77. Ezumi K, Yamamoto H, Murata K, Higashiyama M, Damdin- 60. Shipitsin M, Campbell LL, Argani P, Weremowicz S, suren B, Nakamura Y, et al. Aberrant expression of connexin 26 Bloushtain-Qimron N, Yao J, et al. Molecular definition of breast is associated with lung metastasis of colorectal cancer. Clinical tumor heterogeneity. Cancer Cell. 2007;11:259–73. cancer research: an official journal of the American Association 61. Busby M, Hallett MT, Plante I. The complex subtype-dependent for. Cancer Res. 2008;14:677–84. role of connexin 43 (GJA1) in breast cancer. Int J Mol Sci. 78. Mehta PP, Perez-Stable C, Nadji M, Mian M, Asotra K, Roos 2018;19:693. BA. Suppression of human prostate cancer cell growth by forced 62. Thiagarajan PS, Sinyuk M, Turaga SM, Mulkearns-Hubert EE, expression of connexin genes. Dev Genet. 1999;24:91–110. Hale JS, Rao V, et al. Cx26 drives self-renewal in triple-negative 79. Crespin S, Fromont G, Wager M, Levillain P, Cronier L, Mon- breast cancer via interaction with NANOG and focal adhesion voisin A, et al. Expression of a gap junction protein, connexin43, kinase. Nat Commun. 2018;9:578. in a large panel of human gliomas: new insights. Cancer Med. 63. Uhlen M, Zhang C, Lee S, Sjostedt E, Fagerberg L, Bidkhori G 2016;5:1742–52. et al. A pathology atlas of the human cancer transcriptome. 80. Kotini M, Barriga EH, Leslie J, Gentzel M, Rauschenberger V, Science. 2017;357:pii: eaan2507. Schambon A, et al. Gap junction protein Connexin-43 is a direct 64. Teleki I, Szasz AM, Maros ME, Gyorffy B, Kulka J, Meggye- transcriptional regulator of N-cadherin in vivo. Nat Commun. shazi N, et al. Correlations of differentially expressed gap 2018;9:3846. junction connexins Cx26, Cx30, Cx32, Cx43 and Cx46 with 81. Pogoda K, Kameritsch P, Retamal MA, Vega JL. Regulation of breast cancer progression and prognosis. PLoS One. 2014;9: gap junction channels and hemichannels by phosphorylation and e112541. redox changes: a revision. BMC Cell Biol. 2016;17(Suppl 1):11. 65. Murphy SF, Varghese RT, Lamouille S, Guo S, Pridham KJ, 82. King TJ, Bertram JS. Connexins as targets for cancer chemo- Kanabur P, et al. Connexin 43 inhibition sensitizes chemoresis- prevention and chemotherapy. Biochim Biophys Acta. tant glioblastoma cells to temozolomide. Cancer Res. 2005;1719:146–60. 2016;76:139–49. 83. Yu ML, Zhang CL, Yuan DD, Tong XH, Tao L. Panax noto- 66. Haass NK, Ripperger D, Wladykowski E, Dawson P, Gimotty ginseng saponins enhances the cytotoxicity of cisplatin via PA, Blome C, et al. Melanoma progression exhibits a significant increasing gap junction intercellular communication. Biol Pharm impact on connexin expression patterns in the epidermal tumor Bull. 2012;35:1230–7. microenvironment. Histochem Cell Biol. 2010;133:113–24. 84. Willebrords J, Maes M, Crespo Yanguas S, Vinken M. Inhibitors 67. Lin YP, Wu JI, Tseng CW, Chen HJ, Wang LH. Gjb4 serves as a of connexin and pannexin channels as potential therapeutics. novel biomarker for lung cancer and promotes metastasis and Pharmacol Ther. 2017;180:144–60. chemoresistance via Src activation. Oncogene. 2018;38:822–37. 85. Zibara K, Awada Z, Dib L, El-Saghir J, Al-Ghadban S, Ibrik A, 68. Zhu T, Gao YF, Chen YX, Wang ZB, Yin JY, Mao XY, et al. et al. Anti-angiogenesis therapy and gap junction inhibition Genome-scale analysis identifies GJB2 and ERO1LB as prog- reduce MDA-MB-231 breast cancer cell invasion and metastasis nosis markers in patients with pancreatic cancer. Oncotarget. in vitro and in vivo. Sci Rep. 2015;5:12598. 2017;8:21281–9. 86. Wang H, Tian L, Liu J, Goldstein A, Bado I, Zhang W, et al. The 69. Castellana B, Escuin D, Peiro G, Garcia-Valdecasas B, Vazquez osteogenic niche is a calcium reservoir of bone micrometastases T, Pons C, et al. ASPN and GJB2 are implicated in the and confers unexpected therapeutic vulnerability. Cancer Cell. mechanisms of invasion of ductal breast carcinomas. J Cancer. 2018;34:823–39 e827. 2012;3:175–83. 87. Alonso F, Domingos-Pereira S, Le Gal L, Derre L, Meda P, 70. Dubina MV, Iatckii NA, Popov DE, Vasil’ev SV, Krutovskikh Jichlinski P, et al. Targeting endothelial connexin40 inhibits VA. Connexin 43, but not connexin 32, is mutated at advanced tumor growth by reducing angiogenesis and improving vessel stages of human sporadic colon cancer. Oncogene. perfusion. Oncotarget. 2016;7:14015–28. 2002;21:4992–6. 88. Yusubalieva GM, Baklaushev VP, Gurina OI, Gulyaev MV, 71. Jing YM, Guo SX, Zhang XP, Sun AJ, Tao F, Qian HX. Pirogov YA, Chekhonin VP. Antitumor effects of monoclonal Association between C1019T polymorphism in the connexin 37 antibodies to connexin 43 extracellular fragment in induced low- gene and Helicobacter pylori infection in patients with gastric differentiated glioma. Bull Exp Biol Med. 2012;153:163–9. cancer. Asian Pac J Cancer Prev. 2012;13:2363–7. 89. Yusubalieva GM, Baklaushev VP, Gurina OI, Zorkina YA, 72. Naoi Y, Miyoshi Y, Taguchi T, Kim SJ, Arai T, Maruyama N, Gubskii IL, Kobyakov GL, et al. Treatment of poorly differ- et al. Connexin26 expression is associated with aggressive entiated glioma using a combination of monoclonal antibodies to phenotype in human papillary and follicular thyroid cancers. extracellular connexin-43 fragment, temozolomide, and radio- Cancer Lett. 2008;262:248–56. therapy. Bull Exp Biol Med. 2014;157:510–5. 73. Harris LD, De La Cerda J, Tuziak T, Rosen D, Xiao L, Shen Y, 90. Chekhonin VP, Baklaushev VP, Yusubalieva GM, Belorusova et al. Analysis of the expression of biomarkers in urinary bladder AE, Gulyaev MV, Tsitrin EB, et al. Targeted delivery of lipo- cancer using a tissue microarray. Mol Carcinogen. 2008;47:678–85. somal nanocontainers to the peritumoral zone of glioma by 74. Sentani K, Oue N, Sakamoto N, Anami K, Naito Y, Aoyagi K, means of monoclonal antibodies against GFAP and the extra- et al. Upregulation of connexin 30 in intestinal phenotype gastric cellular loop of Cx43. Nanomedicine. 2012;8:63–70. cancer and its reduction during tumor progression. Pathobiology. 91. Nukolova NV, Baklaushev VP, Abakumova TO, Mel’nikov PA, 2010;77:241–8. Abakumov MA, Yusubalieva GM, et al. Targeted delivery of 75. Tang B, Peng ZH, Yu PW, Yu G, Qian F, Zeng DZ, et al. cisplatin by small es, Cyrilliconnexin 43 vector nanogels to the Aberrant expression of Cx43 is associated with the peritoneal focus of experimental glioma C6. Bull Exp Biol Med. metastasis of gastric cancer and Cx43-mediated gap junction 2014;157:524–9. 4448 T. Aasen et al. 92. Bennett MV, Contreras JE, Bukauskas FF, Saez JC. New roles 112. Mesnil M, Piccoli C, Tiraby G, Willecke K, Yamasaki H. for astrocytes: gap junction hemichannels have something to Bystander killing of cancer cells by herpes simplex virus thy- communicate. Trends Neurosci. 2003;26:610–7. midine kinase gene is mediated by connexins. Proc Natl Acad 93. Chi Y, Gao K, Li K, Nakajima S, Kira S, Takeda M, et al. Sci USA. 1996;93:1831–5. Purinergic control of AMPK activation by ATP released through 113. Dilber MS, Abedi MR, Christensson B, Bjorkstrand B, Kidder connexin 43 hemichannels - pivotal roles in hemichannel- GM, Naus CC, et al. Gap junctions promote the bystander effect mediated cell injury. J Cell Sci. 2014;127:1487–99. of herpes simplex virus thymidine kinase in vivo. Cancer Res. 94. Riquelme MA, Kar R, Gu S, Jiang JX. Antibodies targeting 1997;57:1523–8. extracellular domain of connexins for studies of hemichannels. 114. Vrionis FD, Wu JK, Qi P, Waltzman M, Cherington V, Spray Neuropharmacology. 2013;75:525–32. DC. The bystander effect exerted by tumor cells expressing the 95. Zhou JZ, Riquelme MA, Gu S, Kar R, Gao X, Sun L, et al. herpes simplex virus thymidine kinase (HSVtk) gene is depen- Osteocytic connexin hemichannels suppress breast cancer dent on connexin expression and cell communication via gap growth and bone metastasis. Oncogene. 2016;35:5597–607. junctions. Gene Ther. 1997;4:577–85. 96. Zhou JZ, Riquelme MA, Gao X, Ellies LG, Sun LZ, Jiang JX. 115. Touraine RL, Ishii-Morita H, Ramsey WJ, Blaese RM. The Differential impact of adenosine nucleotides released by osteo- bystander effect in the HSVtk/ganciclovir system and its rela- cytes on breast cancer growth and bone metastasis. Oncogene. tionship to gap junctional communication. Gene Ther. 2015;34:1831–42. 1998;5:1705–11. 97. Leithe E, Mesnil M, Aasen T. The connexin 43 C-terminus: a tail 116. Duflot-Dancer A, Piccoli C, Rolland A, Yamasaki H, Mesnil M. of many tales. Biochim Biophys Acta. 2018;1860:48–64. Long-term connexin-mediated bystander effect in highly 98. Tabernero A, Gangoso E, Jaraiz-Rodriguez M, Medina JM. The tumorigenic human cells in vivo in herpes simplex virus thymi- role of connexin43-Src interaction in astrocytomas: A molecular dine kinase/ganciclovir gene therapy. Gene Ther. 1998;5:1372–8. puzzle. Neuroscience. 2016;323:183–94. 117. Mesnil M, Yamasaki H. Bystander effect in herpes simplex 99. Herrero-Gonzalez S, Gangoso E, Giaume C, Naus CC, Medina virus-thymidine kinase/ganciclovir cancer gene therapy: role of JM, Tabernero A. Connexin43 inhibits the oncogenic activity of gap-junctional intercellular communication. Cancer Res. c-Src in C6 glioma cells. Oncogene. 2010;29:5712–23. 2000;60:3989–99. 100. Jaraiz-Rodriguez M, Tabernero MD, Gonzalez-Tablas M, Otero 118. Kong H, Liu X, Yang L, Qi K, Zhang H, Zhang J, et al. All-trans A, Orfao A, Medina JM, et al. A short region of connexin43 retinoic acid enhances bystander effect of suicide gene therapy in reduces human glioma stem cell migration, invasion, and sur- the treatment of breast cancer. Oncol Rep. 2016;35:1868–74. vival through Src, PTEN, and FAK. Stem Cell Rep. 2017;9:451– 119. Dahle J, Mikalsen S, Rivedal E, Steen H. Gap junction inter- 63. cellular communication is not a major mediator in the bystander 101. Kameritsch P, Kiemer F, Beck H, Pohl U, Pogoda K. Cx43 effect in photodynamic treatment of MDCKII cells. Radiat Res. increases serum induced filopodia formation via activation of 2000;154:10. p21-activated protein kinase 1. Biochim Biophys Acta. 120. Prise KM, O’Sullivan JM. Radiation-induced bystander signal- 2015;1853:2907–17. ling in cancer therapy. Nat Rev Cancer. 2009;9:351. 102. Dai P, Nakagami T, Tanaka H, Hitomi T, Takamatsu T. Cx43 121. Wygoda MR, Wilson MR, Davis MA, Trosko JE, Rehemtulla A, mediates TGF-beta signaling through competitive Smads binding Lawrence TS. Protection of herpes simplex virus thymidine to microtubules. Mol Biol Cell. 2007;18:2264–73. kinase-transduced cells from ganciclovir-mediated cytotoxicity 103. Hunter AW, Barker RJ, Zhu C, Gourdie RG. Zonula occludens-1 by bystander cells: the Good Samaritan effect. Cancer Res. alters connexin43 gap junction size and organization by influ- 1997;57:1699–703. encing channel accretion. Mol Biol Cell. 2005;16:5686–98. 122. Saez CG, Velasquez L, Montoya M, Eugenin E, Alvarez MG. 104. Rhett JM, Jourdan J, Gourdie RG. Connexin 43 connexon to gap Increased gap junctional intercellular communication is directly junction transition is regulated by zonula occludens-1. Mol Biol related to the anti-tumor effect of all-trans-retinoic acid plus Cell. 2011;22:1516–28. tamoxifen in a human mammary cancer cell line. J Cell Biochem. 105. Grek CL, Rhett JM, Bruce JS, Abt MA, Ghatnekar GS, Yeh ES. 2003;89:450–61. Targeting connexin 43 with alpha-connexin carboxyl-terminal 123. Carystinos GD, Alaoui-Jamali MA, Phipps J, Yen L, Batist G. (ACT1) peptide enhances the activity of the targeted inhibitors, Upregulation of gap junctional intercellular communication and tamoxifen and lapatinib, in breast cancer: clinical implication for connexin 43 expression by cyclic-AMP and all-trans-retinoic ACT1. BMC Cancer. 2015;15:296. acid is associated with glutathione depletion and chemosensi- 106. Naus CC, Giaume C. Bridging the gap to therapeutic strategies tivity in neuroblastoma cells. Cancer Chemother Pharmacol. based on connexin/pannexin biology. J Transl Med. 2001;47:126–32. 2016;14:330. 124. Wang Y, Zhang C, Zhang S, Zhao Z, Wang J, Song J, et al. 107. Seymour CB, Mothersill C. Radiation-induced bystander effects Kanglaite sensitizes colorectal cancer cells to Taxol via NF- —implications for cancer. Nat Rev Cancer. 2004;4:158. kappaBeta inhibition and connexin 43 upregulation. Sci Rep. 108. Greco R, Oliveira G, Stanghellini MTL, Vago L, Bondanza A, 2017;7:1280. Peccatori J. Improving the safety of cell therapy with the TK- 125. Wu JF, Ji J, Dong SY, Li BB, Yu ML, Wu DD, et al. Gefitinib suicide gene. Front Pharmacol. 2015;6:95. enhances oxaliplatin-induced apoptosis mediated by Src and 109. Elshami AA, Saavedra A, Zhang H, Kucharczuk JC, Spray DC, PKC-modulated gap junction function. Oncol Rep. Fishman GI, et al. Gap junctions play a role in the ‘bystander 2016;36:3251–8. effect’ of the herpes simplex virus thymidine kinase/ganciclovir 126. Wang L, Fu Y, Peng J, Wu D, Yu M, Xu C, et al. Simvastatin- system in vitro. Gene Ther. 1996;3:85–92. induced up-regulation of gap junctions composed of connexin 110. Fick J, Barker FG 2nd, Dazin P, Westphale EM, Beyer EC, Israel 43 sensitize Leydig tumor cells to etoposide: an involvement of MA. The extent of heterocellular communication mediated by PKC pathway. Toxicology. 2013;312:149–57. gap junctions is predictive of bystander tumor cytotoxicity 127. Wang L, Peng J, Huang H, Wang Q, Yu M, Tao L. Simvastatin in vitro. Proc Natl Acad Sci USA. 1995;92:11071–5. protects Sertoli cells against cisplatin cytotoxicity through 111. Pitts JD. Cancer gene therapy: a bystander effect using the gap enhanced gap junction intercellular communication. Oncol Rep. junctional pathway. Mol Carcinog. 1994;11:127–30. 2015;34:2133–41. Connexins in cancer: bridging the gap to the clinic 4449 128. Liu D, Zhou H, Wu J, Liu W, Li Y, Shi G, et al. Infection by 147. Glass AM, Snyder EG, Taffet SM. Connexins and pannexins in Cx43 adenovirus increased chemotherapy sensitivity in human the immune system and lymphatic organs. Cell Mol Life Sci. gastric cancer BGC-823 cells: not involving in induction of cell 2015;72:2899–910. apoptosis. Gene. 2015;574:217–24. 148. Potente M, Gerhardt H, Carmeliet P. Basic and therapeutic 129. Huang RP, Hossain MZ, Huang R, Gano J, Fan Y, Boynton AL. aspects of angiogenesis. Cell. 2011;146:873–87. Connexin 43 (cx43) enhances chemotherapy-induced apoptosis 149. Wang WK, Chen MC, Leong HF, Kuo YL, Kuo CY, Lee CH. in human glioblastoma cells. Int J Cancer. 2001;92:130–8. Connexin 43 suppresses tumor angiogenesis by down-regulation 130. Raza A, Ghoshal A, Chockalingam S, Ghosh SS. Connexin-43 of vascular endothelial growth factor via hypoxic-induced factor- enhances tumor suppressing activity of artesunate via gap 1alpha. Int J Mol Sci. 2014;16:439–51. junction-dependent as well as independent pathways in human 150. Gartner C, Ziegelhoffer B, Kostelka M, Stepan H, Mohr FW, breast cancer cells. Sci Rep. 2017;7:7580. Dhein S. Knock-down of endothelial connexins impairs angio- 131. Sato H, Senba H, Virgona N, Fukumoto K, Ishida T, Hagiwara H, genesis. Pharmacol Res. 2012;65:347–57. et al. Connexin 32 potentiates vinblastine-induced cytotoxicity in 151. Haefliger JA, Allagnat F, Hamard L, Le Gal L, Meda P, Nardelli- renal cell carcinoma cells. Mol Carcinog. 2007;46:215–24. Haefliger D, et al. Targeting Cx40 (Connexin40) expression or 132. Tanaka M, Grossman HB. Connexin 26 gene therapy of human function reduces angiogenesis in the developing mouse retina. bladder cancer: induction of growth suppression, apoptosis, and Arterioscler Thromb Vasc Biol. 2017;37:2136–46. synergy with Cisplatin. Hum Gene Ther. 2001;12:2225–36. 152. de Wit C, Wolfle SE, Hopfl B. Connexin-dependent commu- 133. Tanaka M, Grossman HB. Connexin 26 induces growth sup- nication within the vascular wall: contribution to the control of pression, apoptosis and increased efficacy of doxorubicin in arteriolar diameter. Adv Cardiol. 2006;42:268–83. prostate cancer cells. Oncol Rep. 2004;11:537–41. 153. Segal SS. Integration and modulation of intercellular signaling 134. Chen W, Wang D, Du X, He Y, Chen S, Shao Q, et al. Glioma underlying blood flow control. J Vasc Res. 2015;52:136–57. cells escaped from cytotoxicity of temozolomide and vincristine by 154. Thuringer D, Jego G, Berthenet K, Hammann A, Solary E, communicating with human astrocytes. Med Oncol. 2015;32:43. Garrido C. Gap junction-mediated transfer of miR-145-5p from 135. Gielen PR, Aftab Q, Ma N, Chen VC, Hong X, Lozinsky S, et al. microvascular endothelial cells to colon cancer cells inhibits Connexin43 confers Temozolomide resistance in human glioma angiogenesis. Oncotarget. 2016;7:28160–8. cells by modulating the mitochondrial apoptosis pathway. Neu- 155. Thuringer D, Boucher J, Jego G, Pernet N, Cronier L, Hammann ropharmacology. 2013;75:539–48. A, et al. Transfer of functional microRNAs between glio- 136. Yulyana Y, Endaya BB, Ng WH, Guo CM, Hui KM, Lam PY, blastoma and microvascular endothelial cells through gap junc- et al. Carbenoxolone enhances TRAIL-induced apoptosis tions. Oncotarget. 2016;7:73925–34. through the upregulation of death receptor 5 and inhibition of 156. Elzarrad MK, Haroon A, Willecke K, Dobrowolski R, Gillespie gap junction intercellular communication in human glioma. Stem MN, Al-Mehdi AB. Connexin-43 upregulation in micro- Cells Dev. 2013;22:1870–82. metastases and tumor vasculature and its role in tumor cell 137. Yang J, Qin G, Luo M, Chen J, Zhang Q, Li L, et al. Reciprocal attachment to pulmonary endothelium. BMC Med. 2008;6:20. positive regulation between Cx26 and PI3K/Akt pathway confers 157. Villares GJ, Dobroff AS, Wang H, Zigler M, Melnikova VO, acquired gefitinib resistance in NSCLC cells via GJIC- Huang L, et al. Overexpression of protease-activated receptor-1 independent induction of EMT. Cell Death Dis. 2015;6:e1829. contributes to melanoma metastasis via regulation of connexin 138. Mulkearns-Hubert EE, Torre-Healy LA, Silver DJ, Eurich JT, 43. Cancer Res. 2009;69:6730–7. Serbinowski E, Hitomi M et al. Development of a Cx46 targeting 158. Stoletov K, Strnadel J, Zardouzian E, Momiyama M, Park FD, strategy for cancer stem cells. bioRxivorg. 2018. Kelber JA, et al. Role of connexins in metastatic breast cancer 139. Falk S. Carbenoxolone as a novel therapy for attenuation of and melanoma brain colonization. J Cell Sci. 2013;126:904–13. cancer-induced bone pain. Pain. 2018;159:1127–36. 159. Thuringer D, Berthenet K, Cronier L, Solary E, Garrido C. Pri- 140. Mehta PP, Bertram JS, Loewenstein WR. Growth inhibition of mary tumor- and metastasis-derived colon cancer cells differ- transformed cells correlates with their junctional communication ently modulate connexin expression and function in human with normal cells. Cell. 1986;44:187–96. capillary endothelial cells. Oncotarget. 2015;6:28800–15. 141. Gleisner MA, Navarrete M, Hofmann F, Salazar-Onfray F, Tit- 160. Karpinich NO, Caron KM. Gap junction coupling is required for tarelli A. Mind the gaps in tumor immunity: impact of connexin- tumor cell migration through lymphatic endothelium. Arter- mediated intercellular connections. Front Immunol. 2017;8:1067. ioscler Thromb Vasc Biol. 2015;35:1147–55. 142. Mendoza-Naranjo A, Saez PJ, Johansson CC, Ramirez M, 161. Lin JH, Takano T, Cotrina ML, Arcuino G, Kang J, Liu S, et al. Mandakovic D, Pereda C, et al. Functional gap junctions facil- Connexin 43 enhances the adhesivity and mediates the invasion itate melanoma antigen transfer and cross-presentation between of malignant glioma cells. J Neurosci. 2002;22:4302–11. human dendritic cells. J Immunol. 2007;178:6949–57. 162. Sin WC, Aftab Q, Bechberger JF, Leung JH, Chen H, Naus CC. 143. Saccheri F, Pozzi C, Avogadri F, Barozzi S, Faretta M, Fusi P, Astrocytes promote glioma invasion via the gap junction protein et al. Bacteria-induced gap junctions in tumors favor antigen connexin43. Oncogene. 2016;35:1504–16. cross-presentation and antitumor immunity. Sci Transl Med. 163. Oliveira R, Christov C, Guillamo JS, de Bouard S, Palfi S, 2010;2:44ra57. Venance L, et al. Contribution of gap junctional communication 144. Tittarelli A, Janji B, Van Moer K, Noman MZ, Chouaib S. The between tumor cells and astroglia to the invasion of the brain selective degradation of synaptic connexin 43 protein by parenchyma by human glioblastomas. BMC Cell Biol. 2005;6:7. hypoxia-induced autophagy impairs natural killer cell-mediated 164. Hong X, Sin WC, Harris AL, Naus CC. Gap junctions modulate tumor cell killing. J Biol Chem. 2015;290:23670–9. glioma invasion by direct transfer of microRNA. Oncotarget. 145. Tittarelli A, Mendoza-Naranjo A, Farias M, Guerrero I, Ihara F, 2015;6:15566–77. Wennerberg E, et al. Gap junction intercellular communications 165. Lin Q, Liu Z, Ling F, Xu G. Astrocytes protect glioma cells from regulate NK cell activation and modulate NK cytotoxic capacity. chemotherapy and upregulate survival genes via gap junctional J Immunol. 2014;192:1313–9. communication. Mol Med Rep. 2016;13:1329–35. 146. Aucher A, Rudnicka D, Davis DM. MicroRNAs transfer from 166. Lai SW, Huang BR, Liu YS, Lin HY, Chen CC, Tsai CF. et al. human macrophages to hepato-carcinoma cells and inhibit pro- Differential characterization of temozolomide-resistant human liferation. J Immunol. 2013;191:6250–60. glioma cells. Int J Mol Sci. 2018;19:127. 4450 T. Aasen et al. 167. Menachem A, Makovski V, Bodner O, Pasmanik-Chor M, Stein 185. Melo SA, Sugimoto H, O’Connell JT, Kato N, Villanueva A, R, Shomron N, et al. Intercellular transfer of small RNAs from Vidal A, et al. Cancer exosomes perform cell-independent astrocytes to lung tumor cells induces resistance to chemother- microRNA biogenesis and promote tumorigenesis. Cancer Cell. apy. Oncotarget. 2016;7:12489–504. 2014;26:707–21. 168. de Martel C, Plummer M, Vignat J, Franceschi S. Worldwide 186. Soares AR, Martins-Marques T, Ribeiro-Rodrigues T, Ferreira burden of cancer attributable to HPV by site, country and HPV JV, Catarino S, Pinho MJ, et al. Gap junctional protein Cx43 is type. Int J Cancer. 2017;141:664–70. involved in the communication between extracellular vesicles 169. Khoo NK, Zhang Y, Bechberger JF, Bond SL, Hum K, Lala PK. and mammalian cells. Sci Rep. 2015;5:13243. SV40 Tag transformation of the normal invasive trophoblast 187. Lazar I, Clement E, Ducoux-Petit M, Denat L, Soldan V, Dau- results in a premalignant phenotype. II. Changes in gap junc- villier S, et al. Proteome characterization of melanoma exosomes tional intercellular communication. Int J Cancer. 1998;77:440–8. reveals a specific signature for metastatic cell lines. Pigment Cell 170. Khan Z, Yaiw KC, Wilhelmi V, Lam H, Rahbar A, Stragliotto G, Melanoma Res. 2015;28:464–75. et al. Human cytomegalovirus immediate early proteins promote 188. Martins-Marques T, Pinho MJ, Zuzarte M, Oliveira C, Pereira P, degradation of connexin 43 and disrupt gap junction commu- Sluijter JP, et al. Presence of Cx43 in extracellular vesicles nication: implications for a role in gliomagenesis. Carcinogen- reduces the cardiotoxicity of the anti-tumour therapeutic esis. 2014;35:145–54. approach with doxorubicin. J Extracell Vesicles. 2016;5:32538. 171. McNutt NS, Weinstein RS. Carcinoma of the cervix: deficiency 189. Soung YH, Ford S, Zhang V, Chung J. Exosomes in cancer of nexus intercellular junctions. Science. 1969;165:597–9. diagnostics. Cancers. 2017;9:8. 172. zur Hausen H. Papillomaviruses in the causation of human 190. Poyet C, Buser L, Roudnicky F, Detmar M, Hermanns T, cancers - a brief historical account. Virology. 2009;384:260–5. Mannhard D, et al. Connexin 43 expression predicts poor 173. Aasen T, Graham SV, Edward M, Hodgins MB. Reduced progression-free survival in patients with non-muscle invasive expression of multiple gap junction proteins is a feature of cer- urothelial bladder cancer. J Clin Pathol. 2015;68:819–24. vical dysplasia. Mol Cancer. 2005;4:31. 191. Balla P, Maros ME, Barna G, Antal I, Papp G, Sapi Z, et al. 174. Aasen T, Hodgins MB, Edward M, Graham SV. The relationship Prognostic impact of reduced connexin43 expression and gap between connexins, gap junctions, tissue architecture and tumour junction coupling of neoplastic stromal cells in giant cell tumor invasion, as studied in a novel in vitro model of HPV-16- of bone. PLoS One. 2015;10:e0125316. associated cervical cancer progression. Oncogene. 192. Naoi Y, Miyoshi Y, Taguchi T, Kim SJ, Arai T, Tamaki Y, et al. 2003;22:7969–80. Connexin26 expression is associated with lymphatic vessel 175. Klymenko T, Gu Q, Herbert I, Stevenson A, Iliev V, Watkins G. invasion and poor prognosis in human breast cancer. Breast et al. RNASeq analysis of differentiated keratinocytes reveals a Cancer Res Treat. 2007;106:11–17. massive response to late events during human papillomavirus 193. Teleki I, Krenacs T, Szasz MA, Kulka J, Wichmann B, Leo C, type 16 infection, including loss of epithelial barrier function. J et al. The potential prognostic value of connexin 26 and 46 Virol. 2017;91:e01001–17. expression in neoadjuvant-treated breast cancer. BMC Cancer. 176. Sun P, Dong L, MacDonald AI, Akbari S, Edward M, Hodgins 2013;13:50. MB, et al. HPV16 E6 controls the gap junction protein Cx43 in 194. Chasampalioti M, Green AR, Ellis IO, Rakha EA, Jackson AM, cervical tumour cells. Viruses. 2015;7:5243–56. Spendlove I et al. Connexin 43 is an independent predictor of 177. Oelze I, Kartenbeck J, Crusius K, Alonso A. Human papillom- patient outcome in breast cancer patients. Breast Cancer Res virus type 16 E5 protein affects cell-cell communication in an Treat. 2018. epithelial cell line. J Virol. 1995;69:4489–94. 195. Nomura S, Maeda K, Noda E, Inoue T, Fukunaga S, Nagahara 178. Ablasser A, Schmid-Burgk JL, Hemmerling I, Horvath GL, H, et al. Clinical significance of the expression of connexin26 in Schmidt T, Latz E, et al. Cell intrinsic immunity spreads to colorectal cancer. J Exp Clin Cancer Res. 2010;29:79. bystander cells via the intercellular transfer of cGAMP. Nature. 196. Knosel T, Emde A, Schluns K, Chen Y, Jurchott K, Krause M, 2013;503:530–4. et al. Immunoprofiles of 11 biomarkers using tissue microarrays 179. Ariazi J, Benowitz A, De Biasi V, Den Boer ML, Cherqui S, Cui identify prognostic subgroups in colorectal cancer. Neoplasia. H, et al. Tunneling nanotubes and gap junctions-their role in 2005;7:741–7. long-range intercellular communication during development, 197. Inose T, Kato H, Kimura H, Faried A, Tanaka N, Sakai M, et al. health, and disease conditions. Front Mol Neurosci. Correlation between connexin 26 expression and poor prognosis 2017;10:333. of esophageal squamous cell carcinoma. Ann Surg Oncol. 180. Wang X, Veruki ML, Bukoreshtliev NV, Hartveit E, Gerdes HH. 2009;16:1704–10. Animal cells connected by nanotubes can be electrically coupled 198. Tanaka T, Kimura M, Ishiguro H, Mizoguchi K, Takeyama H. through interposed gap-junction channels. Proc Natl Acad Sci Connexin 43 expression is associated with poor survival in USA. 2010;107:17194–9. patients with esophageal squamous cell carcinoma. Mol Clin 181. Pasquier J, Guerrouahen BS, Al Thawadi H, Ghiabi P, Maleki Oncol. 2016;4:989–93. M, Abu-Kaoud N, et al. Preferential transfer of mitochondria 199. Liu X, Furuya T, Li D, Xu J, Cao X, Li Q, et al. Connexin 26 from endothelial to cancer cells through tunneling nanotubes expression correlates with less aggressive phenotype of intestinal modulates chemoresistance. J Transl Med. 2013;11:94. type-gastric carcinomas. Int J Mol Med. 2010;25:709–16. 182. Polak R, de Rooij B, Pieters R, den Boer ML. B-cell precursor 200. Danos K, Brauswetter D, Birtalan E, Pato A, Bencsik G, Krenacs acute lymphoblastic leukemia cells use tunneling nanotubes to T, et al. The potential prognostic value of connexin 43 expres- orchestrate their microenvironment. Blood. 2015;126:2404–14. sion in head and neck squamous cell carcinomas. Appl Immu- 183. Thayanithy V, Dickson EL, Steer C, Subramanian S, Lou E. nohistochem Mol Morphol. 2016;24:476–81. Tumor-stromal cross talk: direct cell-to-cell transfer of oncogenic 201. Wang ZS, Wu LQ, Yi X, Geng C, Li YJ, Yao RY. Connexin-43 microRNAs via tunneling nanotubes. Transl Res. 2014;164:359– can delay early recurrence and metastasis in patients with hepa- 65. titis B-related hepatocellular carcinoma and low serum alpha- 184. Osswald M, Solecki G, Wick W, Winkler F. A malignant cellular fetoprotein after radical hepatectomy. BMC Cancer. 2013;13:306. network in gliomas: potential clinical implications. Neuro Oncol. 202. Ito A, Koma Y, Uchino K, Okada T, Ohbayashi C, Tsubota N, 2016;18:479–85. et al. Increased expression of connexin 26 in the invasive Connexins in cancer: bridging the gap to the clinic 4451 component of lung squamous cell carcinoma: significant corre- 212. Fukushima M, Hattori Y, Yoshizawa T, Maitani Y. Combination lation with poor prognosis. Cancer Lett. 2006;234:239–48. of non-viral connexin 43 gene therapy and docetaxel inhibits the 203. Du G, Yang Y, Zhang Y, Sun T, Liu W, Wang Y, et al. growth of human prostate cancer in mice. Int J Oncol. Thrombocytosis and immunohistochemical expression of con- 2007;30:225–31. nexin 43 at diagnosis predict survival in advanced non-small-cell 213. Shishido SN, Delahaye A, Beck A, Nguyen TA. The anticancer lung cancer treated with cisplatin-based chemotherapy. Cancer effect of PQ1 in the MMTV-PyVT mouse model. Int J Cancer. Chemother Pharmacol. 2013;71:893–904. 2014;134:1474–83. 204. Brockmeyer P, Jung K, Perske C, Schliephake H, Hemmerlein 214. Shishido SN, Nguyen TA. Gap junction enhancer increases B. Membrane connexin 43 acts as an independent prognostic efficacy of cisplatin to attenuate mammary tumor growth. PLoS marker in oral squamous cell carcinoma. Int J Oncol. One. 2012;7:e44963. 2014;45:273–81. 215. Li S, Gao Y, Pu K, Ma L, Song X, Liu Y. All-trans retinoic acid 205. Liang QL, Wang BR, Chen GQ, Li GH, Xu YY. Clinical sig- enhances bystander effect of suicide-gene therapy against nificance of vascular endothelial growth factor and connexin43 medulloblastomas. Neurosci Lett. 2011;503:115–9. for predicting pancreatic cancer clinicopathologic parameters. 216. Wu L, Zhou WB, Shen F, Liu W, Wu HW, Zhou SJ, et al. Med Oncol. 2010;27:1164–70. Connexin32mediated antitumor effects of suicide gene therapy 206. Bijnsdorp IV, Rozendaal L, van Moorselaar RJ, Geldof AA. A against hepatocellular carcinoma: In vitro and in vivo anticancer predictive role for noncancerous prostate cells: low connexin-26 activity. Mol Med Rep. 2016;13:3213–9. expression in radical prostatectomy tissues predicts metastasis. 217. Ryu CH, Park KY, Kim SM, Jeong CH, Woo JS, Hou Y, et al. Br J Cancer. 2012;107:1963–8. Valproic acid enhances anti-tumor effect of mesenchymal stem 207. Benko G, Spajic B, Demirovic A, Stimac G, Kru Sbreve Lin B, cell mediated HSV-TK gene therapy in intracranial glioma. Tomas D. Prognostic value of connexin43 expression in patients Biochem Biophys Res Commun. 2012;421:585–90. with clinically localized prostate cancer. Prostate Cancer Pro- 218. Touraine RL, Vahanian N, Ramsey WJ, Blaese RM. Enhance- static Dis. 2011;14:90–95. ment of the herpes simplex virus thymidine kinase/ganciclovir 208. Xu N, Chen HJ, Chen SH, Xue XY, Chen H, Zheng QS, et al. bystander effect and its antitumor efficacy in vivo by pharma- Reduced Connexin 43 expression is associated with tumor cologic manipulation of gap junctions. Hum Gene Ther. malignant behaviors and biochemical recurrence-free survival of 1998;9:2385–91. prostate cancer. Oncotarget. 2016;7:67476–84. 219. Liu L, Li H, Guo Z, Ma X, Cao N, Zheng Y, et al. The com- 209. Bui MM, Han G, Acs G, Reed D, Gonzalez RJ, Pasha TL, et al. bination of three natural compounds effectively prevented lung Connexin 43 is a potential prognostic biomarker for ewing sar- carcinogenesis by optimal wound healing. PLoS One. 2015;10: coma/primitive neuroectodermal tumor. Sarcoma. e0143438. 2011;2011:971050. 220. Ito A, Morita N, Miura D, Koma Y, Kataoka TR, Yamasaki H, 210. Buhrmann C, Shayan P, Kraehe P, Popper B, Goel A, Shakibaei et al. A derivative of oleamide potently inhibits the spontaneous M. Resveratrol induces chemosensitization to 5-fluorouracil metastasis of mouse melanoma BL6 cells. Carcinogenesis. through up-regulation of intercellular junctions, epithelial-to- 2004;25:2015–22. mesenchymal transition and apoptosis in colorectal cancer. 221. Ohba Y, Kanao Y, Morita N, Fujii E, Hohrai M, Takatsuji M, Biochem Pharmacol. 2015;98:51–68. et al. Oleamide derivatives suppress the spontaneous metastasis 211. Harikumar KB, Kunnumakkara AB, Sethi G, Diagaradjane P, by inhibiting connexin 26. Int J Cancer. 2007;121:47–54. Anand P, Pandey MK, et al. Resveratrol, a multitargeted agent, 222. Miura D, Kida Y, Nojima H. Camellia oil and its distillate can enhance antitumor activity of gemcitabine in vitro and in fractions effectively inhibit the spontaneous metastasis of mouse orthotopic mouse model of human pancreatic cancer. Int J melanoma BL6 cells. FEBS Lett. 2007;581:2541–8. Cancer. 2010;127:257–68.