INTERCELLULAR COMMUNICATION AND ITS ROLE IN CANCER
MAKSIM SINYUK
Bachelor of Science in Biology
Cleveland State University
December 2010
submitted in partial fulfillment of requirements for the degree
DOCTOR OF PHILOSOPHY IN REGULATORY BIOLOGY
at the
CLEVELAND STATE UNIVERSITY
December 2018
We hereby approve this thesis/dissertation for Maksim Sinyuk
Candidate for the Doctor of Philosophy in Regulatory Biology degree for the Department of Biological, Geological, and Environmental Sciences and the CLEVELAND STATE UNIVERSITY College of Graduate Studies by
______Date: ______
Dissertation Chairperson, Justin Lathia, PhD Department of Cellular and Molecular Medicine, Cleveland Clinic
______Date: ______
Dissertation Committee member, Ofer Reizes, PhD Department of Cellular and Molecular Medicine, Cleveland Clinic
______Date: ______
Dissertation Committee member, Alex Almasan, PhD Department of Cancer Biology, Cleveland Clinic
______Date: ______
Dissertation Committee member, Keith McCrae, MD Department of Cellular and Molecular Medicine, Cleveland Clinic
______Date: ______
Dissertation Committee member, Aaron Severson, PhD Department of Biological, Geological, and Environmental Sciences, Cleveland State University
______Date: ______
Dissertation Committee member, Angela Ting, PhD Genomic Medicine Institute, Cleveland Clinic
______Date: ______
Dissertation Committee member, Zihua Gong, MD, PhD Department of Cancer Biology, Cleveland Clinic
Date of Defense: October 23, 2018
DEDICATION
I dedicate this work, first and foremost, to all the men and women who have lost their fight with cancer. Even though research is a slow and incremental process, I've never forgotten that the true purpose of my career is to ultimately help humanity in any way I can. I'm humbled and honored that for the past five years, I've had the opportunity to begin to fulfill this calling. I also dedicate this work to my entire family, in particular my mother Natalia
Sinyuk and father Apollinariy Zaginey. They never stopped believing in me, even though there have been many moments when I did so myself. I will always be grateful for their support and love throughout my life, without which none of this would have been possible.
Lastly I dedicate this work to Nicole McCullough whose never-ending optimism, trust, and love have been a light in my life.
ACKNOWLEDGEMENTS
I thank my advisor/mentor Dr. Justin D. Lathia. He took a chance on a Research Technician and I'm grateful for all of the opportunities he has provided me since then. Dr. Lathia has been an inspiration to me and I hold him in highest regard as both a scientist and person.
He has only ever sought to support my development as a researcher, a critical thinker, and a colleague. His genuine nature and eternal optimism have been a bright beacon for me throughout my PhD. Without his guidance, patience, and kindness during this early part of my career, I would be lesser for it. From the bottom of my heart, I thank Dr. Lathia for enabling me to succeed and grow in his lab as well as his dedication in training the next generation of scientists.
Thank you, Dr. James Hale for your friendship and support during my years in the lab. You have been my go-to for advice, scientific or otherwise, the entire time. All of the things you do in lab are invaluable for every single person and your absence is always felt keenly.
I'm honored that I was able to attend your wedding and meet your children five days after they were born, even though I knew you were nervous about letting people around them so early. Thank you for always lending an ear and listening to me complain about failed experiments or appliances.
Thank you, Dr. Erin Mulkearns-Hubert for being my "lab sister." You managed to keep me grounded and always spoke the truth, regardless of whether I wanted to hear it or not.
Thank you for listening to me when I proposed improperly controlled experiments and for helping me get back on track. Thank you simply for all of your life advice, without which
I would have made even more mistakes than usual. Most of all thank you for "fixing my commas and apostrophes" both in manuscripts and life. You've always seen how the small
things can affect the whole and I will be grateful always for our sometimes silly conversations.
Thank you, Dr. Daniel Silver for driving me to be the best scientist I can be. You never shied away from asking me to explain my methods and experimental design to ensure that my research was sound. Your insights into philosophy, dad-jokes, and cancer have been a constant source of inspiration over the years. I could not have asked for a more inquisitive mind to work with and I can't wait until we can talk about Book 10 in the Stormlight
Archive.
Thank you to all my committee members who have guided my Ph.D.; Dr. Ofer Reizes, Dr.
Aaron Severson, Dr. Alex Almasan, Dr. Keith McKrae, and Dr. Justin Lathia. All of your suggestions and guidance enabled me to navigate my own doctoral thesis. Thank you for taking a new graduate student and help him to defend his work over the years.
Thank you to the entire Lathia Lab for making my Ph.D. experience something that I will remember as extraordinary. I wish I could write about how each individual person made an impact on me but that would require a book. I will never forget anyone one of you and
I hope that when you stumble upon this dissertation in the annals of the internet in years to come that you will remember me as well. Even though we have always been known as the
Lathia Lab, I will treasure our time together and will think of you as my Lathia Family as
I leave to follow the next steps in my career. I will never forget you.
INTERCELLULAR COMMUNICATION AND ITS ROLE IN CANCER
MAKSIM SINYUK
ABSTRACT
To ensure proper coordination of normal tissue function, rapid intercellular communication is required between individual cells. Cells utilize adhesion complexes called gap junctions
(GJs), specialized intercellular connections composed of connexin (Cx) proteins that allow direct transport of small molecules and ions between the cytoplasm of adjacent cells to communicate. GJ intercellular communication (GJIC) is an integral communication mechanism by which homeostasis is maintained to allow for precise cellular signaling as well as for the initiation and integration of signaling cascades and cellular function.
However, connexin function in cancer biology remains poorly understood. To interrogate how connexins and GJIC contribute to tumorigenesis, I utilized cellular models of leukemia and ovarian cancer. These tumor types display common characteristics, including the need to engage in intercellular communication between individual cancer cells to drive proliferation and survival. To characterize communication between leukemia cancer cells,
I employed preclinical in vitro models and measured GJ function through dye transfer assays. Clinically relevant GJ inhibitors, carbenoxolone (CBX) and 1-octanol, were utilized to uncouple the communicative capability of leukemia cells. A qRT-PCR screen identified Cx25 as a promising adjuvant therapeutic target in leukemia, and Cx25 but not
Cx43 reduction via RNA interference decreased GJIC and sensitized cells to chemotherapy. In ovarian cancer, I sought to address how pharmacological inhibition of direct intercellular communication impacts tumor cell survival to serve as an additive novel therapeutic in both chemotherapy-sensitive and-resistant ovarian cancer. To validate that ovarian cancer cells are capable of coupling, I used transmission electron microscopy to
vi directly visualize GJ-like structures between adjacent cells. I assessed dye transfer via surrogate cell-cell transfer assays and confirmed that ovarian cancer cells were capable of coupling and communication. To evaluate the consequence of blocking GJIC, I utilized a clinically relevant inhibitor with GJ disruption properties, mefloquine, and found that treatment reduced cisplatin- sensitive and-resistant ovarian cancer cell proliferation and increased apoptosis across all available models. These findings suggest that disruption of
GJIC may be promising for both leukemia and ovarian cancer patients and demonstrate the means by which GJ-inhibiting strategies can be exploited for the development of novel anti-cancer therapies.
vii
TABLE OF CONTENTS
Page
ABSTRACT……………………………………………………………..………..….. vi
LIST OF TABLES………………………………………………..…………...….…... xi
LIST OF FIGURES …………………………………………………………..…...…. xii
NOMENCLATURE ……………………………………………….…..………….… xiv
CHAPTER I. INTRODUCTION
1.1 Introduction……….…………………….…………………….... 1
1.2 Connexin Mutations and Disease ……………...... 8
1.3 Connexins of the Heart...... …………..……… 9
1.4 Connexin Expression and Neurological Disorders...... 10
1.5 Connexins, Communication, and Deafness...... 11
1.6 Connexins and Cancer...... 12
1.7 Connexins and Cancer Stem Cells ...... 15
1.8 GJs, the Bystander Effect, and Cancer Therapy...... 15
1.9 Connexins, GJIC, and Cancer...... 20
1.10 Hemichannels and Cancer Biology...... 22
1.11 Connexins, Cancer, and their Protein Partners...... 27
II. CONNEXINS AND LEUKEMIA
2.1 Introduction ...... 31
2.2 Methods ...... 34
2.2.1 Cell Culture and Preparation of Culture Medium……… 34
2.2.2 Isolation of CD34+ HSCs ……...... 35
2.2.3 Dye Transfer Assay ...... 35
viii
2.2.4 Transwell Assay ……...... …. 36
2.2.5 Proliferation Assay...... 37
2.2.6 Quantitative Real-Time PCR ...... 37
2.2.7 Detection of Cx25 by Immunoblotting...... 38
2.2.8 Cx25 Knockdown ...... 39
2.2.9 Cx25 Immunofluorescence ...... 40
2.2.10 Bioinformatic Analysis …...... ………. 40
2.2.11 Statistical Analysis……...... …………...…………... 41
2.3 Results……………………………………………....…………. 41
2.3.1 Leukemia cells are capable of direct communication….. 41
2.3.2 Effects of gap junction inhibition on dye transfer…….. 46
2.3.3 Bioinformatic Screen of Connexins in Leukemia……… 49
2.3.4 Cx25 is Highly Expressed in Leukemia Cell Lines Compared to Other Tumors...... 49
2.3.5 Cx26 Knockdown Inhibits Leukemia Cell-Cell Communication but not Proliferation...... 53
2.3.6 Cx25 Knockdown Sensitizes Leukemia Cells to Chemotherapy………………………………………..... 57
2.4 Discussion ...... 63
III. THE ROLE OF GJIC IN OVARIAN CANCER
3.1 Introduction ...... 70
3.2 Materials and Methods ...... 75
3.2.1 Cell Culture and Preparation of Culture Medium……… 75
3.2.2 Dye Transfer Assay……...... 75
3.2.3 Transmission Electron Microscopy...... 76
3.2.4 Proliferation Assay...... 77
ix
3.2.5 Apoptosis Assay……...... 78
3.2.6 Statistical Analysis...... 78
3.3 Results...... 79
3.3.1 Cisplatin Sensitivity Validation...... 79
3.3.2 Ovarian Cancer Cells are Capable of Communication… 81
3.3.3 Inhibition of GJIC...... 85
3.3.4 Bioinformatics Analysis...... 90
3.4 Discussion...... 92
IV. CONCLUDING REMARKS…………………………………………….... 96
REFERENCES……………………………………………………………………….. 104
x
LIST OF TABLES
Table Page
I. Connexin function in cancer…………………………....……………..……..... 30
II. Connexin expression in ovarian cancer………...... …....……………………… 91
III. Mefloquine IC50 values ……………………………....……………………... 100
xi
LIST OF FIGURES
Figure Page
1. Schematic of connexin function ...... 7
2. Functional gap junction activity in leukemia…………………………...... 42
3. Histogram of Calcein intensity…….………………………………………….. 43
4. Direct physical contact is required for GJIC……………………….……….… 45
5. Pharmacological blockade attenuates GJIC …………………………………… 47
6. GJ inhibition decreases cell communication…………………………………… 48
7. TCGA RNA-Seq Analysis ……………………………………...... 50
8. Cx25 expression is increased in leukemia ……………………...... 51
9. qRT-PCR analysis of connexin expression in leukemia ……….……...... 52
10. RNAi of Cx25………………...... …….…………….. 54
11. Cx25 KD and Jurkat proliferation…………………...………...……………..... 55
12. Cx25 KD reduces dye transfer………………………….………....………….. 56
13. Cx25 KD increases leukemia cell sensitivity ………..………...……………… 59
14. Cx43 KD does not chemosensitize leukemic cells …..…...... …..……………… 60
15. Cx43 KD does not affect dye transfer …………...... ……………………. 61
16. 1-Octanol does not affect leukemia cell proliferation ……………………….... 62
17. Ovarian cancer cell cisplatin validation ……...... …… 80
xii
18. Ovarian cancer cells are capable of GJIC…...... ……….. 82
19. Dye transfer visualization ………...... …………………...... 83
20. Transmission electron microscopy of ovarian cancer cell junctions …...... …… 84
21. Pharmacological inhibition of GJIC ...... 86
22. Cell growth following GJIC inhibition ...... 87
23. Direct cell counts after mefloquine treatment ...... 88
24. Mefloquine treatment increases apoptosis ...... 89
xiii
NOMENCLATURE
AAP10 Anti-arrhythmic peptide
ALL Acute lymphoblastic leukemia
AML Acute myeloid leukemia
AP-1 Activator protein 1
Ara-C Arabinofuranosyl cytidine
CBX Carbenoxolone
CLL Chronic lymphocytic leukemia
CCOC Clear cell ovarian cancer
CML Chronic myeloid leukemia
CMTX1 Charcot-Marie-Tooth neuropathy x type 1
CSC Cancer stem cell
Cx Connexin
CxRE Connexin-responsive elements
GJ Gap junction
GJIC Gap junction intercellular communication
ECM Extracellular matrix
EGF Epidermal growth factor
ENOC Endometrioid ovarian cancer
xiv
EOC Epithelial ovarian cancer
ERAD Endoplasmic reticulum associated degradation
ERK Extracellular signal-regulated kinase
FBS Fetal bovine serum
GA 18-α-glycyrrhetinic acid
GBM Glioblastoma
HGPRTase Hypoxanthine phosphoribosyltransferase
HGSOC High-grade serous ovarian cancer
HSC Hematopoietic stem cell
Sp1/3 Sp transcription factor
Jak/STAT Janus kinases (JAKs), Signal Transducer and Activator of
Transcription proteins (STATs) miRNA microRNA
MOC Mucinous ovarian cancer
NaBu Sodium butyrate
NO Nitric oxide
OS Overall survival
PI3K Phosphatidylinositide 3-kinases
PFS Progression-free survival
xv
ROS Reactive oxygen species
RNS Reactive nitrogen species
SCF Stem cell factor
TCGA The Cancer Genome Atlas
TGF-β1 Transforming growth factor-β1
TNBC Triple negative breast cancer
TNF-α Tumor necrosis factor-α
xvi
CHAPTER I
INTRODUCTION1
1.1 Introduction To ensure proper coordination of tissue function, rapid intercellular communication is required between individual cells as well as their microenvironment (Marusyk et al.,
2012). Within normal physiological conditions, cells respond to a number of external stimuli including soluble mediators (Ratajczak et al., 2006), extracellular matrix (ECM) surroundings (Brownlee, 2002), and their neighboring stroma (Dorshkind et al., 1993). As such, cell-cell and cell-ECM interactions are integral communication mechanisms by which homeostasis is maintained to allow for precise signaling in response to both external and internal stimuli. On a cellular level, adhesion molecules play a critical phenotypic role, as evidenced by their multifunctionality in providing structural support and mediating cytoskeletal organization (Giancotti and Ruoslahti, 1999).
The ability for cells to communicate and exchange ions through low-resistance pathways was first described in myocardium, as adjacent cells were found to be capable of
1 A part of this chapter has been submitted for publication in Frontiers in Oncology under the title "Cancer Connectors: Connexins, Gap Junctions, and Communication." Authors: Maksim Sinyuk, Erin Mulkearns-Hubert, Ofer Reizes, and Justin Lathia. Submission Date: September 15, 2018.
1 transmitting electrical signals amongst each other (Weidmann, 1952). In 1953 visual evidence for the existence of GJ structures was demonstrated via electron microscopy in squid and crayfish Schwann cells (Robertson, 1953). In the intervening decades, it has been found that GJs are unique structural components of cellular plasma membranes, facilitating communication between adjacent cells (Kumar and Gilula, 1996). Importantly, accumulating evidence has demonstrated that most cells and tissues of the body utilize GJs as a means for communication during development and normal physiology which can be co-opted by cancer cells (Goodenough and Paul, 2009). As mentioned above, GJ channels can consist of different connexons which are themselves made up of identical or differential connexin subunits, regulating their physiological properties, conductive capacity, and permeability (Gong and Nicholson, 2001, Goldberg et al., 2004). The central function of
GJIC is to share metabolic demands across multiple cells in order to control for spatial gradients of nutrients and signaling molecules found in the extracellular environment or coming about as a result of cell stress which is also vital for tumor cell survival. In addition,
GJIC can help cells respond to somatic mutations when a critical metabolic enzyme or ion channel becomes dysfunctional. The loss of one such biologically vital function in a particular cell can be compensated by their connexin family counterparts (Goodenough and
Paul, 2009).
Disruption of adhesion complexes interferes with normal tissue function and results in initiation events for pathophysiological studies (Dbouk et al., 2009). Intercellular communication, mediated by GJs, is vital for the maintenance of cell survival in a variety of different tissues (Vinken et al., 2006). Electron microscopy analyses revealed that GJs often present as distinct crystalline-like plaques on cell membranes. These plaques are
2 composed of a family of proteins termed connexins. Thus far, 20 different connexin genes have been characterized in mice and 21 in humans (Esseltine and Laird, 2016). Each connexin has tissue and developmentally-specific function in mammalian biology, although redundancy does exist between subunits (Belousov et al., 2017). Different connexin isoforms display spatial and temporal specificity which is modulated by transcription factors including the Sp transcription factors (Sp1 and Sp3), activator protein
(AP-1), and members of the Jak/STAT pathway (Miyoshi et al., 2001). Furthermore, cell- specific transcription factors such as Nkx2, HNF-1, Mist1, NFκB, can regulate connexin gene expression, allowing for precise expression of connexins during development and homeostasis (Oyamada et al., 2005).
Connexin proteins are designated by molecular mass, while their respective genes are classified by sequence homology at the nucleotide and amino acid levels. At least three subgroups of connexins have been described and are classified as α, β, or γ. Thus, a 26 kDa connexin protein is referred to as connexin 26 (Cx26) or gap junction β-2 (GJB2).
Structurally, all connexin protein subunits share a comparable topology composed of cytoplasmic N- and C- terminal domains along with four transmembrane regions, two extracellular loops, and one intracellular loop (L. Harris, 2002). However, different isoforms exhibit variability in their cytoplasmic domains, which allows for a variety of different interactions and biological roles (Hervé et al., 2007). Most connexins are modified post-translationally through phosphorylation, primarily on serines, which regulates a variety of connexin processes such as trafficking to membranes, assembly, degradation, and gating of functional GJ channels (Lampe and Lau, 2000). During their short half-life of approximately 2-4 hours, six connexin proteins form a hexameric arrangement in the
3 endoplasmic reticulum or Golgi body and are then trafficked as connexons, or hemichannels, to cellular membranes (Laird, 2006). Connexons can be composed of the same connexin subunit to form homomeric connexons or different subunits to form heteromeric hemichannels. However, not all connexin combinations are capable of forming functional channels, and all channels do not have an equal capability to dock with one another (Dedek et al., 2006). Thus, specific arrangements confer different properties of conductance and regulation in the resulting channels, which allows for a level of control for intercellular communication.
Connexons from one cell can dock with connexons of adjacent cells to form GJ intercellular channels which allow for the passage of ions, second messengers, microRNA
(miRNAs) (Lim et al., 2011), and other small molecules directly between the cytoplasm of joined cells without contact with the extracellular environment (Aasen et al., 2016). This allows cells to quickly coordinate their behavior and regulate signaling during development and normal physiology in various organ systems including brain, heart, lens, liver, ovary, breast, skin, and others (Kar et al., 2012). The function of connexins and by extension gap junctional intercellular communication (GJIC) is of critical importance for normal physiology as evidenced by the former’s ubiquitous expression in nearly every mammalian cell, (summarized in Goodenough et al., 1996). Furthermore, many cell types co-express two or more members of the connexin family which may have overlapping or distinct functions. For example keratinocytes have been shown to express Cx26 (Salomon et al.,
1994), Cx43 (Wiszniewski et al., 2000), Cx31.1 (A. and L., 1994), and Cx30 (Di et al.,
2005). Additionally, cardiomyocytes have been found to express Cx40 (Moreno, 2004),
Cx43 (Garcia-Dorado et al., 2002), and Cx45 (Gross and Jongsman, 1996) while
4 hepatocytes primarily express Cx26 and Cx32 (Zhang and Nicholson, 1994). In this manner, co-expression of multiple connexin family members within the same cell type allows for compensatory communication mechanisms should the expression of one subunit become perturbed.
Historically, studies of connexin function have focused on their role in the formation of GJs to enable GJIC between cells. However, during the 1990s evidence began to emerge suggesting an alternative role for GJs in the form of undocked hemichannels
(Goodenough and Paul, 2003). In nascent stages of investigation, it was thought that unopposed connexin hemichannel activity would drown cells in Na+ and Ca2+ and lead to the loss of metabolites necessary for cellular function. However open hemichannels were eventually described in Xenopus oocytes, mediated in part by Cx46 (Paul et al., 1991). It was found that oocytes rapidly deteriorated and died unless high amounts of Ca2+ was utilized to keep the hemichannel in a closed position. Further studies found that Cx44
(Gupta et al., 1994) and Cx56 (Ebihara and Steiner, 1993) were also able to form conductive hemichannels in Xenopus oocytes. Other subunits such as Cx35 (W. et al.,
1999), Cx32 (Gómez-Hernández et al., 2003), and Cx52.6 (Zoidl et al., 2004) were later shown to have similar capability. Thus a second important role for connexins has quickly become apparent and warrants closer scrutiny which will be explored further in this dissertation.
Lastly, GJIC can be modulated by connexin-associating proteins including regulatory phosphatases, cytoskeletal elements, and enzymes. Thus apart from facilitating
GJIC and hemichannel activity, GJs have increasingly been perceived as signaling complexes that are important in the regulation of cell function and transformation (Dbouk
5 et al., 2009). As such, a complete understanding of connexin biology and subsequent GJ function can only be achieved through the identification of binding partners which may play critical roles in GJ formation, gating, and connexin transport (Spagnol et al., 2018,
Ambrosi et al., 2016).
Based on the studies described above, connexin biology can broadly be classified by three different criteria, namely cell-cell communication, hemichannel activity, and direct connexin-protein interaction to activate downstream signaling pathways and affect cellular phenotypes as shown in Figure 1. Each function plays critical roles in normal physiology and is necessary for proper cellular behavior during development. Furthermore, connexin dysfunction in each of the described axes can contribute to a wide variety of disease states including cancer. Thus, it is critical to understand connexin multifunctionality in normal physiology as well as pathology. As such, the purpose of this dissertation is to characterize connexin function in multiple cancer types and understand their role as it relates to intercellular communication as schematized in Figure 1.
6
Figure 1. The Three Main Functions of Connexins. Six connexin subunits are able to oligomerize into membrane-spanning structures termed connexons. Connexons from adjacent cells are capable of docking and forming channels through which ions, second messengers, miRNAs, and other small signaling molecules can passively diffuse between coupled cells without coming into contact with the extracellular environment (Panel 1). Furthermore, individual connexons can function as hemichannels to allow molecules from the ECM to enter or exit the cellular cytoplasm in a diffuse manner (Panel 2). Lastly, connexin subunits have an intracellular C-terminal domain, allowing for connexin-protein interactions and impacting downstream1.2. Connexin signaling Mutations events viaand GJ Disease-independent mechanisms (Panel 3).
7
1.2 Connexin Mutation and Disease
Multicellular organisms require intercellular communication to coordinate complex behavioral mechanisms in response to internal and external stimuli. GJ channels, comprised of connexin proteins, are the most common means of cell-cell communication.
Thus, it is not surprising that connexin dysfunction is associated with disease states. As of
November 2, 2018 there were 2325 peer-reviewed articles implicating connexins with a wide variety of diseases (PubMed Search: “connexins” AND “disease”) and as such it is virtually impossible to cover each example in detail. Due to their specificity in function, cells are required to tightly control connexin expression during all stages of development and homeostasis. When this process goes awry as a result of heritable or acquired mutations, aberrant connexin expression can be associated with a variety of different pathologies. However, mutations in connexin genes that result in disease have diverse effects on connexin protein expression. In some cases mutant connexins do not move past cellular quality control mechanisms and are thus forced towards endoplasmic reticulum associated degradation (ERAD) or get arrested in the Golgi apparatus (Laird et al., 2017).
Mutated connexins can also lose the ability to complex into functional hemichannels or
GJs due to dysfunction in the channel pore. Likewise, in some cases connexin mutations can result in the protein acquiring an aberrant half-life, causing it to be cleared before it can complete its normal function. Some mutations can also cause connexins to lose their capability of associating with the interactome, leading to disease formation. Lastly, mutations can result in gain-of-function mechanisms causing affected connexins to oligomerize with subunits that they would not normally interact with. The resulting
8 aberrant interactions can lead to improperly activated hemichannels as well as dead or leaky GJ channels which can contribute to cellular pathologies (Laird et al., 2017).
1.3 Connexins of the Heart
Cardiac cells are known to contain several connexins alongside Cx43, namely
Cx40, and Cx45 (Delmar et al., 2017). Moreover, gap junctions have been thought to play a role in a variety of different cardiac pathologies, resulting in both electrical disturbances and structural abnormalities. Each of the three connexin genes have been deleted via embryonic stem cell targeting and it has been shown that all three genes are necessary for heart conduction (Wei et al., 2004). As such, conditional deletion of Cx43, only in adult myocardiocytes, impacted heart conduction and suggested that lack of Cx43 could provide an arrhythmogenic phenotype which can contribute to heart dysfunction (Gutstein et al.,
2001, Veerataghavan et al., 2018). Complete knockout of Cx45 or Cx43 in mice led to early death during gestation, due to conduction block, endocardial cushion defects, or cardiac malformation (Kumai et al., 2001, Eckardt et al., 2004). In contrast Cx40 knockout mice were embryonic viable but showed evidence of slowed conduction and a partial atrioventricular block (Kirchhoff et al., 1998). Furthermore, knockin gene replacement studies during which the coding region of the GJA1 gene, encoding Cx43, was replaced by the coding regions for Cx32 or Cx40, rescued the embryonic lethality of Cx43-deficient mice (Plum et al., 2000). However animals with Cx43 replacements did exhibit mild tissue morphological abnormalities demonstrating that each connexin subunit had different function depending on its resident cell and tissue (Plum et al., 2000). Gap junctions are also involved in determining infarct size following ischemia. Originally, uncoupling and inhibition of GJIC was thought to have a beneficial effect in cardiac cells by preventing
9 the spread of tissue damage. Uncoupling cardiac cells with a broad spectrum gap junction inhibitor, heptanol, resulted in a decrease in arrhythmia scores during ischemia and reperfusion. In addition infarct size due to ischemia was reduced and thus heptanol- mediated uncoupling was shown to confer cardioprotective effects in a rat model of cardiac cell death (Chen et al., 2005).
1.4 Connexin Expression and Neurological Disorders
GJs have also been found to contribute to diseases of the nervous system. Within the mammalian peripheral nervous system, GJs are mainly associated with myelinating
Schwann cells. Cx32 forms GJs between myelin lamellae connecting the Schwann cell cytoplasm with the adaxonal cell compartment inside the myelin sheath (Bergoffen et al.,
1993). This arrangement allows for the diffusion of ions and small molecules across adjacent cell membranes which form the myelin sheath. Thus, Cx32 plays a crucial role in the maintenance and homeostasis of myelinated axons by forming functional GJs (Delmar et al., 2017). Indeed, mutations in Cx32 were implicated in human disease, namely
Charcot-Marie-Tooth neuropathy X type 1 (CMTX1), a progressive peripheral neuropathy defined by a mixture of demyelination and axonal degeneration (Bone et al., 1997). More than 400 mutations have been found in the GJB1 gene encoding Cx32 while both in vitro and in vivo models of the disease confirm that most Cx32 mutations result in the inability of the connexin to form a functional GJ (Kleopa et al., 2012). In addition, oligodendrocytes, the main myelin sheath-creating cells in the CNS, have been found to express Cx32,
Cx29/31.1, and Cx47. Loss of both Cx32 and Cx47 was further associated with severe CNS demyelination and mortality in mice (Menichella et al., 2003).
10
1.5 Connexins, Communication, and Deafness
While connexin gain-of-function mutations feature prominently in skin diseases, the opposite is true of autosomal recessive, nonsyndromic, sensorineural deafness.
Interestingly, over 90 unique genes have been found to be associated with deafness overall, although mutations in GJB2, encoding Cx26, are thought to account for almost 50% of all hearing loss from severe to profound (Delmar et al., 2017). Mutations in Cx30 have also been found to be associated with deafness; however such instances are much less frequent when compared to Cx26 alterations. Most Cx26 mutations correlated with deafness are deletions, truncations, and frameshifts, indicating that hearing loss is mainly a result of loss of GJ or hemichannel activity (Xu and Nicholson, 2013). However, there are a variety of dominant, missense mutations that concomitantly produce functional GJs and lead to both deafness as well as skin dysfunction, such as KID syndrome as reviewed above. Cx26 and
Cx30 are mainly expressed in the supporting mechanosensory hair cells in the organ of
Corti as well as in the lateral wall which contains the stria vascularis (Lautermann et al.,
1998, Forge et al., 2003). The latter is necessary for the production of endolymph and generation of the endocochlear potential (EP), which is required for proper signaling in hair cells and subsequent auditory function (Mittal et al., 2017). Cx26 and Cx30 are expressed in wide overlapping patterns and demonstrate abundant GJIC, indicating that extensive networks are present in the hair cells. Interestingly, Cx26 expression has also been found in basolateral and apical areas of supporting cells, specifically in regions lacking cell-cell contacts, suggesting that cells may be able to use hemichannels as a means of communication with the perlimyphatic and endolymphatic compartments (Majumder et al., 2010). Multiple studies utilizing genetic ablation and transgenic approaches against
11
Cx26 have shown that loss of function, even when only limited to supporting epithelial cells, can lead to the death of hair cells, resulting in hearing loss (Kalsell et al., 1997, Kudo et al., 2003, García et al., 2016). However, it is not completely clear how mutations in Cx26 and Cx30 could function in tandem to lead to deafness. As such, double Cx26+/-/Cx30+/- heterozygous animals were developed and were demonstrated to exhibit EP reduction and hearing loss, although their cochlea displayed normal physiology (Mei et al., 2017). This was found to be in direct contrast to Cx26-/- mice which displayed aberrant cochlear development. Moreover, Cx26+/- or Cx30+/- animals showed no hearing loss or EP reduction suggesting that only digenic Cx26 and Cx30 mutations can impair coupling in the cochlear lateral wall and lead to deafness (Mei et al., 2017).
1.6 Cancer and Connexins
GJIC via connexins is critical for normal cellular and tissue physiology.
Furthermore, dysfunction in connexin biology via mutations can lead to a wide variety of pathophysiologies depending on altered hemichannel activity or aberrant gap junction formation. In the context of cancer, connexins were historically thought to act as tumor suppressors. Early studies interrogating liver tumor cells found that intercellular communication was absent (Loewenstein and Kanno, 1966).
Follow up studies, investigating connexin and gap junction levels in additional tumor cell types as well as in vivo tumors likewise concluded that connexins have broad tumor suppressive properties (Omori et al., 2001, Naus and Laird, 2010). Early studies to characterize intercellular communication in the context of cancer utilized rat C6 glioma cells, known to express low levels of Cx43. Moreover, their ability to dye couple was found to be deficient, indicating that their communicative ability was impaired (Zhu et al., 1991).
12
However, when intracranially implanted into rat brains, C6 cells were able to give rise to large glioma-like lesions. Interestingly, when tumor cells were transfected with full-length cDNA encoding Cx43, it was shown that dye coupling was positively correlated with Cx43 expression leading to decreased cell growth when compared to non-transfected controls
(Chi et al., 2012). In addition, Cx43 transfected C6 glioma cells were found to be less tumorigenic in vitro with growth rates that were inversely related to the amount of Cx43 expression (Naus et al., 1992). Likewise, transgenic mice lacking Cx32 (Temme et al.,
1997) or Cx43 (Plante et al., 2010) demonstrate an increased likelihood of tumorigenesis as a result of radiation or chemical induction, supporting the hypothesis that connexins have tumor suppressor capabilities.
In the intervening decades our understanding of GJIC and connexins in cancer has grown considerably and has proven more complex than originally hypothesize. As such, connexins are now accepted to have a wide range of functions outside of their canonical tumor suppressing roles. In mouse melanoma cells it was reported that Cx26 expression played critical roles in intravasation and extravasation of tumor cells via heterologous gap junction communication with endothelial cells linking connexin expression with invasion
(Ito et al., 2000). Likewise, in multiple prostate cancer cell lines, Cx43 expression was correlated with tumor cell migration and invasion (Zhang et al., 2015). Thus, far from simply acting as repressors of tumor growth, high connexin expression can also be associated with negative prognostic value.
However, while connexin expression has been found to correlate with increased or decreased cancer cell growth, depending on both the neoplasm and connexin subunit, there is still a lack of understanding about exactly which molecules are exchanged between
13 tumor cells. Furthermore, the size of the connexin family makes it difficult to pinpoint the exact role of an individual connexin in cancer biology. Over 20 connexin subunits are known to exist in humans and each may have differential roles in tumor initiation, progression, and/or metastasis. Thus, it is imperative to consider each member of the family in the context of the cancer type and its functional role as it relates to the tumor cell phenotype. In addition, it is important to remember that epidemiological studies comparing functional connexin status to cancer onset and progression are largely absent (Delmar et al., 2017). As such, while connexin expression can be linked to patient prognosis, it is challenging to study the temporal distribution of connexin-mediated signaling and GJIC through the course of cancer development and growth. Likewise, it is largely unknown how mutations in connexin proteins which lead to pathophysiological conditions outside of cancer may inform cancer risk. For example, as mentioned above, mutations in Cx26 are one of the leading causes of sensorineural hearing loss (Wingard and Zhao, 2015).
However, there are no reports that such patients are at increased risk of cancers in organs with high Cx26 expression such as the liver, gallbladder, or colon (Uhlén et al., 2015). That is not to say that such associations do not exist, but to demonstrate that an epidemiological gap exists between connexin expression and cancer risk. It is still largely unknown whether connexins represent a practical target for the prevention or treatment of cancer. While some connexins, in particular Cx43, are being investigated for chronic wound repair to reduce edema, inflammation, and lesion spread (Becker et al., 2016), there are limited current clinical trials relating to connexin function or GJIC in cancer based on searches in the
United States National Library of Medicine (clinicaltrials.gov). Indeed, caution should be taken when targeting connexins in cancer as potentially harmful secondary effects on
14 normal tissue function could result. However, as technology advances and personalized medicine becomes more widespread, targeting connexins as an adjunct anti-tumor strategy for cancer could have promising therapeutic value.
1.7 Connexins and Cancer Stem Cells
In recent years it has been demonstrably proven that multiple tumor types are composed of a heterogeneous population of cells with a small tumor-recapitulating subset at their hierarchical apex, termed cancer stem cells (CSCs), (reviewed by Dick, 2008,
Lathia et al., 2015, Batlle and Clevers, 2017). It is of little surprise that CSCs also require intercellular communication in order to proliferate and maintain their self-renewal properties through the transfer of non-coding RNA and exosomes (Patel et al., 2016). As such, it was shown that in hepatocellular carcinoma, Cx32-mediated GJIC was critical for the expansion and self-renewal of CSCs (Kawasaki et al., 2011). In other tumors characterized by the presence of CSCs, such as glioblastoma (GBM) (Hitomi et al., 2015,
Mulkearns-Hubert et al., 2018, Murphy et al., 2016, Munoz et al., 2014), and liver (Shen et al., 2018) it was found that connexin-mediated communication was an integral part of
CSC function and resulted in tumor progression. However, additional work is critical to describe whether particular connexin subunits are responsible for aberrant CSC phenotypes which would contribute to tumor growth and recapitulation.
1.8 GJs, the Bystander Effect, and Cancer Therapy
It is important to understand whether targeting connexins could serve as a beneficial target for anti-tumor therapy. While loss of GJIC is often thought of as marker for early- stage tumors, this is not an effective prognostic indicator which may reliably demonstrate
15 efficacy (Naus and Laird, 2010a). Furthermore, it is not completely evident which individual connexin subunits may be targetable in a particular cancer as few mimetic peptides are currently under consideration (Wang et al., 2013). However, when considered along with combinatorial therapy, there is reason to believe that connexins are promising molecular targets. Of particular note is the role of the "bystander effect" in facilitating the transfer of damage signals between adjacent cells. As such, one targeted cell can spread radiation, its harmful byproducts, and chemotherapy to a population of coupled tumor cells, thus minimizing the damage to normal cells that are not capable of communicating with their malignant counterparts. Thus, the bystander effect may be a promising mechanism by which drug delivery systems can be designed to specifically target cancer cells while not affecting healthy tissue. A particular example of such offsetting activity is Lesch-Nyhan syndrome, which results from impaired activity of hypoxanthine phosphoribosyltransferase (HGPRTase), a necessary enzyme in the nucleotide salvage pathway encoded by the HPRT1 gene (Bell et al., 2016). Dysfunction in HGPRTase results in overproduction of uric acid causing dystonia, gout, and self-mutilation. However, when fibroblast cells from patients with Lesch-Nyhan were cultured with normal fibroblasts, allowing for cell-cell contact, the mutant phenotype was corrected, in a process coined metabolic cooperation (Cox et al., 1970). While a specific GJ defect was not reported, intercellular communication may be hypothesized to play an important role in the disorder.
Interestingly, metabolic cooperation is also thought to play an important part in heterozygous female Lesh-Nyhan carriers, as HPRT1 is located on the X chromosome.
Thus, random X-inactivation leads to differential populations of cells that are mutant and normal. As such, females with HPRT1 mutations are largely asymptomatic due to
16 metabolic rescue of mutant cells by adjacent wild-type cells utilizing GJIC (Subak-Sharpe et al., 1969).
Communication between cells, mediated by GJs, is usually associated with beneficial correlatives enabling coordinated behavior and rapid response to a wide variety of differential catalysts. The ability of coupled cells to act in a concerted manner amongst each other may act as a mitigating factor to distribute stressors and damage responses in a given tissue, minimizing the burden on individual cells (Klammer et al., 2015). Thus, GJIC is thought to have a protective role in normal physiology. However, in some instances intercellular communication via GJs can have harmful potential (Mancuso et al., 2012).
While the exchange of metabolites, secondary messengers, and electrical signals may have benefits for both donor and acceptor cells in response to stimuli, accumulating evidence has also suggested a potentially detrimental “side-effect” as it regards GJIC. Following the discovery of X-rays by Wilhelm Röntgen, subsequent studies quickly uncovered their cytotoxic and carcinogenic nature (Ward, 1988). Moreover, it was further shown that ionizing radiation was generally damaging to cells due to its ability to generate single and double strand breaks in DNA. The utility of using radiation to destroy rapidly proliferating cancer cells was quickly recognized and heralded a new era of cancer treatment in the form of radiotherapy. In its infancy, radiation was only thought to be lethal to those cells which were directly exposed to high-energy particles. However, in the past two decades it has been demonstrated that intercellular communication can promote DNA damage responses in a phenomenon termed the “bystander effect” (Prise and O'Sullivan, 2009). Put simply this process describes the mechanism by which intercellular communication, mediated by
GJs, is able to induce DNA damage in those cells which have not been directly irradiated
17 but rather have been in contact with those that have (Gaillard et al., 2009). Thus it is becoming more widely recognized that direct irradiation may not be required in order to force cells into a DNA damage response state, especially when considered in the context of cancer. Irradiated cells can exert influence on their unirradiated “bystander” neighbors
(Sokolov et al., 2005). A common feature of genomic instability in response to radiation is the formation of DNA double-strand breaks (DSBs), evidenced by the induction of γ-
H2AX foci at DSB sites. When irradiated human fibroblasts where exposed to 20 α- particles and mixed with non-irradiated cells, the bystander cells were found to have 3.7- fold increased γ-H2AX foci, indicative of DNA damage. Inhibition of GJIC via lindane prevented the bystander effect when cells were grown in conditions enabling them to couple, while this effect was not recapitulated in media transfer studies (Sokolov et al.,
2005). As such, direct cell-cell coupling and GJIC is a likely means by which harmful signals can be transferred among cells. While direct transfer of irradiated particles is not thought to be a major contributor to the bystander effect, the formation and spread of free radicals amongst adjacent cells in response to DNA damage has been implicated (Mancuso et al., 2012). Additional evidence utilizing X-ray studies with immortalized human keratinocytes (HPV-G) showed that plating efficiencies were decreased when cells were grown as microcolonies compared to single cell irradiation (Cummins et al., 1999). A subsequent study replicated this effect and found that when 10 individual cell nuclei in a culture of normal human urothelium were irradiated with 103 He2+ ions, approximately
2000-6000 cells could become damaged as a result of the bystander effect (Belyakov et al.,
2003). It was also shown that when fibroblasts and epithelial cells were exposed to low streams of α-particles their nonirradiated bystander counterparts, growing in the same
18 culture, were found to exhibit similar DNA damage responses (Azzam et al., 2001).
Furthermore, when cells were genetically compromised in their ability to participate in
GJIC, mediated by Cx43, they were no longer able to induce p21 expression after exposure to radiation, further demonstrating that the bystander effect is at least in part due to the inhibition of GJIC (Azzam et al., 2001).
As a general rule the inhibition of communication is most often associated with harmful cellular phenotypes due to decreased responsiveness to external and internal stimuli. Furthermore, perturbation of GJIC during development can negatively affect normal tissue function and in many cases mutations in connexin subunits can lead to lethality. However, under base line conditions GJIC enables cells to rapidly respond to a huge variety of different signals in order to adjust their own internal conditions to best suit the ever-changing extracellular milieu. Similarly to interactions between individual people, proper communication at the cellular level is critical for the overall well-being and function of tissues to ensure appropriate coordination and subsequent regulation of life-processes.
However, a necessary factor involved in connexin studies is the development of genetic protocols and chemicals which are able to induce or inhibit specific connexins in individual neoplasms, or in even more complex situations such as individual tumor cells.
This is particularly important as systemic upregulation of connexins in normal host tissues which undergo continual renewal, could have harmful effects for the development of cancer (Naus and Laird, 2010b). In reference to cancer, enhanced connexin expression may also lead to increased cancer cell metastasis (Stoletov et al., 2013) (Kanczuga‐Koda et al.,
2006) (Kotini and Mayor, 2015). Therefore, it becomes necessary to define which connexins play roles in normal development and which play communication roles in late
19 stage tumor progression for a given tissue. As such, it is possible to imagine that a strategy may be developed through the use of genetic inhibition to manipulate connexin function and combine it with other therapeutic means to provide overall benefit in cancer patients.
1.9 Connexins, GJIC, and Cancer
Despite early evidence that connexins function as tumor suppressors, exceptions have been found to exist in recent years showing that increased connexin expression may lead to tumors with more aggressive phenotypes (Cronier et al., 2009). Evidence was first seen in melanoma cells transfected with cDNA coding for Cx26 which increased cellular metastatic capability in subcutaneous models of disease (Ito et al., 2000). The authors conjectured that this was due to a more effective way of facilitating cellular intravasation and extravasation, dependent on Cx26, which was found to aid GJIC between tumor cells and normal endothelium (Ito et al., 2000). Multiple other studies have further confirmed that increased connexin expression within tumors can lead to greater metastatic potential, migration, and invasion (Pollmann et al., 2005, Elzarrad et al., 2008). However, Cx43 has also been found to reverse EMT and prevent resistance to cisplatin therapy in A549 lung adenocarcinomas (Yu et al., 2014) while increased adhesion and GJIC has long been known to play similarly critical roles in other highly metastatic lung carcinomas (el-Sabban and Pauli, 1994). Moreover, individual connexins can have dual roles in the same tissue, acting as a tumor suppressors in primary progression while facilitating cancer growth in later stages of disease (Aasen et al., 2016).
As such, it is not difficult to surmise that molecular mechanisms associated with
GJIC and tumor suppressive roles are linked to particular signals that are exchanged among
20 healthy cells and their cancerous counterparts. Furthermore, it should be understood that like healthy cells, tumor cells have the capability of interaction with their microenvironment to affect their phenotype, although the exact contents of these interactions are still being puzzled out. Some molecules such as glutathione, a tripeptide with high permeability through gap junction channels (Goldberg et al., 1999), have been shown to have antioxidant properties to protect cells from ROS and DNA damage
(Balendiran et al., 2004). Radiation is the most widely cited inducer of the bystander effect in cancer cells, creating molecules such as reactive oxygen species (ROS), reactive nitrogen species (RNS), protein factors, and DNA molecules which can spread via GJIC from the originally perturbed cell to damage its surrounding neighbors (Verma and Tiku,
2017), although exosomal signaling and hemichannel activity cannot be completely discounted as part of the mechanism. However, while it is commonly accepted that GJIC is one of the most important mechanisms behind the bystander effect, additional molecules such as necrosis factor-α (TNF-α) (Shareef et al., 2007), transforming growth factor-β1
(TGF-β1) (Iyer and Lehnert, 2000), interleukin-6 (IL-6) (Chou et al., 2007), IL-8 (Facoetti et al., 2006), and nitric oxide (NO) (Matsumoto et al., 2001) have also been described, adding further complexity when considering cell-cell communication. Paradoxically, positive signals can also be released by targeted cells and spread to adjacent non-targeted cells to induce bystander responses. Thus, GJIC has remained a critical component of normal physiology as well as cancer biology. It is necessary to understand how the bystander effect may be manipulated to target tumor cells via gap junctions although promising results are quickly manifesting for cancer therapeutics.
21
1.10 Hemichannels and Cancer Biology
Traditionally, GJs have been associated with cell-cell coupling and communication.
As such, connexon hemichannels were first thought to be simple structural precursors to
GJ channels before they had the opportunity to dock with their counterparts on adjacent cells (Dbouk et al., 2009). However, unlike the investigation of GJIC, the quantification of connexin-mediated hemichannel activity presents with more challenges because multiple, unrelated mechanisms exist to facilitate how cells open the membrane pores to communicate with their extracellular environment. Nevertheless, evidence for functional connexon hemichannel activity was first elucidated in catfish retina cells due to their permeability to Lucifer dye (DeVries and Schwartz, 1992). Further studies sought to better define the particular connexin subunits responsible for the formation of permeable channels in cells. As such, upon induction of Cx46 expression in Xenopus oocytes, it was found that cells similarly became water permeable and underwent lysis unless osmotically buffered with Ficoll (Paul et al., 1991). While initial observations reached the conclusion that hemichannels remained closed until connexons docked with one another to form GJs, subsequent work would describe a variety of different regulatory mechanisms to facilitate pore activity. Among these are intracellular and extracellular factors such as changes in the ionic concentration of the microenvironment, in particular Ca2+ gradients (Gómez-
Hernández et al., 2003), although Na+ and K+ have also been implicated in hemichannel regulation (Srinivas et al., 2006). Furthermore, membrane depolarization has been found to induce single hemichannel opening in HeLa cells ectopically expressing Cx43
(Contreras et al., 2003). Likewise, metabolic inhibition was sufficient to open heterologously-expressed Cx43 hemichannels in cardiac cells upon exposure to calcium-
22 free media conditions (John et al., 1999). Importantly, differences between ionic concentrations in cells and their microenvironment are not the only method by which hemichannel activity can be affected, as fluid flow shear stress was found to induce Cx43 translocation to osteocyte membrane surfaces in order to serve as a release mechanism for prostaglandin E2 in response to mechanical strain (Cherian et al., 2005).
With the use of connexin mutants resulting in dead channels, as well as pharmacological inhibitors of GJIC, it has been determined that connexins display their tumor suppressive properties in a hemichannel-dependent manner (Lee et al., 2002, Zhang et al., 2003). The mechanisms by which this occurs are still in the early stages of investigation, although it was been noted that the tumor suppressive capabilities of Cx43 in keratinocytes may be linked to its interaction with caveolin 1, another factor associated with tumor suppression (Langlois et al., 2010). Moreover, it has been shown that Cx43 expression has also been implicated in prostate cancer and is correlated with its metastatic potential. However, only direct Cx43 knockdown but not gap junction channel formation was seen to decrease cell migration and invasion, indicating that hemichannel activity was critical for cellular function (Zhang et al., 2015). Furthermore, the cytoplasmic C-terminal domain of Cx43 was shown to be sufficient in suppressing neuroblastoma progression via
Src signaling (Moorby and Patel, 2001). Interestingly, connexin hemichannel activity was found to be crucial for the regulation of the actin cytoskeleton in human glioma cells
(Crespin et al., 2010). Two Cx43 mutants were created, one without the C-terminal domain and one without the entire transmembrane domain, and a similar reduction in glioma proliferation was described (Crespin et al., 2010). Truncation of Cx43 did not alter gap junction coupling and it was demonstrated that the Cx43 C-terminal domain was sufficient
23 to induce glioma cell migration, associated with a lamellipodia-type migration and actin cytoskeleton regulation (Crespin et al., 2010). Moreover, it has been shown that Cx43 is associated with increased sensitivity to sunitinib-induced cytotoxicity in malignant mesothelioma cells, an effect that is independent of channel formation but is rather a result of its interaction with the apoptotic factor Bax (Uzu et al., 2015). Cx43 hemichannel cancer function was also investigated in transfected HeLa cells and primary human foreskin fibroblasts. 18-α-glycyrrhetinic acid (GA) was used to reduce gap junctional coupling but not overall Cx43 protein level expression. Sodium butyrate (NaBu) or anti-arrhythmic peptide (AAP10) were used to enhance Cx43 expression in the cells (Grek et al., 2015).
With time-lapse microscopy it was shown that Cx43 levels delayed mitotic duration, corresponding with an accumulation of cells in G1, further leading to increased levels of p21waf1/cip1, a cell cycle inhibitor (Johnstone et al., 2010). Overall it is suggested that the upregulation of Cx43 delays the cell cycle rate through the delay of G1, pointing to more roles relating to the GJIC-independent function of connexins.
Importantly, numerous studies have continued to reveal that connexin-mediated hemichannel activity plays an integral role in cell-microenvironment communication in a variety of different tissues and aspects of cell life (Evans et al., 2006). Functional studies have demonstrated that hemichannels play important roles in Ca2+ signaling (Quist et al.,
2000), cell proliferation (Song et al., 2010), apoptosis (Decrock et al., 2009), as well as the normal development of a variety of cell types (Chandrasekhar and Bera, 2012). For instance neurite outgrowth in PC12 cells, derived from a pheocromocytoma of the rat adrenal medulla and used to study neuronal development (Greene and Tischler, 1976), was found to be mediated by hemichannels after stimulation with nerve growth factors.
24
Furthermore, it was found that hemichannel-mediated ATP release and its subsequent interaction with puringergic receptors was sufficient to induce growth and neuronal differentiation in the same cell line (Belliveau et al., 2006). Likewise, Cx43 hemichannels have been implicated in the back and forth movement of NAD+ which is thought to regulate Ca2+ gradients via CD38 transmembrane glycoproteins in 3T3 fibroblasts
(Bruzzone et al., 2001). Likewise, in heart ventricular myocytes, hemichannel activity has been described and shown to have osmoregulatory properties which result in both positive and negative impacts as they relate to myocardial infarctions and normal cardiac physiology (John et al., 2003).
In general, connexin hemichannels in the open position are thought to be detrimental to cells due to their impact on the efficiency of cell metabolism and the maintenance of ionic balance between the interior and exterior of cellular membranes
(Evans et al., 2006). As such, it is maintained that should hemichannels remain open to allow the passive diffusion of extracellular material into the cell or the egress of cytoplasmic contents into the extracellular environment, cells will not be able to sustain normal homeostasis and will thus undergo cell death. Put simply, open hemichannels can be thought to behave as pathogenic pores as they play important balancing roles between cell death as a result of necrosis or apoptosis via the release of ATP (Nicotera et al., 1998).
Furthermore, the depletion of cellular ATP may also activate connexin hemichannels, creating a feedback loop by opening otherwise closed channels (Vergara et al., 2003).
Staurosporin, an ATP-competitive kinase inhibitor, was shown to open Cx43 hemichannels and induce apoptosis which was itself inhibited by truncating the C-terminal tail of the connexin, thereby forming non-functional hemichannels (Hur et al., 2003).
25
Following hemichannel closure via pharmacological means, apoptosis was found to be slowed demonstrating their role in microenvironmental communication and cell function (Kalvelyte et al., 2003). In contrast, alendronate, a bisphosphonate drug used to treat osteoporosis was found to be capable of inhibiting apoptosis by opening rather than closing Cx43 hemichannels, which in turn activated Src kinase and Erk, promoting cell survival (Plotkin et al., 2002). In other organ systems such as the heart, release of ATP through different methods, including connexin hemichannels, has been demonstrated to be a stress response that is capable of vasodilation and the facilitation of increased delivery of oxygen and energy (Headrick et al., 2003). Thus, continued release of ATP would result in harmful consequences and eventually lead to cell death. As such, phosphorylation of mitochondrial Cx43 has been implicated in Cx43 hemichannel communication and cardioprotection (Boengler et al., 2006). Likewise, in tissues of the nervous system, connexin hemichannels can play both protective and harmful roles via diffusion of necrotic or apoptotic signals from injured cells to healthy ones or by allowing for the diffusion of ions and protective signals from healthy to injured cells (Kirchoff et al., Contreras et al.,
2004). Thus it becomes imperative to understand whether pharmaceutical tools can be used for cardio-and neuro-protection by targeting specific connexin hemichannels. While unspecific GJ inhibitors such as halothane, 1-octanol, carbenoxolone, and mefloquine may reduce injury in specific animal models, the need to characterize connexin hemichannel function and its roles in ATP-release or Ca2+ signaling remains under-studied. Thus, designing new and specific connexin mimetic peptides may serve as a promising and strategic means by which adjacent cells can be protected from injury as a result of
26 myocardial infarction, stroke, or other dysfunctions for which cellular communication has an important component.
1.11 Connexins, Cancer, and their Protein Partners
While connexins have traditionally been associated with communication, whether through GJIC or hemichannel formation, increasing evidence is supportive of connexin
GJ-independent roles through a diverse set of interacting partner proteins (Zhou and Jiang,
2014). Among the first of such reports, it was identified that Cx26 has suppressor function in tumor-derived mammary epithelial cells (Lee et al., 1991). Furthermore, induced expression of Cx26 in GJ-deficient MCF-7 breast cancer cells resulted in decreased proliferation, invasion, and in-vivo tumor growth, although their communicative capacity was not yet investigated (Momiyama et al., 2003). However, it did raise interesting questions regarding the role of connexin proteins on tumor growth. Subsequent studies confirmed and further demonstrated that Cx26 can inhibit breast cancer cell migration and overall tumorigenesis in the MDA-MD-435 tumor cell line, independent of GJ function.
Moreover, the mechanism was determined to result from the regulation of β1-integrin and
MMP levels, indicating that communication was not the sole function of the connexins
(Kalra et al., 2006). Interestingly, further work in TNBC, detailed that Cx26 is capable of interacting with NANOG and focal adhesion kinase (FAK) to drive tumor progression and
CSC self-renewal (Thiagarajan et al., 2018). Other connexin subunits have also been assayed for their tumor suppressive function via interaction partners. Cx32 was found to decrease tumor growth, invasion, and metastasis of renal cell carcinoma cells lines via multiple modulators, including Src, tight junction proteins, and vascular endothelial growth factor (VEGF), independent of GJIC function (Sato et al., 2007). In addition, ectopic
27 expression of mutant Cx43, without intrinsic GJ function, is able to prevent cell growth independent of GJIC but rather through the association of its cytoplastic carboxyl domain with proteins such as ZO-1 and c-Src (Moorby and Patel, 2001). Moreover, a number of different studies have evidenced that alteration of connexin expression, either through forced expression or deletion, can lead to changes in downstream gene expression in seemingly unrelated pathways. As such Cx43 deletions were studied in the context of astrocytes in the neonatal brain and it was found that large numbers of genes were statistically changed in mice with decreased expression of the connexin (Iacobas et al.,
2004). Likewise, it was detailed that Cx43 deletions in spinal cords of mice with experimental autoimmune encephalomyelitis (EAE) were associated with alterations in a wide range of functions within cells (Iacobas et al., 2004).
However, the mechanisms behind connexin GJ-independent are still being investigated. One proposed solution is thought to involve genomic connexin-responsive elements (CxRE) which contain Sp1/Sp3 complexes that assemble at promoter regions in a GJ-dependent manner. To demonstrate the molecular mechanisms of GJ-sensitive transcriptional regulation, the rat osteocalcin promoter was utilized. This CT-rich connexin
CxRE was able to confer GJ sensitivity to the promoter as repression of osteocalcin transcription was shown when GJIC was perturbed (Stains et al., 2003). Thus, the functional consequences of inhibiting CxREs is the regulation of genes that have the promoter element and respond to differential connexin regulation. Indeed, the effect of connexins on gene expression has been investigated in multiple studies wherein re- expression of connexins in deficient tumors is sufficient to affect their characteristics, namely tumorigenicity, which cannot simply be the result of cytoplasmic exchange (Dbouk
28 et al., 2009). Cell differentiation has also been implicated in gap junction-independent functions of connexins as evidenced by the ability of Cx45.6, but not Cx43 or Cx56, to stimulate lens cell formation regardless of its ability to form GJ channels. Furthermore, it was demonstrated that its C-terminal domain was, by itself, enough to induce lens cell differentiation (Gu et al., 2003). Likewise, Cx43 was shown to control the directional motility of cardiac neural crest cells via the actin-binding protein vinculin and drebrin (Xu et al., 2006), demonstrating that connexins should not only be thought of in the context of cellular communication but should rather be considered in a cell-specific manner. Adding weight to such conclusions are observations that embryonic neurons are capable of migrating using Cx43 and Cx26 as they provide cytoskeletal contacts with radial fibers without the exchange of cytoplasmic contents (Elias et al., 2007).
Thus, our understanding of connexin function is constantly being updated as individual cells and tissues are interrogated. As such this dissertation focuses on connexin function in the context of leukemia and ovarian cancer. It is becoming more appreciated that no longer can connexins be studied in the context of cell-cell communication, but rather in a wider milieu involving disparate and seemingly unrelated processes. Additional work is necessary to define how connexins are able to carry out each of their roles and whether one function is more important than each other. However, a GJ-independent role for connexins opens up a world of novel observations that could be critical for normal physiology or in the context of cancer. Table 1 summarizes each of the three main functions of connexins in cancer and the resulting data will be the focus of the described studies.
29
Connexin Cancer Function Tumor Activity References
Connexin 25 Leukemia GJIC Pro-tumorigenic (Sinyuk et al., 2015)
Connexin 26 Breast, Cervical, GJIC Anti-tumorigenic (Kalra et al., 2006)
Cervical Hemichannel Anti-tumorigenic (Polusani et al., activity 2016) Breast Protein- Pro-tumorigenic (Thiagarajan et al., connexin 2018) interaction Connexin 30 Glioma, Gastric GJIC Anti-tumorigenic (Princen et al., 2001) (Sentani et al., 2010) Connexin 31.1 Head and neck GJIC Anti-tumorigenic (Broghammer et al., squamous cell 2004) carcinoma Non-small cell Protein- Anti-tumorigenic (Zhu et al., 2015) lung cancer connexin interaction Connexin 32 Breast GJIC Pro-tumorigenic (Luiza Kanczuga- Koda, 2007) Renal cell Protein- Anti-tumorigenic (Yano et al., 2004) carcinoma, connexin (Wu et al., 2017) Ovarian interaction Connexin 36 Cervical Protein- Pro-tumorigenic (Lu and Li, 2006) connexin interaction Connexin 37 Liver, Protein- Pro-tumorigenic (Saito et al., 1997) Insulinoma connexin and anti- (Burt et al., 2008) interaction tumorigenic Connexin 43 Brain GJIC Pro-tumorigenic (Grek et al., 2018)
Breast Hemichannel Anti-tumorigenic (Ferrati et al., 2017) activity Ovarian Protein- Anti-tumorigenic (Qiu et al., 2016 connexin interaction Connexin 46 Brain GJIC Pro-tumorigenic (Mulkearns-Hubert et al., 2018) (Hitomi et al., 2015) Connexin 50 Cervical GJIC Pro-tumorigenic (Liu et al., 2015)
Table I. Connexin Function in Cancer 30
CHAPTER II
CONNEXINS AND LEUKEMIA2
2.1 Introduction
Leukemia consists of a broad diagnostic category of hematological malignancies accounting for over 52,000 new cancer diagnoses in the United States, resulting in over
24,000 deaths annually (Howlader et al., 2010). Leukemia mainly affects the adult population, with a median age of 66 at diagnosis, but is also the most common cancer among children, accounting for almost 1 out of 3 pediatric cancers (Howlader et al., 2010).
Additionally, while the term “leukemia” is used to describe the four major types of leukemia, acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), chronic lymphocytic leukemia (CLL), and chronic myeloid leukemia (CML), these diseases share several defining features such as uncontrolled proliferation and self-renewal of immature lymphoblasts or myeloblasts that subsequently interfere with normal hematopoiesis.
2 This chapter has been published in Oncotarget under the title "Connexin 25 contributes to leukemia cell communication and chemosensitivity." Authors: Maksim Sinyuk, Alvaro G. Alvarado, Pavel Nesmiyanov, Jeremy Shaw, Erin E. Mulkearns-Hubert, Jennifer T. Eurich, James S. Hale, Anna Bogdanova, Masahiro Hitomi, Jaroslaw Maciejewski, Alex Y. Huang, Yogen Saunthararajah, and Justin D. Lathia. Oncotarget. 2015; 6:31508-31521
31
The standard of care for leukemia includes chemotherapy, radiation, or both in combination with drug therapy, often followed by autologous or allogenic stem cell transplantation
(Fernandez, 2010). While progress has been made in recent years toward understanding the underlying molecular mechanisms of leukemia, the average five-year survival rate is
57% in the United States and is largely dependent on the patient’s age at diagnosis, gender, race, and type of leukemia (Xie et al., 2008).
The hematopoietic microenvironment is critical in the initiation and progression of leukemia (Colmone et al., 2008). The bone marrow niche comprises a heterogeneous population of hematopoietic and non-hematopoietic stromal cells as well as their extracellular products and cytokines. These cells produce a variety of different ECM molecules including numerous interstitial and basal lamina collagens. In the close confines of the bone marrow, cell-cell and cell-ECM communication plays a large role in promoting hematopoietic progenitor cell (HPC) survival, expansion, and differentiation (Gillette and
Lippincott-Schwartz, 2009). These interactions are also thought to be essential for the tumorigenesis and progression of leukemia. Recently, evidence has shown that direct contact of leukemia cells with the surrounding stroma inhibits their apoptosis and enhances their survival following chemotherapy (Zhang et al., 2012). Leukemia cells have also been shown to communicate with endothelial cells via exosomal miRNA (Umezu et al., 2013) and may contribute to the angiogenic activity of endothelial cells, suggesting that cytoplasmic signals originating in one cell type are capable of entering another while impacting the phenotype in the recipient cell. However, there has been little evidence for the mechanism of direct homotypic communication between individual leukemia cells themselves and the implications this may cause for clinical treatment.
32
A common means for intercellular signals to quickly travel between two adjacent cells is through gap junction-mediated mechanisms. Gap junction communication is present in virtually all solid tissues and has been demonstrated to be essential for normal embryonic development (Warner et al., 1984), electric coupling in cardiac muscle (Beauchamp et al.,
2012), and neurons (Allison et al., 2006), as well as for normal hematopoiesis (Gonzalez-
Nieto et al., 2012) Additionally, connexin expression in non-excitable tissues has key roles in organ development (Kamiya et al., 2014), skeletal development and function (Stains and
Civitelli, 2005), and growth control (Kihara et al., 2010). Gap junctions are defined as cell- cell junctions at which two plasma membranes from adjacent cells link. Gap junctions are specialized intercellular connections formed by a family of at least 20 human proteins called connexins that allow for the diffusion of small molecules and ions directly between the cytoplasm of adjoining cells (Söhl and Willecke, 2004). Individual connexins show tissue-, cell-type-, or developmental stage-specific expression, and most organs as well as many cell types express more than one connexin. Six individual connexin proteins may co- oligomerize with the same (homomeric) or mixed (heteromeric) connexins into connexons, or hemichannels, although only certain combinations are possible (Hervé and Derangeon,
2013). Two hemichannels are then able to come together, forming homotypic or heterotypic gap junction channels between contacting cells, depending on connexon isotype, that facilitate cell-cell communication through the direct transfer of small molecules up to 1 kDa such as Ca2+, cyclic adenosine monophosphate (cAMP), and inositol triphosphate (IP3) (Goldberg et al., 1999).
The importance and variable nature of cell-cell communication in cancer has only recently begun to be carefully investigated and, as a result, current clinical models do not fully
33 appreciate or take into account the effects this may have on the tumor phenotype. Given that many advanced cancers such as breast, prostate, and leukemia often present as histologically dense entities, it is reasonable to consider that cell-cell communication occurs and is able to confer a survival advantage to tumor cells. Additionally, while current therapies are effective at eradicating rapidly dividing cells, they have no way of distinguishing friend from foe, and many of the side-effects associated with treatment are the direct result of normal cell death. Therefore, the identification of tumor-specific gap junction function is paramount to develop effective, complimentary anti-tumor therapies aimed at removing malignant tissue while sparing normal healthy tissue. In the current study, we demonstrate that direct cell-cell communication with ion exchange between leukemia cells occurs, in part, due to a Cx25-dependent mechanism that enables enhanced tumor cell proliferation and allows for the identification of additional potential specific gap junction protein functions as targets for anti-tumor therapy.
2.2 Methods
2.2.1 Cell Culture and Preparation of Culture Medium
Two different in vitro systems modeling ALL (Jurkat cells) and AML (THP1 and MV4-
11) along with two primary patient-derived AML specimens (AML1 and AML2) were utilized to study gap junction communication in leukemia. Jurkat cells were cultured in
RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum (FBS; Sigma), 2 mM sodium pyruvate, 5x10-5 M 2-mercaptoethanol, penicillin, streptomycin, L-glutamine, and 0.1 mM non-essential amino acids at 37°C in a humidified atmosphere of 20% oxygen and 5% CO2. The primary AML cells used in this study (AML1, AML2) have previously been described (Ng et al., 2011) and were obtained from patients with relapsed/refractory
34
AML according to approved Cleveland Clinic IRB protocols. These cells had a myelomonocytic morphology (M4) and multiple chromosomal abnormalities including t(8;18)(q22:q23) and t(11;13)(q21:q12). Primary AML cells were cultured in IMDM supplemented with 10% FBS and 10 ng/ml of the following human cytokines: stem cell factor (SCF), FLT3 ligand, thrombopoietin, interleukin-3 (IL-3), and interleukin-6 (IL-6)
(Life Technologies). The THP1 and MV4-11 cell lines were cultured in RPMI supplemented with 10% heat-inactivated fetal bovine serum (FBS; Sigma), penicillin, streptomycin, and L-glutamine.
2.2.2 Isolation of CD34+ HSCs
CD34+ cells from bone marrow aspirates were immunopositively purified using a magnetic cell sorting system (Miltenyi Biotec) per the manufacturer’s instructions and used as a negative control. Briefly, bone marrow aspirates were gathered and frozen in 10% dimethyl sulfoxide (DMSO) and FBS. Cells were stored in liquid nitrogen until needed.
After thawing, cells were purified using a magnetic cell sorting system (CD34 MicroBead
Kit #130-046-702, Miltenyl Biotec Inc., Auburn, CA, USA) according to the manufacturer’s instructions. CD34+ cells were considered normal HSCs and were cultured in IMDM supplemented with 10% fetal bovine serum and 10 ng/ml of the following human cytokines: stem cell factor, FLT3 ligand, thrombopoietin, IL-3, and IL-6.
2.2.3 Dye Transfer Assay
For dye transfer assessments, cells were divided into two groups. One group was labeled with 1 μM Calcein AM (Life Technologies), a cell-permeable dye in its parent form that is converted to the green-fluorescent calcein after acetoxymethyl ester hydrolysis by
35 intracellular esterases, after which it is no longer cell permeable and only able to leave the cell via gap junctions, hemichannels, or exocytosis. The other group of cells was labeled with 3 μM DiI (Life Technologies), a lipophilic membrane stain that diffuses laterally to stain the entire cell and does not leave the cell membrane. In both cases, cells were labeled for 45 minutes with either Calcein AM or DiI according to the manufacturer’s instructions before centrifugation to remove excess dye. These two groups of cells were then mixed at a 1:1 ratio; incubated for 1, 2, or 3 hr, and analyzed by flow cytometry using a BD Fortessa to quantify the percentage of DiI-labeled receptor cells uptaking green fluorescence from
Calcein AM labeled donor cells. Relative dye transfer is represented as the percent of DiI- labeled cells that have acquired green fluorescence after co-incubation. Single DiI- and
Calcein-stained control cells were used to verify the efficacy of the dyes and to set flow cytometry gates. After functional gap junction activity was confirmed, cells were incubated in the same manner as above in the presence or absence of the gap junction inhibitors carbenoxolone (CBX) and 1-octanol and likewise analyzed using flow cytometry. Flow cytometry experiments analyzed four different biological samples (Jurkat cells, one primary patient-derived AML specimen, and two different AML lines, MV4-11 and
THP1). For the dye transfer assay, 20, 000 events were collected from a total of five groups of 1.0x106 cells of each biological sample (one vehicle control and two carbenoxolone and two 1-octanol concentrations).
2.2.4 Transwell Assay
To confirm that dye transfer required direct cell-cell contact, cells were divided into two equal groups. One group was pre-loaded with Calcein AM while the other was pre-loaded with DiI. Cells were subsequently incubated for 1, 2, or 3 hr in 6-well dishes with 12 mm
36
Transwell Inserts containing a 0.4 μM pore (Corning) used to keep the two groups physically separated (Schematic provided in Fig. 4A). While in culture, the Calcein group was incubated in the apical chamber of the insert to ensure that if dye was released into the media, the donor DiI-labeled cells in the basolateral chamber would be capable of taking it up and subsequently fluorescing green. After incubation, both cell populations were mixed, and dye transfer was quantified using flow cytometry as described above.
2.2.5 Proliferation Assay
To quantify leukemia cell proliferation, a CellTiter-GloTM proliferation assay (Promega) was utilized. Briefly, cells were plated in clear-bottom 96-well plates at a density of 1000 cells per well in 100 µL of culture media. In addition, to test whether Cx25 knockdown was capable of sensitizing leukemia cells to chemotherapy, cells were incubated in the presence or absence of arabinofuranosyl cytidine (Ara-C), also known as cytarabine, and proliferation was measured at Day 0 and Day 3. Cells were allowed to recover for at least
2 hr before the first measurement was taken, which served as a loading control. The luminescence intensity of each well was measured using a luminometer (PerkinElmer) at several time points after plating. Each well was run in triplicate to generate statistical variation. To prepare relative growth curves, the data were normalized to day 0 to account for any variation during preparation.
2.2.6 Quantitative Real-Time PCR
Total RNA was isolated using TRIzol (Life Technologies) according to the manufacturer’s instructions. cDNA synthesis and amplification from 1 μg mRNA via PCR was performed using the qScriptTM cDNA Supermix (Quanta Biosciences) with an Eppendorf Vapo
37
Protect Mastercycler Pro (Eppendorf). This kit contains both oligo(dT) and random primers. Gene expression was measured by real-time PCR using the RT2 SYBR Green
ROXTM qPCR Mastermix Kit (SABiosciences). Connexin primers were utilized for 20 different connexins as previously described (Zhang et al., 2015) and validated by verification of single-peak melt curves. The PCR product for Cx25 was further validated by the verification of a single band on an agarose gel. The PCR reaction and detection were performed with the ABI 7000 Sequence Detection System (Applied Biosystems). To minimize variability due to single housekeeping gene adjustment, individual connexin expression levels were normalized to multiple housekeeping genes (β-actin and GAPDH).
After adjustment to housekeeping genes, the difference in cycle numbers for each individual connexin was further normalized to that of HSCs. The fold change was subsequently calculated by squaring the cycle difference between each tumor cell type and
HSCs, and each technical replicate was performed in triplicate.
2.2.7 Detection of Cx25 by Immunoblotting
Immunoblot analysis was performed on the whole-cell extracts of primary AML specimens, Jurkat cells, and primary GBM cells used as a control. Cells were lysed in 10 mM Tris HCl, pH 7.4; 0.5% IGEPAL CA-630 (weight/volume); 150 mM NaCl; 1 mM
EDTA; 2 mM sodium orthovanadate; 1 mM PMSF; and a 1:100 dilution of protease inhibitor cocktail for mammalian cells (P8340 Sigma), followed by protein determination using a Pierce BCA Protein Assay Kit (Thermo Scientific). Protein (20 μg) from cells was separated by SDS-PAGE using a 12% gel and then transferred to a PVDF membrane. After blocking the membrane with 5% milk in TBST, total Cx25 was detected using rabbit polyclonal anti-Cx25 (Sigma, SAB4501629), then incubated with the secondary antibody
38 linked to horseradish peroxidase. The immunoreactive bands were visualized by Pierce
ECL 2 Western Blotting Substrate (Thermo Scientific). Blots were washed and reprobed with an anti-actin antibody (Santa Cruz) and developed in an identical manner to assess β- actin protein levels to ensure even loading.
2.2.8 Cx25 Knockdown
Cx25 inhibition was accomplished by shRNA in Jurkat cells. In short, DNA was isolated from glycerol stocks of bacteria containing shRNA plasmid DNA (Sigma MISSION shRNA) specific to Cx25 (TRCN0000222613 (KD13) and TRCN0000074136 (KD 36)) or a nontargeting control (NT) and used to produce virus. These shRNA plasmids were chosen from a group of five total plasmids as they were shown to reliably reduce the levels of
Cx25, as demonstrated by immunoblotting and qRT-PCR. Additional shRNA constructs against Cx43 were also purified from bacterial glycerol stocks (TRCN0000059775 (KD
75) and TRCN0000059776 (KD 76) and were likewise used to inhibit Cx43 expression in
Jurkat cells as demonstrated by qRT-PCR. Bacterial stocks were expanded, and plasmid
DNA was purified using PureLink HiPure Plasmid Maxiprep Kit (Life Technologies) according to manufacturer’s instructions. 293T cells were cotransfected with the packaging vectors pMD2.G and psPAX2 (Addgene) and shRNA constructs. Media on the 293T cell cultures were changed 18 hr after transfection, and viral supernatants were collected 24,
36, and 48 hr later and filtered for immediate use or concentrated with polyethylene glycol precipitation and stored at -80⁰C for future use. Jurkat cells and MV4-11 cells were subsequently infected with lentiviral particles using the Sigma MISSION shRNA
Spinoculation protocol, and transduced cells were selected after incubation with puromycin.
39
2.2.9 Cx25 Immunofluorescence
To visualize Cx25 expression and localization in leukemia cells, Jurkat and THP1 cells were centrifuged onto glass slides using a Cytospin protocol. Cells were fixed with 4%
PFA for 10 minutes and washed three times with 0.1% PBST (PBS/Triton X-100) for 5 minutes. After washing, cells were blocked with 10% normal goat serum for 30 minutes.
Rabbit polyclonal anti-Cx25 (Sigma, SAB4501629, 1:100) was used to stain cells overnight at 4°C. The following day, cells were washed 3 times with 0.1% PBST, and the appropriate secondary antibody was applied for 2 hr at room temperature (goat anti-rabbit
IgG Alexa Flour 488, Life Technologies, 1:2000). After secondary antibody incubation, cells were washed three times with 0.1% PBST and counterstained with 4',6-diamidino-2- phenylindole (DAPI) for 5 minutes. Afterwards, cells were washed three additional times with 0.1% PBST, and coverslips were mounted using FluorSave Reagent (VWR
International). Cells were imaged with a Leica TCS SP8 confocal microscope, and images were prepared in figure form using Adobe Photoshop.
2.2.10 Bioinformatics Analysis
RNA sequencing for all connexin subunits was evaluated using the Cancer Genome Atlas
(Ley et al., 2013). Expression of Cx25, Cx31.9, Cx40, and Cx43 were interrogated in the
Beroukhim et al., Nature 2010 dataset (Beroukhim et al., 2010) in Oncomine
(www.oncomine.org).
40
2.2.11 Statistical Analysis
Data are represented as mean values +/- standard deviation. Statistical significance was analyzed using one-way ANOVA, with p values less than 0.05 considered statistically significant. p-values: * ≤ 0.05, ** ≤ 0.01, ***≤ 0.001, ****≤ 0.0001.
2.3 Results
2.3.1 Leukemia Cells are Capable of Direct Communication
To assess whether direct cell-cell communication was present in leukemia cells, we used several preclinical models of ALL and AML. After Calcein AM- and DiI-labeled Jurkat cells were mixed and incubated for 1 hr, flow cytometry analysis detected 4.1±0.4%
Calcein transfer from Calcein-labeled cells to DiI-labeled cells (Fig. 2A-B). Increasing the length of time that cells were allowed to mix permitted the Calcein to spread to more cells, indicative that dye transfer is a passive, on-going process. After 3 hr of co-incubation, dye transfer increased to 90.6±3.6%, confirming that connexin channels are functional in Jurkat cells (Fig. 2A-B). The increase in Calcein transfer over time was further validated by the increase of fluorescence intensity over time (Fig. 2C). This suggests that as cells interact with each other for increased lengths of time, Calcein transfers to a greater number of cells, thus amplifying the intensity of the signal. To better represent cell-cell communication in leukemia, dye transfer was also measured in a primary patient-derived AML cell line as well as in two characterized human AML cell lines, MV4-11 and THP1. The direct percentage of dye transfer was variable among all tested cell lines and confirmed that direct cell-cell communication is conserved among leukemia cells (Fig. 3).
41
Figure 2 Sinyuk et al.
Figure 2: Functional gap junction activity is detectable directly between leukemia cells. (A) Jurkat cells were labeled with Calcein AM or DiI and then co-incubated for 1-3 hours. After incubation, cells were analyzed by flow cytometry. Cells indicated in red are
Calcein acceptors. (B) Quantification of the percent of dye transfer between leukemia cells as a function of time (1, 2, and 3 hr). (C) Histogram of Calcein intensity in DiI-labeled Jurkat cells over time.
42
Figure 3 Sinyuk et al.
Figure 3: Histogram of Calcein intensity. Calcein intensity was measured via flow cytometry in primary patient-derived AML cells, THP1, and MV4-11 cell lines.
43
Additionally, to test whether the detected dye transfer was mediated by direct physical contact between cells and was not a result of Calcein uptake from the cell culture media itself, we performed dye transfer assays with 0.4 μm Transwell inserts, thereby keeping the two groups of cells from making contact (Fig. 4A-C ). After 1 and 3 hr of incubation, the control wells without inserts with physical cell-cell contact had a higher relative percentage of dye transfer than those that were separated by the Transwell inserts (Fig. 4B). In addition, to demonstrate that Calcein does not leak out of cells and into the surrounding media, flow cytometry was utilized to measure Calcein signal at 1, 2, and 3 hr. Indeed, leukemia cells did not lose intensity during the time course used for our experiments, arguing that dye uptake requires cell-cell contact. It should be noted that the percentage of transfer was highly variable among the four cell lines used (20-98% after 3 hr) and was dependent on several factors including cell type and the stochastic nature of cell-cell interactions in vitro, as we cannot control which individual cells form gap junction channels. However, taken together, these data demonstrate that functional and quantifiable cell-cell communication exists in leukemia and is dependent on physical cell-cell contact.
44
Figure 4 Sinyuk et al. Sinyuk et al.
Figure 4: Direct physical contact between cells is necessary for gap junction- mediated communication. (A) Schematic detailing how the Transwell assay was performed. (B) dye transfer was measured over time between two groups of cells there were allowed to physically interact and between those that were kept separate with a Transwell insert. (C) Calcein fluorescence was measured by flow cytometry over 1, 2, or 3 hrs and showed no reduction in intensity over time, indicating that no Calcein was leaking form cells.
45
2.3.2 Effects of Gap Junction Inhibition on Dye Transfer
To evaluate whether the observed cell-cell communication was gap junction dependent, two clinically relevant pan-gap junction inhibitors, CBX and 1-octanol, were used to inhibit gap junctions in leukemia cells. CBX is currently approved for the clinical treatment of esophageal and mouth ulcers in the United Kingdom (Drugs.com, 2018), while 1-octanol is currently being investigated for the treatment of essential tremor (Bahab et al., 2011).
Cells were stained with Calcein AM and DiI as previously described, treated with each inhibitor, and allowed to incubate for either 1 or 3 hr, after which the percent of Calcein dye transfer to DiI-labeled cells was measured using flow cytometry. After 1 hr of co- incubation, Jurkat cells showed a 6.1% transfer, with MV4-11 cells showing an 18.5% transfer. However, when co-incubated in the presence of 100 µM CBX or 1 mM 1-octanol, both Jurkat and MV4-11 cells exhibited reduced dye transfer percentages compared with their vehicle controls (Fig. 5). Furthermore, after 3 hr of co-incubation with CBX and 1- octanol, Jurkat and THP1 cells continued to show a reduced percentage of dye transfer
(Fig. 6), while MV4-11 cells did not demonstrate a reduction in dye transfer at this time point, indicating that slower compensatory communication mechanisms likely exist in leukemia cells when gap junctions are inhibited. These data demonstrate that homotypic cell-cell communication relies on functional gap junctions and may be blocked through pharmacological inhibition of connexin subunits.
46
Figure 5 Sinyuk et al.
Figure 5: Pharmacological blockade of gap junctions is sufficient to attenuate communication between Jurkat and MV4-11 cells. Gap junction activity was inhibited by two pan-gap junction inhibitors, CBX and 1-octanol, at pharmacologically relevant concentrations. After 1 hr of incubation, dye transfer was reduced in cells treated with inhibitors. Cells indicated in red are Calcein acceptors.
47
Figure 6 Sinyuk et al.
Figure 6: Gap junction inhibition is sufficient to decrease leukemia cell communication. After 3 hr of incubation with 100 μM of CBX or 1 mM 1-octanol, both Jurkat and THP1 cells showed decreased dye transfer by flow cytometry analysis. However, after 3 hr of co-incubation, MV4-11 cells did not show the same decrease in dye transfer. p-values: * ≤ 0.05, ** ≤ 0.01, ***≤ 0.001, ****≤ 0.0001.
48
2.3.3 Bioinformatic Screen of Connexins in AML and Jurkat Cells versus HSCs
Bioinformatics data using RNA-seq were subsequently generated to narrow down those connexins that were detected in the Cancer Genome Atlas AML dataset (Ley et al., 2013).
Samples were organized by the French American British (FAB) morphological categories, with the group expressing high Cx25/GJB7 consisting of M3 AML. Consequently, Cx25 mRNA expression was found to be present in leukemia via bioinformatic analysis (Fig. 7).
2.3.4 Cx25 is Highly Expressed in Leukemia Cell Lines Compared with Other Tumors
We have shown in vitro that inhibition of Cx25 can sensitize leukemia cells to chemotherapy, however there is limited information as to the role of Cx25 in leukemia in vivo. We interrogated Cx25 expression across multiple tumor cell lines and found significantly higher expression in leukemia cell lines as compared with other tumors (Fig.
8). This was not observed with other highly expressed connexins identified in Fig. 8 (Cx40 and Cx31.9 and Cx43). Taken together, these data demonstrate that Cx25 may have a unique role in leukemia cell communication and may serve as an attractive target for the development of future adjuvant therapeutics. To identify whether connexin 25 is expressed in leukemia, a qRT-PCR screen of known connexin subunits was employed. Normal hematopoietic stem cells (HSCs) were probed to identify tumor-specific connexins important in leukemia cells but not healthy controls. Three connexins were found to be increased in all leukemia cell lines tested: Cx25, Cx40, and Cx31.9 (Fig. 9).
49
Figure 7 Sinyuk et al.
Figure 7: The Cancer Genome Atlas (TCGA) AML RNA-Seq analysis of connexin expression. Gene expression measured by RNA sequencing (TCGA AML, Runx1 wild type n=179). Heat map of gene expression values generated by ArrayStar software.
50
Figure 8 Sinyuk et al. Cancer cell line expression
Figure 8: Cx25 expression is elevated in leukemia cell lines compared with additional tumor cell lines and connexin subunits. Box and whisker plots of connexin subunit expression across multiple cancer cell lines demonstrates that Cx25 is significantly elevated in leukemia cell lines (black box) compared with other tumor types. Data accessed from the Beroukhim et al., Nature 2010 dataset in Oncomine and number of tumor cell lines per groups is indicated for each tumor. Boxes span the 25-75th percentile, line represents median value, bars represent expression range, and *** p<0.001
51
Figure 9 Sinyuk et al.
Figure 9: qRT-PCR analysis of connexin expression in leukemia. mRNA profiles of connexin expression were interrogated in normal HSCs, Jurkat cells, two primary patient- derived AML cell specimens, and two AML cell lines, MV4-11 and THP1. Two connexins were found to be more highly expressed in all leukemia cells versus normal HSCs, Cx25 and Cx40, while primary AML cells lines expressed higher levels of Cx31.9 compared with HSCs. 52
2.3.5 Cx25 Knockdown Inhibits Leukemia Cell-Cell Communication but Not
Proliferation
By PCR-based analysis, Cx25 and Cx40 were identified as potential tumor-promoting connexin subunits expressed in both primary AML cells and Jurkat cells, while Cx31.9 was expressed by primary AML cell lines. To validate our observation at the protein level, immunoblot analysis of Cx25 and Cx31.9 was utilized. Cx25 protein expression was found in all leukemia cell lines tested (Fig. 10A), although Cx31.9 protein expression was undetectable (data not shown). In addition Cx25 expression was visualized in both Jurkat and THP1 cells using immunofluorescence and Cx25 staining was demonstrated on cell membranes (Fig. 11C). To further confirm the role of Cx25 in leukemia, we utilized a genetic approach to disrupt Cx25 by RNA interference (RNAi). We obtained two independent short hairpin RNA (shRNA) constructs to knock down Cx25 expression
(knockdown 13 (KD 13) and knockdown 36 (KD 36)) in Jurkat cells. Compared with a nontargeting (NT) control, both Cx25 knockdown constructs reduced Cx25 expression as evaluated by immunoblotting and qRT-PCR (Fig. 10B). Dye transfer assays were subsequently utilized to measure whether cell-cell communication was disrupted after
Cx25 knockdown. A decrease in dye transfer was observed in Cx25 knockdown cells after
1 hr of incubation (11% dye transfer in KD 13 cells and 76% dye transfer in KD 36 cells) versus the NT control (87% dye transfer) (Fig. 12). However, after 3 hr of incubation, the percent of transfer was similar in all three groups, indicating the presence of additional compensatory communication mechanisms not dependent on Cx25.
53
Figure 10 Sinyuk et al.
Figure 10: Targeting Cx25 by RNA interference. (A) Immunoblot analysis of Jurkat cells, two primary patient-derived AML cell lines, and two AML cell lines to assess Cx25 protein expression. (B) Inhibition of Cx25 by shRNA-mediated knockdown showed decreased protein expression following transduction with two targeting constructs. qRT- PCR was utilized to validate the shRNA constructs and determined that both KD12 and KD36 were able to decrease Cx25 mRNA expression. p-values: * ≤ 0.05, ** ≤ 0.01, ***≤ 0.001, ****≤ 0.0001.
54
Figure 11 Sinyuk et al.
Figure 11: Cx25 KD does not affect proliferation in Jurkat cells. (A) Cx25 expression was observed in two primary AML lines as measured by immunoblotting. (B) Following
Cx25 knockdown, Jurkat cell proliferation was measured, and relative growth was not found to change significantly between the two shRNA constructs and NT control. (C) Micrographs of Jurkat cells and THP1 cells prepared by Cytospin and stained with an anit-Cx25 antibody (green). Yellow arrows indicate cell-cell contact, scale bar represents 10 microns.
55
Figure 12 Sinyuk et al.
Figure 12: Cx25 KD reduces calcein dye transfer. Following inhibition of Cx25, Calcein dye transfer was reduced in Jurkat cells at 1 hr compared with the NT control. Cells indicated in red are Calcein acceptors.
56
2.3.6 Cx25 Knockdown Sensitizes Leukemia Cells to Chemotherapy
Following Cx25 knockdown, the proliferative capability of leukemia cells was interrogated but did not show a reduction compared with NT controls (Fig. 11B), indicating that the disruption of one connexin subunit was not sufficient to induce apoptosis. Interestingly, when Cx25 knockdown cells were incubated in the presence of the chemotherapeutic agent
Ara-C at a concentration much lower than previously reported in the literature (Negrotto et al., 2012), the knockdown cells demonstrated a reduced capability to proliferate compared with their NT counterparts when incubated with 15 nM Ara-C (Fig. 13A) (also known as arabinofuranosyl cytidine), a common chemotherapeutic agent used for the treatment of AML and non-Hodgkin lymphoma, suggesting that gap junction inhibition in combination with chemotherapy may be a potentially viable treatment strategy. These results indicate that Cx25 knockdown sensitizes leukemia cells to chemotherapeutics in vitro and may justify gap junction inhibition as an addition to current standard-of-care regimens. Although Jurkat cell proliferation was not affected by Cx25 knockdown, the disruption of cell-cell communication suggests a therapeutic potential for the targeting of gap junction function in leukemia. To that end, we further interrogated the phenotype of
Cx25 knockdown in Jurkat and MV4-11 cells. We hypothesized that inhibition of Cx25 may sensitize leukemia cells to chemotherapeutics, as they would respond less rapidly to cellular damage. Indeed, when Cx25 was knocked down in Jurkat and MV4-11 cells, both constructs decreased proliferation in the presence of 15 nM Ara-C To test whether this phenomenon was dependent on Cx25, we similarly knocked down Cx43 with two shRNA constructs previously described in prostate cancer (Zhang et al., 2015). The decision to focus on Cx43 as a control rather than the other connexin subunits identified in the qRT-
57
PCR screen was mediated by the fact that Cx43 is heavily studied in a wide variety of normal and pathophysiological contexts and would allow for a phenotypic comparison following shRNA disruption of either subunit. After confirming Cx43 knockdown with qRT-PCR (Fig. 14A), we treated Jurkat cells with 15 nM Ara-C and found that Cx43 inhibition did not sensitize the cells to chemotherapy (Fig. 14B). As Cx43 and its gene
GJA1 have known tumor suppressor functions (Naus and Laird, 2010), the increase in proliferation of Jurkat cells following a 50% decrease in mRNA transcript levels is indicative of their cancer promoting function in leukemia. In addition, we performed dye transfer assays with Cx43-knockdown Jurkat cells and found that dye transfer was not affected at any time point (Fig. 15), indicating that Cx25 rather than Cx43 plays an important role in leukemia cell communication and chemosensitivity. Finally, we wanted to elucidate whether pharmacological inhibition of connexins is a feasible strategy for the development of future therapeutics. All four cell lines tested showed a decreased rate of proliferation following treatment with 100 μM CBX (Fig. 13B). However, 1-octanol was less effective at decreasing proliferation, as treatment with several concentrations did not affect Jurkat cell growth (Fig. 16). These data demonstrate that pan-gap junction inhibition has a negative effect on leukemia cell growth, while disrupting Cx25 is sufficient to chemosensitize both Jurkat and MV4-11 cells.
58
Figure 13 Sinyuk et al.
Figure 13: Cx25 KD increases leukemia cell chemosensitivity, while treatment with a gap junction inhibitors decreases leukemia cell proliferation. (A) Treatment of Cx25-knockdown Jurkat and MV4-11 cells with 15 nM Ara-C reduced their proliferative activity. (B) Inhibiting gap junction activity in Jurkat cells, one primary patient-derived AML specimen, and two AML cells lines with 100 μM CBX reduced cell proliferation. p-values: * ≤ 0.05, ** ≤ 0.01, ***≤ 0.001, ****≤ 0.0001.
59
Figure 14 Sinyuk et al.
Figure 14: Cx43 KD does not chemosensitize Jurkat cells. (A) Cx43 KD was validated with qRT-PCR using two different constructs (KD 75 and KD76). (B) After Cx43 KD, Jurkat cells were not sensitized to treatment with 15 nM Ara-C. Knockdown cells demonstrated increased proliferation, indicating that Cx43 functions as a tumor suppressor in leukemia. p-values: * ≤ 0.05, ** ≤ 0.01, ***≤ 0.001, ****≤ 0.0001.
60
Figure 15 Sinyuk et al.
Figure 15: Cx43 KD does not affect Jurkat cell dye transfer. Jurkat cell dye transfer was not affected by Cx43 KD compared with NT controls at any time point. Cells indicated in red are Calcein acceptors.
61
Figure 16 Sinyuk et al.
Figure 16: 1-Octanol is not effective at reducing leukemia cell proliferation. Jurkat cells were incubated with 1-octanol at concentrations ranging from 50 μM to 1 mM. No concentration of 1-octanol was sufficient to reduce Jurkat cell proliferation.
62
2.4 Discussion
The cellular microenvironment plays a major role in tumorigenesis by providing nutrients and survival signals to cancer cells (Marusyk et al., 2012). Additionally, the microenvironment may protect tumor cells from normal immune responses and promote resistance to therapeutic treatment regimens. Evidence in ALL has demonstrated that osteoblasts secrete an ECM molecule, osteopontin (OPN), that plays an important role in anchoring leukemia blasts in anatomic microenvironmental locations that support tumor cell dormancy and protect them from cytotoxic chemotherapy (Boyerinas et al., 2013). In other cases, tumor cells are even able to co-opt their microenvironment by crowding out normal cells, subsequently fashioning their milieu to provide them with a competitive survival advantage (Colmone et al., 2008). However, in many instances, the direct mechanisms by which tumor cells utilize their surroundings is poorly understood. While tumor cells may secrete various stimulatory growth factors and cytokines such as TNF-α,
IL-6, TGF-β, and IL-10 to avoid immunosurveillance and create a pro-tumorigenic environment, little is yet understood about how direct cell-cell communication between the tumor cells, rather than tumor-stroma interactions, influences malignancy.
Here, we have shown that leukemia cells directly communicate with each other through at least one specific connexin, Cx25, which is increased in primary patient AML cells compared with normal hematopoietic cells. Likewise, Cx25 expression was also detected in Jurkat cells, a model for ALL, as well as in MV4-11 and THP1 cells, both models for
AML. Using RNAi strategies to disrupt Cx25 in Jurkat and MV4-11 cells, we found that the communicative ability of the cells was reduced, which translated to increased chemosensitivity following treatment with Ara-C. This finding is clinically relevant, as one
63 of the major challenges in the treatment of leukemia is the tolerability of current anti- leukemia agents. Children, the elderly, and those in general poor health are particularly susceptible to the side-effects of chemotherapy and are often unable to be treated. Future studies should be aimed at determining if Cx25 is linked to leukemia patient survival across multiple leukemia subtypes. However, this may be challenging as many databases do not contain Cx25 and these data may need to be generated. To further translate our findings, leukemia cell were treated with clinically relevant gap junction inhibitors, CBX and 1- octanol, at physiologically relevant concentrations [Drugs.com, 2018, Nahab et al., 2011), and cell-cell transfer of calcein decreased. However, only CBX was found to be effective at reducing leukemia cell proliferation.
It is important to note that proliferation and cell number were quantified via Cell-Titer
Glo®. The assay measured ATP using firefly luciferase and is one of the most commonly applied methods for estimating the number of viable cells in a high-throughput manner
(Riss et al., 2013). When cells die, their membranes lose integrity and their ability to synthesize ATP is lost. Endogenous ATPases then deplete any remaining cytoplasmic
ATP, enabling the assay readout to only measure ATP from metabolically active cells (Riss et al., 2013). The detection reagent used in the assay contains a detergent to lyse cells,
ATPase inhibitors to ensure that cell lysis is not responsible for the release of molecules that can degrade ATP in live cells, a substrate (luciferin), and a firefly luciferase which catalyzes the production of detectable light photons from the luciferin and ATP reaction.
The luminescent signal from this reaction can then be detected on a wide variety of microplate readers. The ATP assay is the fastest measurement of cell viability and is generally less prone to cellular artifacts than other viability measurements. Importantly, it
64 has been shown to have sensitivity for the detection of fewer than 10 wells in a 1536-well format (Riss et al., 2013). However, using ATP as a surrogate measure of cell number is not without its own challenges. Decreased intracellular ATP concentrations can result from nonlethal perturbations that can cause cessation of proliferation. Senescence or contact inhibition can also decrease the amount of cellular ATP as mitochondrial respiration is impacted (Méry et al., 2017). As such, the measurement of metabolic activity via ATP may not accurately reflect cell viability. To account for these factors, alternative cell proliferation experiments, such as directly counting viable cells should be utilized to validate and discriminate between cytotoxic or antiproliferative effects (Méry et al., 2017).
The inhibition of gap junction communication in leukemia cells followed by treatment with chemotherapeutics may allow for a much lower chemotherapy dose to be used, achieving the desired results while sparing patients from the harmful effects inherent to cancer treatment. However additional in vitro and in vivo studies are critical to determine the exact mechanisms behind gap junction-mediated cellular communication in leukemia. A wider screen of chemotherapeutics would also be necessary to investigate which combinatorial adjuvants best complement each other in pre-clinical models of leukemia.
Physical contact between cells was also found to be necessary for functional gap junction activity, as prohibiting this interaction did not allow for Calcein transfer. The ability of leukemia cells to communicate with each other and their surrounding microenvironment through gap junctions has the potential to affect the malignancy of leukemia. Cells are better able to respond to external stimuli and escape damage from sources such as chemotherapeutics and radiation by exchanging information amongst themselves. In addition, gap junctions may allow for the release of potentially lethal intercellular
65 components such as reactive oxygen species (ROS) in response to cell damage or may facilitate the uptake of molecules that protect from ROS-induced DNA damage (Upham and Trosko, 2009). Recent work on gap junctions in HSCs has confirmed that Cx43- deficient HSCs and HPCs displayed decreased survival and increased senescence mediated by their inability to transfer ROS to the hematopoietic microenvironment following myeloablation (Taniguchi et al., 2012). These results demonstrate that Cx43 plays a protective role during stressful conditions such as hematopoietic recovery.
Targeting gap junction-mediated communication in leukemia and other cancers is emerging as an exciting prospective strategy with easily translatable results (Trosko and
Ruch, 2002, Zhang et al., 2015). Specifically, the additive survival advantage that gap junction inhibition in combination with chemotherapy confers in animal models of glioblastoma is particularly promising (Hitomi et al., 2015). However, several caveats remain to be addressed regarding both CBX and 1-octanol before clinical trials are implemented. Both agents demonstrate remarkable efficacy for inhibiting gap junctions and tumor cell growth in vitro and in vivo. However, their mechanism of action is poorly understood. In particular, it should be noted that these compounds do not specifically block individual connexin subunits or gap junctions. Rather, they are pan-gap junction inhibitors, ostensibly blocking the function of all connexins and making it difficult to study the particular connexin subunits that are involved in tumor biology. It is also important to remember that these inhibitors are capable of blocking additional hemichannels, such as pannexins, necessating careful examination of the molecular signaling pathways behind the mechanisms of CBX and 1-octanol function, as current reports remain contradictory.
Less is known about connexin hemichannel activity, although both CBX and 1-octanol
66 have been shown to be capable of disrupting hemichannels in keratinocytes (Barr et al.,
2013) and spinal cord astrocytes (Chen et al., 2014) inhibiting ATP and chemokine release, respectively. Functional connexin and pannexin hemichannels have also been described to play a role in leukocyte adhesion (Véliz et al., 2008), as demonstrated by decreased adhesion to venular endothelium after treatment with CBX. However, a contradictory study reported that reducing Cx40 levels instead increased leukocyte adhesion to mouse endothelial cells (Chadjichristos et al., 2010). These contrasting conclusions are likely a result of the tissue-specific function of connexins in vivo and highlight the complexity of connexin function. It is also important to note that blocking all gap junctions may have unintended off-target effects that need to be addressed before considering clinical trials.
Additionally, the exact method by which the agents inhibit connexin function is an ongoing area of investigation. It has been hypothesized that both CBX and 1-octanol act on cell membranes to alter fluidity and disrupt the transmembrane domains of connexin proteins, rendering them inert. However, this explanation has yet to be fully investigated and remains speculative.
The ablation of highly expressed connexins is not sufficient to target and destroy leukemia cells, nor is it sufficient to completely prevent cell-cell communication. Targeting specific connexins may also have secondary effects or cause a concomitant increase in the expression of other connexin proteins, leading to unintended phenotypic consequences. It is prudent to consider that connexins may possess additional functions that have yet to be fully described. To this end, cytoplasmic partners capable of interacting with the intracellular domains of connexin proteins may provide a potential means of specifically targeting individual subunits. The ablation of one universal connexin may have unintended
67 secondary effects or no effects at all, as compensatory mechanisms likely exist among various connexin proteins. Rather, gap junction inhibition strategies should be contextualized in light of the overall tumor subtype or, even more effectively, in light of the cell-of-origin of the tumor, to target the root of the malignancy rather than the branches.
Of paramount importance is the development of novel mimetic peptides or agents capable of disrupting individual connexin subunits to minimize the harm to normal tissue in the course of treatment. Cancer therapy as a whole is moving away from a “one-size-fits-all” paradigm and toward a more individualized model. Targeting specific connexin subunits, depending on tumor subtype, is therefore complementary to the emerging trends regarding cancer care and should be considered for further attention.
Following radiation therapy, gap junctions also play an important role in the bystander effect, in which cells that are not directly exposed to radiation but are in the vicinity also respond to the exposure and display increased levels of genetic change and lethality. Recent work using genetic approaches to downregulate Cx43 demonstrated that gap junction- mediated communication is crucial for the transmission of harmful free radicals as a result of radiation, causing tumor responses in the distal CNS in areas not exposed to direct radiation therapy. Cx43 was found to be upregulated in nontargeted tissue following irradiation, which may allow for the transduction of potentially oncogenic signals to remote tissue through the bystander effect (Mancuso et al., 2011). In addition, Cx43 knockdown in leukemia was found to increase cancer cell proliferation, demonstrating its tumor suppressive role as opposed to Cx25 (Fig. 14). An attractive additional strategy for radiotherapy would be the identification of tumor-specific gap junctions through which
“death signals” may be transferred into the extracellular microenvironment to affect
68 adjacent tumor cells that may have been protected from radiation. Alternatively, tumor- specific antigens may be released through gap junctions after radiotherapy or chemotherapy, priming the innate immune response to identify and eradicate unaffected tumor cells. This underlines the need to study gap junctions in the context of cancer and develop novel connexin inhibitors for pharmacological use (Naus and Laird, 2010).
Additional work is needed to determine the molecular mechanisms of gap junction signaling to gain a better understanding of the downstream signaling components behind connexin activation or inhibition. We have taken the first steps toward characterizing connexin function in leukemia, although many hurdles remain before gap junction inhibition will become an important clinical tool for cancer therapy. However, our results are an important proof-of-principle example, demonstrating the value in exploring and exploiting novel concepts in the fight against cancer.
69
CHAPTER III
THE ROLE OF GJIC IN OVARIAN CANCER
3.1 Introduction
To further study GJIC and its role in cancer, a second tumor model was necessary to compare how solid tumors respond to inhibition of communication compared to hematological malignancies. Ovarian cancer (OC) remains the most lethal gynecological malignancy in the United States, with a predicted 22,240 new cases in 2018 that will account for approximately 14,070 deaths (Torre et al., 2018). A woman’s overall lifetime risk of developing OC is 1 in 75 while her chance of dying of the malignancy is 1 in 100
(Reid et al., 2017). Several factors play important roles in the lethality of OC and its subsequent high mortality rate. The disease is usually diagnosed at a late stage during which tumor cells have metastasized from their tissue of origin and invaded the abdominal cavity and other distal organs. The average 5-year survival rate for such diagnoses is only
29% while diagnosis of localized OC is relatively high at 92% (Howlader et al., 2011).
Worldwide, the 5-year survival rate remains 30-40%, which has only increased modestly
(2-4%) since 1995, demonstrating the need for more effective treatment strategies
(Allemani et al., 2015).
70
Making therapeutic intervention more difficult in OC is the cellular heterogeneity seen in the tumors. Most benign and malignant tumors arise from three main cell types: epithelial, stromal, or germ cells. In the developed world, over 90% of ovarian tumors are of epithelial origin, while 5-6% arise from sex cord-stromal tumors, and the remainder are thought to develop from germ cells, (as reviewed by Chen et al., 2003). As such, due to its prevalence, most epidemiologic research is concentrated on epithelial OC (EOC) which is itself composed of distinct histologic subtypes that have unique properties such as their cellular origin, pathogenesis, and prognosis (McCluggage, 2011; Prat, 2012). EOC is divided into
2 categories, designated type I and type II. Type I tumors are low-grade neoplasms that arise in a stepwise manner from borderline tumors, while type II tumors have no morphologically characterized precursor lesions, suggesting that they arise in a de novo manner (Shih Ie and Kurman, 2004). Malignant EOC is further classified into five main subtypes: high-grade serous (HGSOC) accounting for 70% of cases, endometrioid (ENOC) and clear cell (CCOC) each comprising 10% of cases, mucinous (MOC) accounting for 3% of cases, with mixed or undifferentiated subtypes making up the remainder of the diagnoses
(Bell, 2005). Furthermore, type I tumors, composed of ENOC, CCOC, and MOC, are correlated with molecular changes not typically found in type II tumors, such as HGSOC.
These include mutations in BRAF (Singer et al., 2003), KRAS (Mok et al., 1993), β-catenin
(Wu et al., 2001), and PTEN (Obata et al., 1998), although there is limited data regarding
HGSOC aside from frequent p53 mutations and altered BRCA expression as a result of mutation or promoter methylation, making them genetically unstable and harder to molecularly characterize (Kurman and Shih Ie, 2011). The exact cellular origin and pathophysiology or OC remains under investigation, although the majority of tumors have
71 been shown to originate from gynecological tissues other than the ovaries, which are involved after tumor cells begin metastasizing through the extracellular microenvironment, detailed in (Vang et al., 2013). In mouse models of OC where ovarian tumor-initiating cells were implanted into animals via intrauterine or intraperitoneal injection, it was shown that tumor formation was most extensive in areas of fatty deposits, such as the omentum, diaphragm, and peritoneum (Yang-Hartwich et al., 2014). However, over 30% of mice injected with fluorescently labeled OC cells were found to develop ovarian tumors which was increased to 100% following superovulation of mice with pregnant mares’ serum and human chorionic gonadotropin, demonstrating that ovulation may play an important role in OC cell homing and tumorigenicity (Yang-Hartwich et al., 2014). ENOC and CCOC have also been demonstrated to arise from endometriotic cysts (Veras et al., 2009) while
MOC tumors have been shown to originate from transitional cell nests at the tubal- mesothelial junction (Seidman and Khedmati, 2008).
Correct diagnosis of the subtype and stage of EOC is of paramount importance because each subtype has a differential response to standard-of-care therapies. Type I tumors are relatively genetically stable, are contained within the ovary at presentation, and respond well to treatment. Type II tumors, are highly invasive, aggressive, and are generally diagnosed at an advanced stage, making them particularly difficult to treat due to their dissemination within the abdominal cavity (van Zyl et al., 2018). However, despite the genetic and molecular differences of EOC subtypes, treatment remains similar and consists of initial surgical removal, or debulking, followed by six courses of platinum-based chemotherapy (Tomao et al., 2017). While most patients with advanced disease initially respond well to such a regimen, approximately 80% will experience recurrence within five
72 years of diagnosis with 70% of women relapsing within 18 months (Luvero et al., 2014).
Adding to the lethality of recurrent OC is the ability of tumor cells to develop platinum resistance. This is important for patient prognosis as platinum-sensitive OC has a median survival of two years compared to 9-12 months for platinum-resistant OC (Davis et al.,
2014). Thus, it is critical to develop novel strategies targeting both sensitive and resistant
OC cells to maximize tumor eradication while minimizing the risk of recurrence as tumor cells respond to chemotherapy.
In normal physiology, ovarian follicles of mammals are typically made up of a single oocyte surrounded by layers of somatic granulosa and theca cells, producing sex steroids, androgens, and growth factors for proper egg development (Winterhager and Kidder,
2015). Throughout folliculogenesis, GJs enable coupling of granulosa cells between themselves and the oocyte, establishing a metabolic relationship that is critical for oogenesis, (reviewed by Kidder and Vanderhyden, 2010). Animal studies have demonstrated that nutrients such as amino acids and glucose pass from granulosa cells to the oocyte through GJs, supporting the latter’s metabolic state (Winterhager and Kidder,
2015). In addition, GJIC between the ooycyte and ovarian follicular somatic cells is required for meiotic arrest prior to ovulation to reduce the number of chromosomes from diploid to haploid (Wigglesworth et al., 2013).
As GJs are composed of connexin family proteins, several including Cx37 and Cx43 have been consistently identified in growing and mature follicles (Gershon et al., 2008).
Additional subunits have been detected but their roles in oogenesis are unclear as knockout studies reveal that Cx32-/- females remained fertile (Nelles et al., 1996) while Cx26-/- animals were embryonic lethal (Gabriel et al., 1998). However, through genetic ablation
73 strategies, Cx37 and Cx43 were found to be essential for early follicle growth (Kidder and
Vanderhyden, 2010). Cx37 was found to form GJs at the surface of the oocyte enabling ooctye-granulosa coupling (Simon et al., 1997) while Cx43 was found to facilitate GJIC between the granulosa cells themselves (Gittens and Kidder, 2005). However, given their importance for normal oocyte development, the role for connexins in OC has remained undefined. The upregulation of cytoplasmic, rather than membrane-bound, Cx32 was found to be correlated with cisplatin resistance. Furthermore, knockdown of Cx32 was found to resensitize cisplatin resistant ovarian cancer cells, demonstrating a role for connexins in chemoresistance (Wu et al., 2017). Cx43 was also found to be upregulated by epidermal growth factor (EGF) in ovarian cancer cells. Knockdown of Cx43 increased basal and EGF-induced cell proliferation while Cx43 overexpression had the opposite effect (Qiu et al., 2016). However, treatment with a pan GJ–inhibitor, carbenoxolone, did not alter Cx43-mediated suppressive effects, indicating a GJ-independent mechanism in ovarian cancer (Qiu et al., 2016). Thus GJIC and its role in OC remains relatively undefined. Given that ovarian cancers often present as histologically dense entities, it is reasonable to speculate that intercellular communication occurs to potentially confer survival advantages for tumor cells. In the current study, we demonstrate that GJIC communication takes place between ovarian cancer cells and that inhibition of the process is detrimental for tumor cell survival.
74
3.2 Materials and Methods
3.2.1 Cell Culture and Preparation of Culture Medium
Six different in vitro systems were used to model ovarian cancer. OV81 (cisplatin sensitive) and CP10 (cisplatin resistant) are patient-derived xenografts representing high grade serous ovarian cancer (Nagaraj et al., 2015). A2780 (cisplatin sensitive) and CP10 (cisplatin resistant) cells are established endometrioid cell lines that are commercially available.
TOV112D cells are also representative of endometrioid ovarian cancer while OVCAR-3 are cisplatin resistant epithelial adenocarcinoma cells. OV81, CP10, A2780, and CP10 were cultured in DMEM supplemented with 10% heat-inactivated fetal bovine serum
(FBS; Sigma), L-glutamine, penicillin, streptomycin, and sodium pyruvate at 37⁰ C in a humidified atmosphere of 20% oxygen and 5% CO2. TOV112D cells were cultured in a 1:1 mixture of MCDB 105 and Medium 199 supplemented with 15% heat-inactivated FBS, penicillin, and streptomycin. OVCAR-3 cells were cultured in RPMI 1640 supplemented with 10% heat-inactivated FBS, penicillin, streptomycin, and L-glutamine.
3.2.2 Dye Transfer Assay
For dye transfer assessments, each cell type was divided into two groups. The donor group was labeled with 1 μM Calcein AM (Life Technologies), a dye that is cell-membrane permeable in its parent form that is then converted to a green-fluorescent Calcein after acetoxymethyl ester hydrolysis by intracellular esterases. After conversion the dye is no longer cell-permeable and can only leave cells by GJs, hemichannels, or exocytosis. The
Calcein-labeled group was also exposed to 3 μM DiI (Life Technologies), a lipophilic membrane dye that stains the entire cell and does not spread. Cells were labeled for 45
75 minutes with each dye according to the manufacturer’s instructions before being centrifuged to remove excess dye. The double-labeled group was then mixed at a 1:1 ratio with adherent non-labeled cells and incubated for 1, 2, 3, and 4 hours to allow for cell-cell coupling. After incubation cells were analyzed by flow cytometry using a BD Fortessa to quantify the percentage of un-labeled acceptor cells acquiring green Calcein AM from
DiI/Calcein-labeled donor cells. Single DiI-and Calcein-stained control cells were used to verify dye efficacy and set flow cytometry gates. After functional GJ was confirmed, OV81 and CP10 cells were incubated in the same manner in the presence of 0.1 μM, 1.0 μM, 5.0
μM, and 10 μM mefloquine for 3 hours and dye transfer was likewise analyzed by flow cytometry to validate GJ inhibition. Dye transfer experiments analyzed six cell lines
(A2780, CP70, OV81, CP10, TOV112D, and OVCAR-3) while dye transfer inhibition experiments utilized OV81 and CP70. For all dye transfer experiments, 10,000 cells were collected from a total of 6.0x105 of each biological sample (one vehicle control, four mefloquine concentrations, as well as single Calcein and DiI controls for each concentration). To visualize dye transfer, OV81 and CP10 cells were labeled and incubated as indicated above and imaged using an AMG EVOs f1 digital inverted microscope with
GFP and RFP filters to track donor (DiI and Calcein double-labeled) and acceptor (Calcein- labeled) cells at 10X magnification.
3.2.3 Transmission Electron Microscopy
OV81 and CP10 cells were grown on plastic chamber slides (Lab-Tek®) until they reached approximately 90% confluency. Cells were washed in 1X phosphate buffered saline (PBS) at 37⁰C for five minutes. Cells were then fixed with a 4% paraformaldehyde/2.5% glutaraldehyde in 0.1M sodium cacodylate for five minutes at 37⁰C followed by overnight
76 fixation at 4⁰C. Post fixation, cells were incubated with 1% Osmium Tetroxide in H2O for
60 minutes in 40 C and subsequently washed in sodium cacodylate buffer 2 times for 5 minutes each. Cells were rinsed with maleate buffer which was then changed to 1% uranyl acetate in maleate buffer for 60 minutes. Uranyl acetate was then removed and cells were once again washed in maleate buffer 3 times for 5 minutes each. Cells were then dehydrated by washing in 30%, 50%, 75%, and 90% cold ethanol for 5 minutes. Afterwards they were placed in a 100% ethanol solution for 10 minutes, followed by incubation with propylene oxide 3 times for 15 minutes. For infiltration the propylene oxide was removed and cells were treated with a 1:1 propylene oxide/Eponate 12 medium at room temperature overnight. Pure Eponate 12 medium was then used to incubate cells for 4-6 hours at room temperateure. Ultra-thin 85nm sections where then cut with a diamond knife, stained with uranyl acetate and lead citrate, and observed with a Tecnai G2 electron microscope operated at 60 kV. OV81 and CP10 cell junctions were captured at a magnification of x30000.
3.2.4 Proliferation Assay
To quantify ovarian cancer cell proliferation, a CellTiter-Glo™ proliferation assay
(Promega) and direct cell counts, were utilized. Briefly, 1000 cells were plated in clear- bottom 96-well plates in 100 μL of culture media. A2780, CP70, OV81, CP10, TOV112D, and OVCAR-3 were treated with 2.5 μM, 5 μM, and 10 μM mefloquine or 0.001 μM, 0.005
μM 0.01 μM , 0.10 μM , 1.0 μM , 5.0 μM, 10 μM, or 25 μM cisplatin for 72 hours. Cells were allowed to recover for at least 30 minutes before the first measurement was taken, to serve as a loading control. Luminescence readings were taken 30 minutes after applying
CellTiter-Glo™ reagent. The luminescence intensity of each well was measured using a
77 luminometer (PerkinElmer) and each sample was run in triplicate to evaluate variability.
To prepare relative growth curves, data was normalized to the vehicle control. To validate the CellTiter-Glo™, assay direct cell counts were also utilized. Briefly, 5.0x104 A2780,
CP70, OV81, and CP10 cells were plated in clear-bottom 6-well plates. Cells were then treated with 2.5 μM, 5 μM, and 10 μM for 72 hours and counted using a TC20™ Automated
Cell Counter (BioRad).
3.2.5 Apoptosis Assay
To measure cellular apoptosis, a Caspase-Glo® 3/7 Assay system was utilized. Briefly,
1000 A2780, CP70, OV81, CP10, TOV112D, and OVCAR-3 cells were plated in 96-well clear bottom plates. Cells were treated with 2.5 μM, 5 μM, and 10 μM mefloquine for 24 hours. Luminescence readings were taken 60 minutes after applying Caspase-Glo reagent.
The luminescence intensity of each well was measured using a luminometer (PerkinElmer) and each sample was run in triplicate to evaluate variability. To quantify relative caspase activity, data was normalized to Cell-Titer Glo™.
3.2.6 Statistical Analysis
Data are represented as mean values +/- standard deviation. Statistical significance was analyzed using Student’s T-test or one way ANOVA where appropriate. P values less than
0.05 where considered statistically significant. p-values: * ≤ 0.05, ** ≤ 0.01, ***≤ 0.001,
****≤ 0.0001.
78
3.3 Results
3.3.1 Cisplatin Sensitivity Validation
To ensure that A2780 and OV81 cell lines were cisplatin sensitive while their counterparts
CP70 and CP10, respectively were cisplatin resistant, proliferation assays were carried out to measure their IC50 values. After three days of treatment with cisplatin, cell growth was measured via CellTiter-Glo™. A2780 cell were found to have an IC50 value of 0.74 μM while CP70 cells had a value of 5.42 μM. OV81 cells had a similar IC50 of 1.25 μM while
CP10 cells had a value of 6.95 μM, validating that each of the cell lines utilized were cisplatin sensitive or cisplatin resistant (Fig. 17)
79
Figure 17 Sinyuk et al. In preparation Figure 17
OV81 IC50: 1.25µM CP10 IC50: 6.95 µM 1.2
0.8
Relative ATP Relative 0.4
0
-2 -1 0 1 2 log[Cisplatin], μM
A2780 IC50: 0.74µM CP70 IC50: 5.42 µM
1.2
0.8
0.4 Relative ATP Relative
0 -2 -1 0 1 2 log[Cisplatin], μM
Figure 17: A2780/OV81 cells were cisplatin sensitive while CP70/CP10 cells are cisplatin resistant. After 3 days of treatment with different concentrations of cisplatin, OV81 and A2780 cells displayed increased chemosensitivity when compared to CP70 and CP10 cells, their cisplatin resistant counterparts. IC50 values are reported for each cell type.
80
3.3.2 Ovarian Cancer Cells are Capable of GJIC
To assess whether direct cell-cell communication was present in ovarian cancer cells, dye transfer assays were utilized as a surrogate. After Calcein AM and DiI-labeled donor cells were mixed with unlabeled acceptor cells, flow cytometry analysis detected a mean rate of
28.38±21.83% with a range of 52.8% across all cell lines, indicating that short-term dye transfer was variable across each cancer cell line. After four hours of co-incubation dye transfer was increased to a mean of 90.78±4.96% with a range of 13.0% across all tested cell lines, demonstrating that transfer is a time-dependent process (Fig. 18). As each cell line was distinct, the variability in the kinetics of dye transfer was not unexpected and was likely a result of the stochastic nature of GJIC. Additionally, as cells interacted with each other for longer periods of time, calcein was able to spread to a greater number of cells when visualized via fluorescence microscopy (Fig. 19). Thus, while the direct percentage of dye transfer was variable, all tested ovarian cancer cells lines were found to be capable of coupling. Further when ovarian cell junctions were investigated using transmission electron microscopy, GJ-like structures were identified between adjacent cells (Fig. 20), demonstrating that dye transfer was a result of GJIC between adjacent cells.
81
Figure 18 Sinyuk et al. In preparation
100 OV81
CP10
80 A2780
CP70 60 OVCAR-3
TOV112-D 40
20 % Dye Transfer Dye %
0 Unstained 1 2 3 4 Time (Hours)
Figure 18: Ovarian cancer cell are capable of coupling and dye transfer. Cisplatin sensitive and cisplatin resistant cells lines were capable of communication, quantified by the spread of green Calcein AM from DiI and Calcein AM double-labeled donor cells to unlabeled acceptor cells. Dye transfer was measured by flow cytometry at 1, 2, 3, and 4 hours. At 1 hour the mean percentage across all cell lines was 28.38±21.83% with a range of 52.8%. At 2 hours the mean percentage increased to 44.14±21.38 with a range of 52.8%. At 3 hours the mean percentage was 67.15±19.39 with a range of 53.1%. At 4 hours the mean percent of dye transfer increased to 90.78±4.96% with a range of 13%.
82
Figure 19 Sinyuk et al. In preparation
OV81 CP10
Figure 19: Dye transfer validation. Cisplatin sensitive (OV81) and cisplatin resistant (CP10) cells lines were each divided into two groups. One was double-labeled with DiI and Calcein, while the other remained unlabeled. Both groups were then co-incubated for three hours and dye transfer was visualized. Double positive (Red and Green) cells are donors while single positive (Green) cells are acceptors.
83
Figure 20 Sinyuk et al. In preparation
OV81 CP10
Figure 20: Transmission electron microscopy of cell junctions between adjacent ovarian cancer cells. Cisplatin sensitive (OV81) and cisplatin resistant (CP10) cell junctions were imaged at x30000 magnification. Yellow arrows indicate GJ-like structures between adjacent cells.
84
3.3.3 Inhibition of GJIC Communication
When GJIC communication was blocked pharmacologically by mefloquine, an anti- malarial agent with GJ inhibition properties, dye transfer was inhibited in OV81 and CP10 cells at 5 μM and 10 μM (Fig. 21), suggesting that GJIC is a targetable cellular process.
Indeed when all of the tested ovarian cancer cell lines were treated with mefloquine for three days, a decrease in cell growth was seen using both CellTiter-Glo™ and direct cell counts (Fig. 22 and Fig. 23). To interrogate whether the decrease in cell growth was associated with increased cell death, apoptosis was measured via caspase 3/7 activity after
24 hours of mefloquine treatment. In all ovarian cancer cells utilized, apoptosis was significantly increased at 5 and 10 μM, demonstrating that blocking GJIC could induce cell death (Fig. 24). Importantly, it should be noted that all of the ovarian cells responded to mefloquine in a similar manner, regardless of cisplatin sensitivity, suggesting that mefloquine could be a promising adjuvant therapeutic for recurrent ovarian cancer.
However, more work is necessary to elucidate the mechanisms by which mefloquine is killing cells and animal studies are required to investigate its in vivo efficacy against ovarian cancer.
85
Figure 21 Sinyuk et al. In preparation
OV81 CP10
** 1.5 1.5 ** * ***
1.0 1.0
0.5 0.5
Relative dye transfer dye Relative Relative dye transfer dye Relative 0.0 0.0 Vehicle 0.1 µM 1 µM 5 µM 10 µM Vehicle 0.1 µM 1 µM 5 µM 10 µM
Figure 21: Gap junction inhibition with mefloquine. Dye transfer assays were utilized with cisplatin sensitive (OV81) and cisplatin resistant (CP10) cell lines. Prior to co- incubation cells were treated with different concentrations of mefloquine to inhibit GJs. Dye transfer was quantified by flow cytometry. Both cell lines showed significant reduction of Calcein AM transfer at 5 and 10 μM. p-values: * ≤ 0.05, ** ≤ 0.01, ***≤ 0.001, ****≤ 0.0001.
86
Figure 22 Sinyuk et al. In preparation
A2780 OV81 TOV112D **** ** **** *** * *** 1.2 1.2 * 1.2
0.8
0.8 0.8 Relative ATP Relative Relative ATP Relative 0.4 Relative ATP Relative 0.4 0.4
0.0 0.0 0.0 Vehicle 2.5 5.0 10 Vehicle 2.5 5.0 10 Vehicle 2.5 5.0 10 Mefloquine (μM) Mefloquine (μM) Mefloquine (μM)
CP70 CP10 OVCAR3 *** ** *** ** * 1.2 ** 1.2 1.2
0.8 0.8 0.8
0.4 Relative ATP Relative
0.4 ATP Relative 0.4 Relative ATP Relative
0.0 0.0 0.0 Vehicle 2.5 5.0 10 Vehicle 2.5 5.0 10 Vehicle 2.5 5.0 10 Mefloquine (μM) Mefloquine (μM) Mefloquine (μM)
Figure 22: Ovarian cancer cell growth is inhibited upon GJ inhibition. Six cell lines were treated with different concentration of mefloquine for three days. All tested cell lines displayed inhibition of growth as assayed by Cell-Titer Glo™. IC50 values for each cell line are reported. p-values: * ≤ 0.05, ** ≤ 0.01, ***≤ 0.001, ****≤ 0.0001.
87
Figure 23 Sinyuk et al. In preparation
A2780 OV81 **
*
) ) 2.0 6 2.0 6 ** * ** 1.5 * 1.5
1.0 1.0
0.5 0.5 Relative cell cell (1.0x10 number Relative Relative cell cell (1.0x10 number Relative 0.0 0.0 Vehicle 2.5 5.0 10 Vehicle 2.5 5.0 10 Mefloquine (μM) Mefloquine (μM)
CP70 CP10
**
**
) 6 ) 2.0 ** 2.0 6 ** * ** 1.5 1.5
1.0 1.0
0.5 0.5 Relative cell cell (1.0x10 number Relative
Relative cell cell (1.0x10 number Relative 0.0 0.0 Vehicle 2.5 5.0 10 Vehicle 2.5 5.0 10 Mefloquine (μM) Mefloquine (μM)
Figure 23: Mefloquine treatment decreases ovarian cancer cell proliferation. Cisplatin sensitive (A2780, OV81) and cisplatin resistant (CP70, CP10), were plated at
equal numbers (5.0x104) and treated with mefloquine for three days. After treatment, cells were directly counted to assess total number after GJ inhibition. p-values: * ≤ 0.05, ** ≤ 0.01, ***≤ 0.001, ****≤ 0.0001.
88
Figure 24 Sinyuk et al. In preparation
A2780 CP70 TOV112D
*** *** *** *** *** * 30 40 8
30 6 20 20 4 10 10 2
0 0 Vehicle 2.5 10.0 0
5.0 Vehicle 2.5 5.0 10.0 Vehicle 2.5 5.0 10.0
Relative Caspase Activity (Fold Change) Caspase Activity(Fold Relative Change) Caspase Activity(Fold Relative Relative Caspase Activity (Fold Change) Caspase Activity(Fold Relative Mefloquine (μM) Mefloquine (μM) Mefloquine (μM)
OV81 CP10 OVCAR3 *** *** *** 30 150 30 *** *** 20 100 20 10 50 10 3 2 3 2 2 1 1 1 0 0 0
Vehicle 2.5 5.0 10.0 Vehicle 2.5 5.0 10.0 Vehicle 2.5 5.0 10.0 Relative Caspase Activity (Fold Change) Caspase Activity(Fold Relative
Mefloquine (μM) Change) Caspase Activity(Fold Relative Mefloquine (μM) Mefloquine (μM) Relative Caspase Activity (Fold Change) Caspase Activity(Fold Relative
Figure 24: Mefloquine treatment increases apoptosis. Six cell lines were treated with mefloquine for 24 hours. Apoptosis was measured via Caspase-Glo® 3/7 Assay. In all cell
lines 5 and 10 μM mefloquine significantly increased apoptosis. p-values: * ≤ 0.05, ** ≤ 0.01, ***≤ 0.001, ****≤ 0.0001.
89
3.3.4 Bioinformatics Analysis
To interrogate whether particular connexin subunits are involved in ovarian cancer GJIC, a bioinformatic survival analysis was conducted on all known connexin subunits using
Kaplan Meir (KM) Plotter and The Cancer Genome Atlas (TCGA) database retrieved from http://kmplot.com/analysis/index.php?p=service&cancer=ovar. Higher expression of
Cx43, Cx46, Cx62, Cx26, and Cx45 was associated with higher overall (OS) and progression-free survival (PFS). Through the Human Protein Atlas, connexin protein expression in ovarian cancer was also investigated and Cx43, Cx62, and Cx26 were found to be present in ovarian tumor tissue, summarized in Table 2. These connexins bear further scrutiny and additional functional studies are necessary to specifically target each subunit, alone and in tandem, to understand which functions, GJIC, hemichannel activity, or connexin-protein associations, are the main drivers in ovarian cancer tumorigenesis.
90
OS PFS Present in normal Present in ovarian tumor ovarian tissue tissue Database TCGA TCGA Human Protein Atlas Human Protein Atlas Connexin Gene (p value) (p value) IHC Staining Grade/Number Cx43 GJA1 0.024 0.039 Not detected Low: 5/12 Cx26 GJB2 0.00023 3.40E-08 Not detected High: 4/12 Cx62 GJA10 0.015 5.20E-07 N/A Medium: 2/11 Cx40 GJA5 0.18 1.20E-05 Ovarian stroma cells Medium: 9/11 Cx46 GJA3 9.90E-05 0.00018 Not detected Not detected Cx45 GJC1 9.40E-10 7.10E-04 Not detected N/A Cx31.3 GJC3 0.18 2.60E-04 Follicle cells High + Medium: 3/12 Cx31 GJB3 0.0088 5.20E-03 N/A Not detected Cx30.3 GJB4 0.44 2.70E-04 N/A N/A Cx31.1 GJB5 0.031 7.80E-05 Not detected Medium: 1/12 Cx30 GJB6 0.07 2.30E-02 N/A N/A Cx47 GJC2 0.04 9.10E-04 N/A N/A Cx32 GJB1 5.50E-09 1.20E-07 ND High: 7/11 Cx36 GJD2 0.056 9.70E-03 N/A Medium: 1/12 Cx31.9 GJD3 0.0037 3.50E-02 N/A N/A Cx39 GJD4 0.22 1.80E-01 Ovarian stroma cells Medium: 6/10 Cx37 GJA4 0.4 0.014 Follicle and stromal cells Low: 4/12 Cx50 GJA8 0.14 0.037 Not detected Not detected
Table II: Connexin expression in ovarian cancer and normal tissues. Bioinformatic data was analyzed using KM Plotter and the TCGA database. All known connexin subunits were interrogated to understand whether higher expression led to poorer or better prognosis. All connexins in red (bolded) represent subunits whose higher expression in ovarian cancer was associated with a poor prognosis. Connexins in blue (italics) represent high expression that is associated with a more favorable prognosis. The Human
Protein Atlas utilized a tissue micro-array based platform and immunohistochemistry to visualize connexin protein expression in normal ovarian tissue and ovarian tumor tissue. An IHC staining grade from low to high was further established depending on protein expression levels of stained tissues.
91
3.4 Discussion
Targeting GJIC in cancer has the potential to be an exciting targeting strategy with easily translatable results (Trosko and Ruch, 2002). In leukemia targeting GJs, and Cx25 in particular, decreased cancer cell growth and sensitized cells to chemotherapy (Sinyuk et al., 2015). Thus GJ inhibition could also confer a survival advantage when used alongside chemotherapy. However, it is important to understand that mefloquine does not have specificity to inhibit individual connexin subunits. While it has been demonstrated that mefloquine can be used to inhibit Cx36 and Cx50 at lower concentrations than Cx43, Cx32, or Cx26 in transfected mouse neuroblastoma (N2A) cells (Cruikshank et al., 2004), this does not exactly equate to drug specificity and can suggest that Cx36 and Cx50 are more sensitive to mefloquine as other connexins could also be blocked, albeit at higher concentrations of mefloquine. In addition, junctional currents were utilized as a means to quantify GJ blockade and there was no evidence that functional communication between cells was inhibited.
Here, we have shown that ovarian cancer cells directly communicate with one another, and that this process could be inhibited by treatment with mefloquine. We have also detected
GJ-like structures between adjacent cells, suggesting that communication is occurring via
GJs. When mefloquine was used to treat ovarian cancer cells, cell growth was suppressed at mefloquine concentrations above 2.5 μM, as demonstrated by ATP assays (Fig. 22) and direct cell counts (Fig. 23). Apoptosis was also increased in ovarian cancer cells in response to 5 or 10 μM mefloquine, demonstrating that the drug was actively killing cells and not simply halting their proliferation (Fig. 24). Toxicity studies are thus vital to find effective anti-cancer doses while minimizing the neurological side-effects seen in humans who
92 receive the drug as a prophylactic for malaria. Nonetheless, mefloquine and its GJIC inhibitory properties show promise in ovarian cancer and should be pursued as an additive strategy in advanced disease for which few additional treatment options are currently available.
While inhibiting GJIC is a promising area of investigation, several caveats must be addressed. It is important to note that blocking all GJs may have unintended off-target effects that need to be addressed before considering clinical trials. Additionally, the exact methods by which mefloquine inhibits connexin function is an ongoing area of investigation. Two other pharmacological agents with GJ inhibition activity, CBX and 1-
Octanol, are thought to act on cell membranes to alter fluidity and disrupt the transmembrane domains of connexin proteins, rendering them inert. However, this explanation has yet to be fully investigated and remains speculative. Mefloquine could also act on cell membranes or it may have additional mechanisms for inhibiting GJIC that have not yet been described.
The last, and possibly most important, point to consider regarding GJ inhibition is the exact mechanism behind ovarian cell death after treatment with mefloquine. Several likely explanations for this phenomenon have therefore been proposed. As previously mentioned, ovarian cancers cells exist in a closely packed microenvironment and communicate predominantly through cell-cell contact mediated by GJs. As such, tumor cells are better able to respond to external stimuli and escape damage from sources such as chemotherapeutics and radiation by exchanging information and rendering themselves less susceptible to perturbation. In addition, GJs may allow for the release of potentially lethal intercellular components, such as reactive oxygen species (ROS), generated in response to
93 cell damage. Conversely, GJ hemichannels may also facilitate the uptake of molecules necessary to protect tumor cells from ROS-induced DNA damage. Work in normal hematopoietic stem cells (HSCs) has supported this concept, as Cx43 deficient HSCs displayed decreased survival and increased senescence as a direct result of their inability to transfer ROS to the hematopoietic microenvironment following myeloablation, demonstrating that Cx43 is able to play a protective role during stressful conditions such as hematopoietic recovery (Taniguchi et al., 2012).
Even though the exact molecular mechanisms behind connexin signaling are only now beginning to be elucidated, the potential to disrupt GJs, and consequently tumor cells, by pharmacologically targeting connexins remains an attractive strategy in a field that has had limited clinical success over the past decades. However it is prudent to consider that connexins may additional functions which have yet to be fully described. To this end, cytoplasmic partners have been thought to be capable of interacting with the intracellular domains of connexin proteins, providing a potential means of specifically targeting individual subunits. The ablation of one universal connexin may have unintended secondary effects or no effects at all, as compensatory mechanisms likely exist among various connexin proteins. Rather, GJ inhibition strategies should be contextualized in light of the overall tumor or. Of paramount importance is the development of novel mimetic peptides or agents capable of disrupting individual connexin subunits to minimize the harm done to normal tissue in the course of treatment. Targeting specific connexin subunits, depending tumor subtype, is can be complementary for therapy. Additional work is necessary to tease out the direct molecular mechanisms responsible for connexin signaling, but efforts are slowly beginning to concentrate on this line of inquiry. With careful
94 investigation, elucidating connexin signaling and GJIC has the potential to make a transformative impact for the development of therapies capable of improving the outcome of patients diagnosed with not only ovarian cancer but also other neoplasias for which little hope currently exists.
95
CHAPTER IV
CONCLUDING REMARKS3
In the past 50 years, remarkable work has been conducted to investigate how connexins function in cell-cell communication, hemichannel activity, and other activities unrelated to
GJIC. Proper cellular adhesion and communication are necessary for the development of multicellular life, without which it would not be possible to coordinate larger-scale behavior and cellular responses. Unlike other junctional molecules, connexins help enable tissue organization and mediate the transfer of signals among cells during development, the maintenance of homeostasis, and pathology. Thus, understanding their biology and regulatory mechanisms at the transcriptional and translational levels will enable additional exploration regarding their three main functions.
In fact, connexins may have additional thus-far undefined roles in cellular development, differentiation, and physiology outside of the scope of this review. Likewise, it is critical
3 A part of this chapter has been submitted for publication in Frontiers in Oncology under the title "Cancer Connectors: Connexins, Gap Junctions, and Communication." Authors: Maksim Sinyuk, Erin Mulkearns-Hubert, Ofer Reizes, and Justin Lathia. Submission Date: September 15, 2018.
96 to remember that connexins do not exist in a vacuum. Rather, their characterization should be considered in temporal and tissue-specific contexts. While one connexin subunit may be required for the growth of a particular tissue, it should not be assumed that this will hold true throughout the lifetime of the organism.
Moreover, the recent push to identify GJ-independent functions of connexins, whether they take the form of hemichannels or connexin-protein interactions, has given rise to novel questions about their role in normal cell physiology. However, their complexity and overall tissue distribution has made it difficult to fully elucidate their function in human development. Nonetheless, with more powerful genetic and pharmacological approaches, it may soon be possible to track their behavior over time. It should also be mentioned that while there is a wide range of knowledge about the specific types of molecules that are shuttled via connexins, as of yet, it is difficult to definitively state how individual connexin subunits are selectively permeable to particular signals. However, that is not to say that such efforts are fruitless. In fact, this should be a larger area of investigation to potentially identify novel means to deliver drugs, chemotherapy, or other pharmacological agents into cells via GJs. While there is still a great deal to discover regarding GJIC in cancer, advances are rapidly gaining momentum to answer such questions.
While connexins have been described to play three main roles in cancer, namely GJIC, hemichannel activity, and connexin-protein interactions, this dissertation has mainly focused on the effects of inhibition of communication between individual tumor cells. This is not meant to exclude or minimize the importance of the latter two functions, but rather demonstrates the need for cancer cells to couple and exchange information regarding environmental and internal conditions to better survive and proliferate.
97
Investigating connexin function and GJIC in leukemia was informative for the identification of individual subunits associated with tumor growth and chemosensitivity in hematological malignancies. However, an additional cancer model, ovarian, was also used to characterize the role of connexins and GJIC in solid tumors. Additionally, a hallmark of ovarian cancer is the effective initial response to stand-of-care modalities, in the form of platinum-based therapies, followed by the recurrence of disease with a large population of tumor cells developing resistance to chemotherapy. To achieve higher survival rates in advanced ovarian cancer, it is critical to develop adjuvant treatment strategies which eliminate chemotherapy-sensitive and chemotherapy-resistant cancer cells. As such, the inhibition of GJIC may serve as a promising novel means by which both populations can be targeted.
While mast cells and a variety of different tumor cells have had their mefloquine toxicity responses measured, summarized in Table 3, it is critical to establish how normal healthy cells respond to the drug to better inform its potential utilization in ovarian cancer. By establishing a strategy that uses mefloquine to target cisplatin-sensitive and cisplatin- resistant ovarian tumor cells, while sparing the surrounding stroma, fewer or weaker doses of chemotherapeutics may be utilized limiting the harmful side-effects of cytotoxic therapy. Further, I have shown that mefloquine can overcome and destroy cisplatin- resistant cells, demonstrating its usefulness when chemotherapy is no longer effective in advanced-stage ovarian cancer. Such a finding should give rationale for a more thorough investigation into the exact molecular mechanisms by which cisplatin-resistant cells die as quickly and effectively as their cisplatin-sensitive counterparts as a result of mefloquine and due, in part, as a result of the inhibition of GJIC.
98
Connexins and GJIC are becoming increasingly attractive targets for cancer therapy as their three canonical functions become better defined. Arguably, the most advantageous feature of anti-connexin strategies is the ubiquitous nature of the proteins. Most normal cells of the body require the ability to communicate in order to carry out tissue and organ-level functions that require precise regulation and rapid response to changing local and systemic conditions. Thus, connexin trafficking and turnover are tightly controlled, as suggested by their rapid half-lives. Such properties of connexins and GJIC enable cells to quickly generate correct connections with their neighbors or microenvironment and coordinate large-scale actions that would not be possible on a single-cell level. Moreover, tumor cells are also able to co-opt such function to facilitate their sustained growth. Intercellular communication is especially important in the context of cancer because tumors should not be thought of as simply an amalgamation of rapidly proliferating cells but rather as discrete entities that are able to manipulate their microenvironment to create conditions that are more conducive for survival.
99
Cell type Mefloquine Toxicity References Concentrations (μM)
Mast Cells Paivandy et al., 2014. Bone-marrow derived mast cells 8.8 ± 1.9 Peritoneal cell-derived mast cells 3.0 ± 0.4 Human mast cell line 14.2 ± 0.6 Cord-blood derived mast cells 16.9 ± 0.3
Leukemia Sukhai et al., 2013 U-937 9.2 THP1 11.4 K562 7.85
GBM Geng et al., 2010 U87 10.0
Breast Cancer Sharma et al., 2012 MDA-MB-231 cells 2.5-5.0 MCF7 5.0-10.0
Prostate Cancer Yan et al., 2013. PC3 5.0-40.0 DU145 10.0-20.0
Non-Tumor Thompson et al., 2007 HEK-293 9.36
Table III. Mefloquine IC50 values as measured in different cell lines
100
The connexin family is composed of over 20 distinct subunits and tumors likely express only a few isoforms for the facilitation of GJIC, hemichannel activity, or connexin-protein associations. Thus, with proper investigation, including bioinformatics and functional studies, connexins could serve as tumor markers based on the pattern of expression on cancer cells. Additionally, their conserved transmembrane regions make connexins and
GJs sensitive to a wide variety of different pharmacological inhibitors. Currently several pan-GJ inhibitors, including some that are FDA-approved for unrelated conditions, are widely utilized, including carbenoxolone, 1-octanol, mefloquine, halothane, histamine, and others (as reviewed in Salameh and Dhein, 2005). However, while these agents are known to be capable of disrupting GJIC and hemichannel function, it is more difficult to understand how connexin-protein interactions are affected due to the intracellular C- terminal domain. As such, whereas blocking certain GJs on tumor cells could have positive consequences in terms of hindering proliferation and growth, deleterious side-effects may also occur as a result of inhibition of GJIC on normal cells. As such, it is necessary to explore peptides that are able to inhibit specific connexins to selectively target tumor- specific proteins while sparing those utilized by healthy, non-cancerous cells. While some connexin-specific peptides such as Gap19 for Cx43, Gap27/40 for Cx40, and Gap 24 for
Cx32 (reviewed by Evans and Leybaert, 2007) exist, their specificity and mechanism of action are still under scrutiny. Likewise, their efficacy in animal settings is still being elucidated, and it is necessary to better understand their pharmacology, toxicity, and anti- tumor function.
Lastly, it is important to consider which of the three aforementioned functions of a connexin is most appropriate for targeting strategies. Blocking or otherwise inhibiting
101
GJIC and hemichannel function is the most straightforward method, as it forces cells to remain in relative isolation and unable to respond to harmful internal or external conditions.
Interestingly it has been found that a Cx43-specific peptide, L2, is capable of keeping Cx43
GJs in an open state while inhibiting hemichannel opening to investigate the therapeutic potential of counteracting excessive activity without disrupting GJIC (Iyyathurai et al.,
2013). Thus, such tools will enable the development of strategies towards specific connexin functions in a more personalized manner. This is most evident when CSC or tumor cell communication is inhibited prior to chemotherapy administration, increasing their sensitivity to current standard-of-care practices, likely as a result of ROS accumulation or the inability to shuttle out toxic molecules (Mulkearns-Hubert et al., 2018). However, inhibiting connexin-protein interactions, while arguably more difficult, could lead to more specific anti-tumor approaches, as only certain proteins are able to associate with connexins within particular neoplasms (Thiagarajan et al., 2018). Additionally, blocking cell-cell contacts could affect critical downstream signaling pathways including PKC, MAPK, and
Src to induce tumor cell death while sparing normal cells lacking the connexin-protein complex. Blocking connexin biosynthesis or trafficking could also represent valid anti- tumor strategies, although this would still lead to GJ, hemichannel, or connexin-protein interaction dysfunction. Thus, in order to consider how connexins may be utilized as potential therapeutic targets in cancer, one must take into account their functional diversity and cellular specificity. Whereas targeting GJIC in one tumor type could limit proliferation or induce death, the same cannot be assumed to be true across different malignancies. The same concept applies to hemichannel activity and connexin-protein interactions, although they do allow for more specificity when targeted. Moreover, it should also be remembered
102 that targeting a single connexin subunit may simply not be sufficient to cause tumor cell death, as different subunits are able to compensate for the loss of a single group of channels.
Thus, careful consideration is warranted when designing connexin and GJIC-mediated therapy. However, this is not cause to abandon targeting intercellular communication in cancer but rather a promising area of development. Ideally, inhibiting connexin function will synergize with current standard-of-care therapies to enable treatment where other options are not available. Thus, connexin function is a crucial, multifaceted process that may enable next-generation anti-tumor modalities through GJIC, hemichannel activity, and connexin-protein associations.
103
REFERENCES
(2013). Genomic and Epigenomic Landscapes of Adult De Novo Acute Myeloid
Leukemia. New England Journal of Medicine 368, 2059-2074.
A., G.J., and L., P.D. (1994). Expression of gap junction proteins Cx26, Cx31.1, Cx37,
and Cx43 in developing and mature rat epidermis. Developmental Dynamics
200(1), 1-13. doi: doi:10.1002/aja.1002000102.
Aasen, T., Mesnil, M., Naus, C.C., Lampe, P.D., and Laird, D.W. (2016). Gap junctions
and cancer: communicating for 50 years. Nature Reviews Cancer 16, 775. doi:
10.1038/nrc.2016.105
Allison, D.W., Ohran, A.J., Stobbs, S.H., Mameli, M., Valenzuela, C.F., Sudweeks, S.N.,
Ray, A.P., Henriksen, S.J., and Steffensen, S.C. (2006). Connexin-36 gap
junctions mediate electrical coupling between ventral tegmental area GABA
neurons. Synapse 60, 20-31.
Allemani, C., Weir, H.K., Carreira, H., Harewood, R., Spika, D., Wang, X.S., Bannon, F.,
Ahn, J.V., Johnson, C.J., Bonaventure, A., Marcos-Gragera, R., Stiller, C.,
Azevedo E Silva, G., Chen, W.Q., Ogunbiyi, O.J., Rachet, B., Soeberg, M.J.,
You, H., Matsuda, T., Bielska-Lasota, M., Storm, H., Tucker, T.C., and Coleman,
M.P. (2015). Global surveillance of cancer survival 1995-2009: analysis of
individual data for 25,676,887 patients from 279 population-based registries in 67
countries (CONCORD-2). Lancet 385, 977-1010.
Ambrosi, C., Ren, C., Spagnol, G., Cavin, G., Cone, A., Grintsevich, E.E., et al. (2016).
Connexin43 Forms Supramolecular Complexes through Non-Overlapping
104
Binding Sites for Drebrin, Tubulin, and ZO-1. PLoS One 11(6), e0157073. doi:
10.1371/journal.pone.0157073.
Azzam, E.I., de Toledo, S.M., and Little, J.B. (2001). Direct evidence for the
participation of gap junction-mediated intercellular communication in the
transmission of damage signals from α-particle irradiated to nonirradiated cells.
Proceedings of the National Academy of Sciences of the United States of America
98(2), 473-478.
Balendiran, G.K., Dabur, R., and Fraser, D. (2004). The role of glutathione in cancer.
Cell Biochem Funct 22(6), 343-352. doi: 10.1002/cbf.1149.
Barr, T.P., Albrecht, P.J., Hou, Q., Mongin, A.A., Strichartz, G.R., and Rice, F.L. (2013).
Air-Stimulated ATP Release from Keratinocytes Occurs through Connexin
Hemichannels. PLoS ONE 8, e56744.
Batlle, E., and Clevers, H. (2017). Cancer stem cells revisited. Nature Medicine 23, 1124.
doi: 10.1038/nm.4409.
Beauchamp, P., Desplantez, T., Mccain, M.L., Li, W., Asimaki, A., Rigoli, G., Parker,
K.K., Saffitz, J.E., and Kleber, A.G. (2012). Electrical Coupling and Propagation
in Engineered Ventricular Myocardium With Heterogeneous Expression of
Connexin43. Circulation Research 110, 1445-1453.
Becker, D.L., Phillips, A.R., Duft, B.J., Kim, Y., and Green, C.R. (2016). Translating
connexin biology into therapeutics. Seminars in Cell & Developmental Biology
50, 49-58. doi: https://doi.org/10.1016/j.semcdb.2015.12.009.
Bell, D.A. (2005). Origins and molecular pathology of ovarian cancer. Mod Pathol 18
Suppl 2, S19-32.
105
Bell, S., Kolobova, I., Crapper, L., and Ernst, C. (2016). Lesch-Nyhan Syndrome:
Models, Theories, and Therapies. Molecular Syndromology 7(6), 302-311.
Belliveau, D.J., Bani-Yaghoub, M., McGirr, B., Naus, C.C., and Rushlow, W.J. (2006).
Enhanced neurite outgrowth in PC12 cells mediated by connexin hemichannels
and ATP. J Biol Chem 281(30), 20920-20931. doi: 10.1074/jbc.M600026200.
Belousov, A.B., Fontes, J.D., Freitas-Andrade, M., and Naus, C.C. (2017). Gap junctions
and hemichannels: communicating cell death in neurodevelopment and disease.
BMC Cell Biol 18(Suppl 1), 4. doi: 10.1186/s12860-016-0120-x.
Bergoffen, J., Scherer, S., Wang, S., Scott, M., Bone, L., Paul, D., et al. (1993). Connexin
mutations in X-linked Charcot-Marie-Tooth disease. Science 262(5142), 2039-
2042. doi: 10.1126/science.8266101.
Boengler, K., Schulz, R., and Heusch, G. (2006). Connexin 43 signalling and
cardioprotection. Heart 92(12), 1724-1727. doi: 10.1136/hrt.2005.066878.
Bone, L.J., Deschênes, S.M., Balice-Gordon, R.J., Fischbeck, K.H., and Scherer, S.S.
(1997). Connexin32 and X-linked Charcot–Marie–Tooth Disease. Neurobiology
of Disease 4(3), 221-230. doi: https://doi.org/10.1006/nbdi.1997.0152.
Boyerinas, B., Zafrir, M., Yesilkanal, A.E., Price, T.T., Hyjek, E.M., and Sipkins, D.A.
(2013). Adhesion to osteopontin in the bone marrow niche regulates
lymphoblastic leukemia cell dormancy. Blood 121, 4821-4831.
Brownlee, C. (2002). Role of the extracellular matrix in cell–cell signalling: paracrine
paradigms. Current Opinion in Plant Biology 5(5), 396-401. doi:
https://doi.org/10.1016/S1369-5266(02)00286-8.
106
Bruzzone, S., Franco, L., Guida, L., Zocchi, E., Contini, P., Bisso, A., et al. (2001). A
self-restricted CD38-connexin 43 cross-talk affects NAD+ and cyclic ADP-ribose
metabolism and regulates intracellular calcium in 3T3 fibroblasts. J Biol Chem
276(51), 48300-48308. doi: 10.1074/jbc.M107308200.
Chandrasekhar, A., and Bera, A.K. (2012). Hemichannels: permeants and their effect on
development, physiology and death. Cell Biochemistry and Function 30(2), 89-
100. doi: doi:10.1002/cbf.2794.
Chen, B.-P., Mao, H.-J., Fan, F.-Y., Bruce, I.C., and Xia, Q. (2005). Delayed Uncoupling
Contributes to the Protective Effect of Heptanol Against Ischaemia in the Rat
Isolated Heart. Clinical and Experimental Pharmacology and Physiology 32(8),
655-662. doi: doi:10.1111/j.0305-1870.2005.04246.x.
Chadjichristos, C.E., Scheckenbach, K.E.L., Van Veen, T.a.B., Richani Sarieddine, M.Z.,
De Wit, C., Yang, Z., Roth, I., Bacchetta, M., Viswambharan, H., Foglia, B.,
Dudez, T., Van Kempen, M.J.A., Coenjaerts, F.E.J., Miquerol, L., Deutsch, U.,
Jongsma, H.J., Chanson, M., and Kwak, B.R. (2010). Endothelial-Specific
Deletion of Connexin40 Promotes Atherosclerosis by Increasing CD73-
Dependent Leukocyte Adhesion. Circulation 121, 123-131.
Chen, G., Park, C.-K., Xie, R.-G., Berta, T., Nedergaard, M., and Ji, R.-R. (2014).
Connexin-43 induces chemokine release from spinal cord astrocytes to maintain
late-phase neuropathic pain in mice. Brain 137, 2193-2209.
Chen, V.W., Ruiz, B., Killeen, J.L., Cote, T.R., Wu, X.C., and Correa, C.N. (2003).
Pathology and classification of ovarian tumors. Cancer 97, 2631-2642.
107
Cherian, P.P., Siller-Jackson, A.J., Gu, S., Wang, X., Bonewald, L.F., Sprague, E., et al.
(2005). Mechanical strain opens connexin 43 hemichannels in osteocytes: a novel
mechanism for the release of prostaglandin. Mol Biol Cell 16(7), 3100-3106. doi:
10.1091/mbc.e04-10-0912.
Chi, J., Li, L., Liu, M., Tan, J., Tang, C., Pan, Q., et al. (2012). Pathogenic Connexin-31
Forms Constitutively Active Hemichannels to Promote Necrotic Cell Death.
PLOS ONE 7(2), e32531. doi: 10.1371/journal.pone.0032531.
Chou, C.H., Chen, P.-J., Lee, P.-H., Cheng, A.-L., Hsu, H.-C., and Cheng, J.C.-H. (2007).
Radiation-Induced Hepatitis B Virus Reactivation in Liver Mediated by the
Bystander Effect from Irradiated Endothelial Cells. Clinical Cancer Research
13(3), 851-857. doi: 10.1158/1078-0432.ccr-06-2459.
Colmone, A., Amorim, M., Pontier, A.L., Wang, S., Jablonski, E., and Sipkins, D.A.
(2008). Leukemic Cells Create Bone Marrow Niches That Disrupt the Behavior of
Normal Hematopoietic Progenitor Cells. Science 322, 1861-1865.
Contreras, J.E., Saez, J.C., Bukauskas, F.F., and Bennett, M.V. (2003). Functioning of
cx43 hemichannels demonstrated by single channel properties. Cell Commun
Adhes 10(4-6), 245-249.
Contreras, J.E., Sanchez, H.A., Veliz, L.P., Bukauskas, F.F., Bennett, M.V., and Saez,
J.C. (2004). Role of connexin-based gap junction channels and hemichannels in
ischemia-induced cell death in nervous tissue. Brain Res Brain Res Rev 47(1-3),
290-303. doi: 10.1016/j.brainresrev.2004.08.002.
Cox, R.P., Krauss, M.R., Balis, M.E., and Dancis, J. (1970). Evidence for Transfer of
Enzyme Product as the Basis of Metabolic Cooperation between Tissue Culture
108
Fibroblasts of Lesch-Nyhan Disease and Normal Cells. Proceedings of the
National Academy of Sciences 67(3), 1573-1579. doi: 10.1073/pnas.67.3.1573.
Crespin, S., Bechberger, J., Mesnil, M., Naus, C.C., and Sin, W.C. (2010). The carboxy-
terminal tail of connexin43 gap junction protein is sufficient to mediate
cytoskeleton changes in human glioma cells. J Cell Biochem 110(3), 589-597.
doi: 10.1002/jcb.22554.
Cronier, L., Crespin, S., Strale, P.O., Defamie, N., and Mesnil, M. (2009). Gap junctions
and cancer: new functions for an old story. Antioxid Redox Signal 11(2), 323-338.
doi: 10.1089/ars.2008.2153.
Cruikshank, S.J., Hopperstad, M., Younger, M., Connors, B.W., Spray, D.C., and
Srinivas, M. (2004). Potent block of Cx36 and Cx50 gap junction channels by
mefloquine. Proceedings of the National Academy of Sciences of the United
States of America 101, 12364-12369.
Cummins, R.J., Mothersill, C., Seymour, C.B., Johns, H., and Joiner, M.C. (1999). The
effect of microcolony size, at time of irradiation, on colony forming ability. Int J
Radiat Biol 75(2), 225-232.
Davis, A., Tinker, A.V., and Friedlander, M. (2014). "Platinum resistant" ovarian cancer:
what is it, who to treat and how to measure benefit? Gynecol Oncol 133, 624-631.
Dbouk, H.A., Mroue, R.M., El-Sabban, M.E., and Talhouk, R.S. (2009). Connexins: a
myriad of functions extending beyond assembly of gap junction channels. Cell
Communication and Signaling 7(1), 4. doi: 10.1186/1478-811x-7-4.
Decrock, E., De Vuyst, E., Vinken, M., Van Moorhem, M., Vranckx, K., Wang, N., et al.
(2009). Connexin 43 hemichannels contribute to the propagation of apoptotic cell
109
death in a rat C6 glioma cell model. Cell Death Differ 16(1), 151-163. doi:
10.1038/cdd.2008.138.
Dedek, K., Schultz, K., Pieper, M., Dirks, P., Maxeiner, S., Willecke, K., et al. (2006).
Localization of heterotypic gap junctions composed of connexin45 and
connexin36 in the rod pathway of the mouse retina. European Journal of
Neuroscience 24(6), 1675-1686. doi: doi:10.1111/j.1460-9568.2006.05052.x.
Delmar, M., Laird, D.W., Naus, C.C., Nielsen, M.S., Verselis, V.K., and White, T.W.
(2017). Connexins and Disease. Cold Spring Harbor Perspectives in Biology. doi:
10.1101/cshperspect.a029348.
DeVries, S.H., and Schwartz, E.A. (1992). Hemi-gap-junction channels in solitary
horizontal cells of the catfish retina. The Journal of Physiology 445, 201-230.
Di, W.-L., Gu, Y., Common, J.E.A., Aasen, T., O'Toole, E.A., Kelsell, D.P., et al. (2005).
Connexin interaction patterns in keratinocytes revealed morphologically and by
FRET analysis. Journal of Cell Science 118(7), 1505-1514. doi:
10.1242/jcs.01733.
Dick, J.E. (2008). Stem cell concepts renew cancer research. Blood 112(13), 4793-4807.
doi: 10.1182/blood-2008-08-077941.
Dorshkind, K., Green, L., Godwin, A., and Fletcher, W. (1993). Connexin-43-type gap
junctions mediate communication between bone marrow stromal cells. Blood
82(1), 38-45.
Ebihara, L., and Steiner, E. (1993). Properties of a nonjunctional current expressed from
a rat connexin46 cDNA in Xenopus oocytes. The Journal of General Physiology
102(1), 59-74. doi: 10.1085/jgp.102.1.59.
110
Eckardt, D., Theis, M., Degen, J., Ott, T., van Rijen, H.V.M., Kirchhoff, S., et al. (2004).
Functional role of connexin43 gap junction channels in adult mouse heart
assessed by inducible gene deletion. Journal of Molecular and Cellular
Cardiology 36(1), 101-110. doi: 10.1016/j.yjmcc.2003.10.006. el-Sabban, M.E., and Pauli, B.U. (1994). Adhesion-mediated gap junctional
communication between lung-metastatatic cancer cells and endothelium. Invasion
Metastasis 14(1-6), 164-176.
Elias, L.A., Wang, D.D., and Kriegstein, A.R. (2007). Gap junction adhesion is necessary
for radial migration in the neocortex. Nature 448(7156), 901-907. doi:
10.1038/nature06063.
Elzarrad, M.K., Haroon, A., Willecke, K., Dobrowolski, R., Gillespie, M.N., and Al-
Mehdi, A.B. (2008). Connexin-43 upregulation in micrometastases and tumor
vasculature and its role in tumor cell attachment to pulmonary endothelium. BMC
Med 6, 20. doi: 10.1186/1741-7015-6-20.
Esseltine, J.L., and Laird, D.W. (2016). Next-Generation Connexin and Pannexin Cell
Biology. Trends in Cell Biology 26(12), 944-955. doi: 10.1016/j.tcb.2016.06.003.
Evans, W H., De Vuyst, E., and Leybaert, L. (2006). The gap junction cellular internet:
connexin hemichannels enter the signalling limelight. Biochemical Journal 397(Pt
1), 1-14. doi: 10.1042/BJ20060175.
Facoetti, A., Ballarini, F., Cherubini, R., Gerardi, S., Nano, R., Ottolenghi, A., et al.
(2006). Gamma ray-induced bystander effect in tumour glioblastoma cells: a
specific study on cell survival, cytokine release and cytokine receptors. Radiation
Protection Dosimetry 122(1-4), 271-274. doi: 10.1093/rpd/ncl431.
111
Fernandez, H.F. (2010). New Trends in the Standard of Care for Initial Therapy of Acute
Myeloid Leukemia. ASH Education Program Book 2010, 56-61.
Forge, A., Becker, D., Casalotti, S., Edwards, J., Marziano, N., and Nevill, G. (2003).
Gap junctions in the inner ear: Comparison of distribution patterns in different
vertebrates and assessement of connexin composition in mammals. Journal of
Comparative Neurology 467(2), 207-231. doi: doi:10.1002/cne.10916.
Gaillard, S., Pusset, D., de Toledo, S.M., Fromm, M., and Azzam, E.I. (2009).
Propagation Distance of the α-Particle-Induced Bystander Effect: The Role of
Nuclear Traversal and Gap Junction Communication. Radiation Research 171(5),
513-520. doi: 10.1667/RR1658.1.
Garcia-Dorado, D., Ruiz-Meana, M., Padilla, F., Rodriguez-Sinovas, A., and Mirabet, M.
(2002). Gap junction-mediated intercellular communication in ischemic
preconditioning. Cardiovascular Research 55(3), 456-465. doi: 10.1016/S0008-
6363(02)00441-8.
García, I.E., Prado, P., Pupo, A., Jara, O., Rojas-Gómez, D., Mujica, P., et al. (2016).
Connexinopathies: a structural and functional glimpse. BMC Cell Biology 17(1),
S17. doi: 10.1186/s12860-016-0092-x.
Giancotti, F.G., and Ruoslahti, E. (1999). Integrin Signaling. Science 285(5430), 1028-
1033. doi: 10.1126/science.285.5430.1028.
Gershon, E., Plaks, V., and Dekel, N. (2008). Gap junctions in the ovary: expression,
localization and function. Mol Cell Endocrinol 282, 18-25.
Gillette, J.M., and Lippincott-Schwartz, J. (2009). Hematopoietic progenitor cells
regulate their niche microenvironment through a novel mechanism of cell-cell
112
communication. Commun Integr Biol 2, 305-307.
Gittens, J.E., and Kidder, G.M. (2005). Differential contributions of connexin37 and
connexin43 to oogenesis revealed in chimeric reaggregated mouse ovaries. J Cell
Sci 118, 5071-5078.
Goldberg, G.S., Lampe, P.D., and Nicholson, B.J. (1999). Selective transfer of
endogenous metabolites through gap junctions composed of different connexins.
Nat Cell Biol 1(7), 457-459.
Goldberg, G.S., Valiunas, V., and Brink, P.R. (2004). Selective permeability of gap
junction channels. Biochimica et Biophysica Acta (BBA) - Biomembranes
1662(1), 96-101. doi: https://doi.org/10.1016/j.bbamem.2003.11.022.
Goldberg, G.S., Lampe, P.D., and Nicholson, B.J. (1999). Selective transfer of
endogenous metabolites through gap junctions composed of different connexins.
Nat Cell Biol 1, 457-459.
Gómez-Hernández, J.M., de Miguel, M., Larrosa, B., González, D., and Barrio, L.C.
(2003). Molecular basis of calcium regulation in connexin-32 hemichannels.
Proceedings of the National Academy of Sciences 100(26), 16030-16035. doi:
10.1073/pnas.2530348100.
Gonzalez-Nieto, D., Li, L., Kohler, A., Ghiaur, G., Ishikawa, E., Sengupta, A., Madhu,
M., Arnett, J.L., Santho, R.A., Dunn, S.K., Fishman, G.I., Gutstein, D.E.,
Civitelli, R., Barrio, L.C., Gunzer, M., and Cancelas, J.A. (2012). Connexin-43 in
the osteogenic BM niche regulates its cellular composition and the bidirectional
traffic of hematopoietic stem cells and progenitors. Blood 119, 5144-5154.
113
Gong, X.Q., and Nicholson, B.J. (2001). Size selectivity between gap junction channels
composed of different connexins. Cell Commun Adhes 8(4-6), 187-192.
Goodenough, D.A., Goliger, J.A., and Paul, D.L. (1996). Connexins, Connexons, and
Intercellular Communication. Annual Review of Biochemistry 65(1), 475-502. doi:
10.1146/annurev.bi.65.070196.002355.
Goodenough, D.A., and Paul, D.L. (2003). Beyond the gap: functions of unpaired
connexon channels. Nature Reviews Molecular Cell Biology 4, 285. doi:
10.1038/nrm1072.
Goodenough, D.A., and Paul, D.L. (2009). Gap Junctions. Cold Spring Harbor
Perspectives in Biology 1(1). doi: 10.1101/cshperspect.a002576.
Greene, L.A., and Tischler, A.S. (1976). Establishment of a noradrenergic clonal line of
rat adrenal pheochromocytoma cells which respond to nerve growth factor. Proc
Natl Acad Sci U S A 73(7), 2424-2428.
Grek, C.L., Rhett, J.M., Bruce, J.S., Abt, M.A., Ghatnekar, G.S., and Yeh, E.S. (2015).
Targeting connexin 43 with alpha-connexin carboxyl-terminal (ACT1) peptide
enhances the activity of the targeted inhibitors, tamoxifen and lapatinib, in breast
cancer: clinical implication for ACT1. BMC Cancer 15, 296. doi:
10.1186/s12885-015-1229-6.
Gross, DB, and Jongsman, HJ. (1996). Connexins in mammalian heart function.
BioEssays 18(9), 719-730. doi: doi:10.1002/bies.950180907.
Gu, S., Yu, X.S., Yin, X., and Jiang, J.X. (2003). Stimulation of Lens Cell Differentiation
by Gap Junction Protein Connexin 45.6. Investigative Ophthalmology & Visual
Science 44(5), 2103-2111. doi: 10.1167/iovs.02-1045.
114
Gupta, V.K., Berthoud, V.M., Atal, N., Jarillo, J.A., Barrio, L.C., and Beyer, E.C. (1994).
Bovine connexin44, a lens gap junction protein: molecular cloning, immunologic
characterization, and functional expression. Investigative Ophthalmology &
Visual Science 35(10), 3747-3758.
Gutstein, D.E., Morley, G.E., and Fishman, G.I. (2001). Conditional Gene Targeting of
Connexin43: Exploring the Consequences of Gap Junction Remodeling in the
Heart. Cell Communication & Adhesion 8(4-6), 345-348. doi:
10.3109/15419060109080751.
Headrick, J.P., Hack, B., and Ashton, K.J. (2003). Acute adenosinergic cardioprotection
in ischemic-reperfused hearts. Am J Physiol Heart Circ Physiol 285(5), H1797-
1818. doi: 10.1152/ajpheart.00407.2003.
Hervé, J.-C., Bourmeyster, N., Sarrouilhe, D., and Duffy, H.S. (2007). Gap junctional
complexes: From partners to functions. Progress in Biophysics and Molecular
Biology 94(1), 29-65. doi: https://doi.org/10.1016/j.pbiomolbio.2007.03.010.
Hervé, J.-C., and Derangeon, M. (2013). Gap-junction-mediated cell-to-cell
communication. Cell and Tissue Research 352, 21-31.
Hitomi, M., Deleyrolle, Loic P., Mulkearns-Hubert, Erin E., Jarrar, A., Li, M., Sinyuk,
M., et al. (2015). Differential Connexin Function Enhances Self-Renewal in
Glioblastoma. Cell Reports 11(7), 1031-1042. doi:
Hur, K.C., Shim, J.E., and Johnson, R.G. (2003). A potential role for cx43-hemichannels
in staurosporin-induced apoptosis. Cell Commun Adhes 10(4-6), 271-277.
115
Iacobas, D.A., Scemes, E., and Spray, D.C. (2004). Gene expression alterations in
connexin null mice extend beyond the gap junction. Neurochem Int 45(2-3), 243-
250. doi: 10.1016/j.neuint.2003.12.008.
Ito, A., Katoh, F., Kataoka, T.R., Okada, M., Tsubota, N., Asada, H., et al. (2000). A role
for heterologous gap junctions between melanoma and endothelial cells in
metastasis. The Journal of Clinical Investigation 105(9), 1189-1197. doi:
10.1172/JCI8257.
Iyer, R., and Lehnert, B.E. (2000). Factors Underlying the Cell Growth-related Bystander
Responses to α Particles. Cancer Research 60(5), 1290-1298.
John, S., Cesario, D., and Weiss, J.N. (2003). Gap junctional hemichannels in the heart.
Acta Physiol Scand 179(1), 23-31. doi: 10.1046/j.1365-201X.2003.01197.x.
John, S.A., Kondo, R., Wang, S.Y., Goldhaber, J.I., and Weiss, J.N. (1999). Connexin-43
hemichannels opened by metabolic inhibition. J Biol Chem 274(1), 236-240.
Johnstone, S.R., Best, A.K., Wright, C.S., Isakson, B.E., Errington, R.J., and Martin, P.E.
(2010). Enhanced connexin 43 expression delays intra-mitoitc duration and cell
cycle traverse independently of gap junction channel function. Journal of cellular
biochemistry 110(3), 772-782. doi: 10.1002/jcb.22590.
Kalra, J., Shao, Q., Qin, H., Thomas, T., Alaoui-Jamali, M.A., and Laird, D.W. (2006).
Cx26 inhibits breast MDA-MB-435 cell tumorigenic properties by a gap
junctional intercellular communication-independent mechanism. Carcinogenesis
27(12), 2528-2537. doi: 10.1093/carcin/bgl110.
116
Kalvelyte, A., Imbrasaite, A., Bukauskiene, A., Verselis, V.K., and Bukauskas, F.F.
(2003). Connexins and apoptotic transformation. Biochem Pharmacol 66(8),
1661-1672.
Kamiya, K., Yum, S.W., Kurebayashi, N., Muraki, M., Ogawa, K., Karasawa, K., Miwa,
A., Guo, X., Gotoh, S., Sugitani, Y., Yamanaka, H., Ito-Kawashima, S., Iizuka,
T., Sakurai, T., Noda, T., Minowa, O., and Ikeda, K. (2014). Assembly of the
cochlear gap junction macromolecular complex requires connexin 26. The
Journal of Clinical Investigation 124, 1598-1607.
Kanczuga‐Koda, L., Sulkowski, S., Lenczewski, A., Koda, M., Wincewicz, A., Baltaziak,
M., et al. (2006). Increased expression of connexins 26 and 43 in lymph node
metastases of breast cancer. Journal of Clinical Pathology 59(4), 429-433. doi:
10.1136/jcp.2005.029272.
Kar, R., Batra, N., Riquelme, M.A., and Jiang, J.X. (2012). Biological role of connexin
intercellular channels and hemichannels. Archives of Biochemistry and Biophysics
524(1), 2-15. doi: https://doi.org/10.1016/j.abb.2012.03.008.
Kawasaki, Y., Omori, Y., Li, Q., Nishikawa, Y., Yoshioka, T., Yoshida, M., et al. (2011).
Cytoplasmic accumulation of connexin32 expands cancer stem cell population in
human HuH7 hepatoma cells by enhancing its self-renewal. Int J Cancer 128(1),
51-62. doi: 10.1002/ijc.25308.
Kelsell, D.P., Dunlop, J., Stevens, H.P., Lench, N.J., Liang, J.N., Parry, G., et al. (1997).
Connexin 26 mutations in hereditary non-syndromic sensorineural deafness.
Nature 387, 80. doi: 10.1038/387080a0.
Kidder, G.M., and Vanderhyden, B.C. (2010). Bidirectional communication between
117
oocytes and follicle cells: ensuring oocyte developmental competence. Can J
Physiol Pharmacol 88, 399-413.
Kihara, A.H., Santos, T.O., Osuna-Melo, E.J., Paschon, V., Vidal, K.S.M., Akamine,
P.S., Castro, L.M., Resende, R.R., Hamassaki, D.E., and Britto, L.R.G. (2010).
Connexin-mediated communication controls cell proliferation and is essential in
retinal histogenesis. International Journal of Developmental Neuroscience 28, 39-
52.
Kirchhoff, F., Dringen, R., and Giaume, C. (2001). Pathways of neuron-astrocyte
interactions and their possible role in neuroprotection. Eur Arch Psychiatry Clin
Neurosci 251(4), 159-169.
Kirchhoff, S., Nelles, E., Hagendorff, A., Krüger, O., Traub, O., and Willecke, K. (1998).
Reduced cardiac conduction velocity and predisposition to arrhythmias in
connexin40-deficient mice. Current Biology 8(5), 299-302. doi: 10.1016/S0960-
9822(98)70114-9.
Klammer, H., Mladenov, E., Li, F., and Iliakis, G. (2015). Bystander effects as
manifestation of intercellular communication of DNA damage and of the cellular
oxidative status. Cancer Letters 356(1), 58-71. doi:
https://doi.org/10.1016/j.canlet.2013.12.017.
Kleopa, K.A., Abrams, C.K., and Scherer, S.S. (2012). How do mutations in GJB1 cause
X-linked Charcot–Marie–Tooth disease? Brain Research 1487, 198-205. doi:
https://doi.org/10.1016/j.brainres.2012.03.068.
118
Kotini, M., and Mayor, R. (2015). Connexins in migration during development and
cancer. Developmental Biology 401(1), 143-151. doi:
https://doi.org/10.1016/j.ydbio.2014.12.023.
Kudo, T., Kure, S., Ikeda, K., Xia, A.-P., Katori, Y., Suzuki, M., et al. (2003). Transgenic
expression of a dominant-negative connexin26 causes degeneration of the organ
of Corti and non-syndromic deafness. Human Molecular Genetics 12(9), 995-
1004. doi: 10.1093/hmg/ddg116.
Kumai, M., Nishii, K., Nakamura, K., Takeda, N., Suzuki, M., and Shibata, Y. (2000).
Loss of connexin45 causes a cushion defect in early cardiogenesis. Development
127(16), 3501-3512.
Kumar, N.M., and Gilula, N.B. (1996). The Gap Junction Communication Channel. Cell
84(3), 381-388. doi: 10.1016/S0092-8674(00)81282-9.
Kurman, R.J., and Shih Ie, M. (2011). Molecular pathogenesis and extraovarian origin of
epithelial ovarian cancer--shifting the paradigm. Hum Pathol 42, 918-931.
L. Harris, A. (2002). Emerging issues of connexin channels: biophysics fills the gap.
Quarterly Reviews of Biophysics 34(3), 325-472. doi:
10.1017/S0033583501003705.
Laird, Dale W. (2006). Life cycle of connexins in health and disease. Biochemical
Journal 394(3), 527-543. doi: 10.1042/bj20051922.
Laird, D.W., Naus, C.C., and Lampe, P.D. (2017). SnapShot: Connexins and Disease.
Cell 170(6), 1260-1260.e1261. doi: https://doi.org/10.1016/j.cell.2017.08.034.
119
Lampe, P.D., and Lau, A.F. (2000). Regulation of Gap Junctions by Phosphorylation of
Connexins. Archives of Biochemistry and Biophysics 384(2), 205-215. doi:
https://doi.org/10.1006/abbi.2000.2131.
Langlois, S., Cowan, K.N., Shao, Q., Cowan, B.J., and Laird, D.W. (2010). The tumor-
suppressive function of Connexin43 in keratinocytes is mediated in part via
interaction with caveolin-1. Cancer Res 70(10), 4222-4232. doi: 10.1158/0008-
5472.can-09-3281.
Lathia, J.D., Mack, S.C., Mulkearns-Hubert, E.E., Valentim, C.L.L., and Rich, J.N.
(2015). Cancer stem cells in glioblastoma. Genes & Development 29(12), 1203-
1217. doi: 10.1101/gad.261982.115.
Lautermann, J., ten Cate, W.-J.F., Altenhoff, P., Grümmer, R., Traub, O., Frank, H.G., et
al. (1998). Expression of the gap-junction connexins 26 and 30 in the rat cochlea.
Cell and Tissue Research 294(3), 415-420. doi: 10.1007/s004410051192.
Lee, H.J., Lee, I.K., Seul, K.H., and Rhee, S.K. (2002). Growth inhibition by connexin26
expression in cultured rodent tumor cells. Mol Cells 14(1), 136-142.
Lee, S.W., Tomasetto, C., and Sager, R. (1991). Positive selection of candidate tumor-
suppressor genes by subtractive hybridization. Proc Natl Acad Sci U S A 88(7),
2825-2829.
Lim, P.K., Bliss, S.A., Patel, S.A., Taborga, M., Dave, M.A., Gregory, L.A., et al. (2011).
Gap junction-mediated import of microRNA from bone marrow stromal cells can
elicit cell cycle quiescence in breast cancer cells. Cancer Res 71(5), 1550-1560.
doi: 10.1158/0008-5472.can-10-2372.
120
Loewenstein, W.R., and Kanno, Y. (1966). Intercellular Communication and the Control
of Tissue Growth: Lack of Communication between Cancer Cells. Nature 209,
1248. doi: 10.1038/2091248a0.
Luvero, D., Milani, A., and Ledermann, J.A. (2014). Treatment options in recurrent
ovarian cancer: latest evidence and clinical potential. Ther Adv Med Oncol 6, 229-
239.
Majumder, P., Crispino, G., Rodriguez, L., Ciubotaru, C.D., Anselmi, F., Piazza, V., et
al. (2010). ATP-mediated cell–cell signaling in the organ of Corti: the role of
connexin channels. Purinergic Signalling 6(2), 167-187. doi: 10.1007/s11302-
010-9192-9.
Mancuso, M., Pasquali, E., Giardullo, P., Leonardi, S., Tanori, M., Majo, V.D., et al.
(2012). The Radiation Bystander Effect and its Potential Implications for Human
Health. Current Molecular Medicine 12(5), 613-624. doi:
http://dx.doi.org/10.2174/156652412800620011.
Mancuso, M., Pasquali, E., Leonardi, S., Rebessi, S., Tanori, M., Giardullo, P., Borra, F.,
Pazzaglia, S., Naus, C.C., Di Majo, V., and Saran, A. (2011). Role of connexin43
and ATP in long-range bystander radiation damage and oncogenesis in vivo.
Oncogene 30, 4601-4608.
Marusyk, A., Almendro, V., and Polyak, K. (2012). Intra-tumour heterogeneity: a looking
glass for cancer? Nat Rev Cancer 12, 323-334.
Matsumoto, H., Hayashi, S., Hatashita, M., Ohnishi, K., Shioura, H., Ohtsubo, T., et al.
(2001). Induction of Radioresistance by a Nitric Oxide-Mediated Bystander
Effect. Radiation Research 155(3), 387-396.
121
Mccluggage, W.G. (2011). Morphological subtypes of ovarian carcinoma: a review with
emphasis on new developments and pathogenesis. Pathology 43, 420-432.
Mei, L., Chen, J., Zong, L., Zhu, Y., Liang, C., Jones, R.O., et al. (2017). A deafness
mechanism of digenic Cx26 (GJB2) and Cx30 (GJB6) mutations: Reduction of
endocochlear potential by impairment of heterogeneous gap junctional function in
the cochlear lateral wall. Neurobiology of Disease 108, 195-203. doi:
https://doi.org/10.1016/j.nbd.2017.08.002.
Menichella , D.M., Goodenough, D.A., Sirkowski, E., Scherer, S.S., and Paul, D.L.
(2003). Connexins Are Critical for Normal Myelination in the CNS. The Journal
of Neuroscience 23(13), 5963-5973. doi: 10.1523/jneurosci.23-13-05963.2003.
Méry, B., Guy, J.-B., Vallard, A., Espenel, S., Ardail, D., Rodriguez-Lafrasse, C., et al.
(2017). In Vitro Cell Death Determination for Drug Discovery: A Landscape
Review of Real Issues. Journal of cell death 10, 1179670717691251-
1179670717691251. doi: 10.1177/1179670717691251.
Mittal, R., Debs, L.H., Nguyen, D., Patel, A.P., Grati, M.h., Mittal, J., et al. (2017).
Signaling in the Auditory System: Implications in Hair Cell Regeneration and
Hearing Function. Journal of Cellular Physiology 232(10), 2710-2721. doi:
doi:10.1002/jcp.25695.
Miyoshi, K., Shillingford, J.M., Smith, G.H., Grimm, S.L., Wagner, K.-U., Oka, T., et al.
(2001). Signal transducer and activator of transcription (Stat) 5 controls the
proliferation and differentiation of mammary alveolar epithelium. The Journal of
Cell Biology 155(4), 531-542. doi: 10.1083/jcb.200107065.
Mok, S.C., Bell, D.A., Knapp, R.C., Fishbaugh, P.M., Welch, W.R., Muto, M.G.,
122
Berkowitz, R.S., and Tsao, S.W. (1993). Mutation of K-ras protooncogene in
human ovarian epithelial tumors of borderline malignancy. Cancer Res 53, 1489-
1492.
Momiyama, M., Omori, Y., Ishizaki, Y., Nishikawa, Y., Tokairin, T., Ogawa, J., et al.
(2003). Connexin26-mediated gap junctional communication reverses the
malignant phenotype of MCF-7 breast cancer cells. Cancer Sci 94(6), 501-507.
Moorby, C., and Patel, M. (2001). Dual functions for connexins: Cx43 regulates growth
independently of gap junction formation. Exp Cell Res 271(2), 238-248. doi:
10.1006/excr.2001.5357.
Moreno, A.P. (2004). Biophysical properties of homomeric and heteromultimeric
channels formed by cardiac connexins. Cardiovascular Research 62(2), 276-286.
doi: 10.1016/j.cardiores.2004.03.003.
Mulkearns-Hubert, E.E., Torre-Healy, L.A., Silver, D.J., Eurich, J.T., Serbinowski, E.,
Hitomi, M., et al. (2018). Development of a Cx46 targeting strategy for cancer
stem cells. bioRxiv. doi: 10.1101/289397.
Munoz, J.L., Rodriguez-Cruz, V., Greco, S.J., Ramkissoon, S.H., Ligon, K.L., and
Rameshwar, P. (2014). Temozolomide resistance in glioblastoma cells occurs
partly through epidermal growth factor receptor-mediated induction of connexin
43. Cell Death Dis 5, e1145. doi: 10.1038/cddis.2014.111.
Murphy, S.F., Varghese, R.T., Lamouille, S., Guo, S., Pridham, K.J., Kanabur, P., et al.
(2016). Connexin 43 Inhibition Sensitizes Chemoresistant Glioblastoma Cells to
Temozolomide. Cancer Research 76(1), 139.
Nagaraj, A.B., Joseph, P., Kovalenko, O., Singh, S., Armstrong, A., Redline, R., et al.
123
(2015). Critical role of Wnt/β-catenin signaling in driving epithelial ovarian
cancer platinum resistance. Oncotarget 6(27), 23720-23734.
Nahab, F., Wittevrongel, L., Ippolito, D., Toro, C., Grimes, G., Starling, J., Potti, G.,
Haubenberger, D., Bowen, D., Buchwald, P., Dong, C., Kalowitz, D., and Hallett,
M. (2011). An Open-Label, Single-Dose, Crossover Study of the
Pharmacokinetics and Metabolism of Two Oral Formulations of 1-Octanol in
Patients with Essential Tremor. Neurotherapeutics 8, 753-762.
Naus, C.C., and Laird, D.W. (2010). Implications and challenges of connexin
connections to cancer. Nat Rev Cancer 10(6), 435-441. doi: 10.1038/nrc2841.
Naus, C.C.G., Elisevich, K., Zhu, D., Belliveau, D.J., and Del Maestro, R.F. (1992). In
Vivo Growth of C6 Glioma Cells Transfected with Connexin43 cDNA. Cancer
Research 52(15), 4208-4213.
Nelles, E., Bützler, C., Jung, D., Temme, A., Gabriel, H.D., Dahl, U., Traub, O., Stümpel,
F., Jungermann, K., Zielasek, J., Toyka, K.V., Dermietzel, R., and Willecke, K.
(1996). Defective propagation of signals generated by sympathetic nerve
stimulation in the liver of connexin32-deficient mice. Proceedings of the National
Academy of Sciences of the United States of America 93, 9565-9570.
Negrotto, S., Ng, K.P., Jankowska, A.M., Bodo, J., Gopalan, B., Guinta, K., Mulloy, J.C.,
Hsi, E., Maciejewski, J., and Saunthararajah, Y. (2012). CpG methylation patterns
and decitabine treatment response in acute myeloid leukemia cells and normal
hematopoietic precursors. Leukemia 26, 244-254.
Ng, K.P., Ebrahem, Q., Negrotto, S., Mahfouz, R.Z., Link, K.A., Hu, Z., Gu, X., Advani,
A., Kalaycio, M., Sobecks, R., Sekeres, M., Copelan, E., Radivoyevitch, T.,
124
Maciejewski, J., Mulloy, J.C., and Saunthararajah, Y. (2011). p53 Independent
epigenetic-differentiation treatment in xenotransplant models of acute myeloid
leukemia. Leukemia 25, 1739-1750.
Nicotera, P., Leist, M., and Ferrando-May, E. (1998). Intracellular ATP, a switch in the
decision between apoptosis and necrosis. Toxicol Lett 102-103, 139-142.
Obata, K., Morland, S.J., Watson, R.H., Hitchcock, A., Chenevix-Trench, G., Thomas,
E.J., and Campbell, I.G. (1998). Frequent PTEN/MMAC mutations in
endometrioid but not serous or mucinous epithelial ovarian tumors. Cancer Res
58, 2095-2097.
Omori, Y., Zaidan Dagli, M.L., Yamakage, K., and Yamasaki, H. (2001). Involvement of
gap junctions in tumor suppression: analysis of genetically-manipulated mice.
Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis
477(1), 191-196. doi: https://doi.org/10.1016/S0027-5107(01)00120-8.
Oyamada, M., Oyamada, Y., and Takamatsu, T. (2005). Regulation of connexin
expression. Biochimica et Biophysica Acta (BBA) - Biomembranes 1719(1), 6-23.
doi: https://doi.org/10.1016/j.bbamem.2005.11.002.
Patel, J.S., Hu, M., Sinha, G., Walker, N.D., Sherman, L.S., Gallagher, A., et al. (2016).
Non-coding RNA as mediators in microenvironment–breast cancer cell
communication. Cancer Letters 380(1), 289-295. doi:
https://doi.org/10.1016/j.canlet.2015.11.016.
Paul, D.L., Ebihara, L., Takemoto, L.J., Swenson, K.I., and Goodenough, D.A. (1991).
Connexin46, a novel lens gap junction protein, induces voltage-gated currents in
125
nonjunctional plasma membrane of Xenopus oocytes. The Journal of Cell Biology
115(4), 1077-1089. doi: 10.1083/jcb.115.4.1077.
Plante, I., Stewart, M.K.G., Barr, K., Allan, A.L., and Laird, D.W. (2010). Cx43
suppresses mammary tumor metastasis to the lung in a Cx43 mutant mouse model
of human disease. Oncogene 30, 1681. doi: 10.1038/onc.2010.551
Plotkin, L.I., Manolagas, S.C., and Bellido, T. (2002). Transduction of Cell Survival
Signals by Connexin-43 Hemichannels. Journal of Biological Chemistry 277(10),
8648-8657. doi: 10.1074/jbc.M108625200.
Plum, A., Hallas, G., Magin, T., Dombrowski, F., Hagendorff, A., Schumacher, B., et al.
(2000). Unique and shared functions of different connexins in mice. Current
Biology 10(18), 1083-1091. doi: 10.1016/S0960-9822(00)00690-4.
Pollmann, M.-A., Shao, Q., Laird, D.W., and Sandig, M. (2005). Connexin 43 mediated
gap junctional communication enhances breast tumor cell diapedesis in culture.
Breast Cancer Research 7(4), R522-R534. doi: 10.1186/bcr1042.
Prat, J. (2012). Ovarian carcinomas: five distinct diseases with different origins, genetic
alterations, and clinicopathological features. Virchows Arch 460, 237-249.
Prise, K.M., and O'Sullivan, J.M. (2009). Radiation-induced bystander signalling in
cancer therapy. Nature reviews. Cancer 9(5), 351-360. doi: 10.1038/nrc2603.
Qiu, X., Cheng, J.C., Klausen, C., Chang, H.M., Fan, Q., and Leung, P.C. (2016). EGF-
Induced Connexin43 Negatively Regulates Cell Proliferation in Human Ovarian
Cancer. J Cell Physiol 231, 111-119.
126
Quist, A.P., Rhee, S.K., Lin, H., and Lal, R. (2000). Physiological Role of Gap-
Junctional Hemichannels: Extracellular Calcium-Dependent Isosmotic Volume
Regulation. The Journal of Cell Biology 148(5), 1063-1074.
Ratajczak, J., Wysoczynski, M., Hayek, F., Janowska-Wieczorek, A., and Ratajczak,
M.Z. (2006). Membrane-derived microvesicles: important and underappreciated
mediators of cell-to-cell communication. Leukemia 20, 1487. doi:
10.1038/sj.leu.2404296.
Reid, B.M., Permuth, J.B., and Sellers, T.A. (2017). Epidemiology of ovarian cancer: a
review. Cancer Biol Med 14, 9-32.
Riss TL, Moravec RA, Niles AL, et al. Cell Viability Assays. 2013 May 1 [Updated 2016
Jul 1]. In: Sittampalam GS, Coussens NP, Brimacombe K, et al., editors. Assay
Guidance Manual [Internet]. Bethesda (MD): Eli Lilly & Company and the
National Center for Advancing Translational Sciences; 2004-.Available from:
https://www.ncbi.nlm.nih.gov/books/NBK144065/
Robertson, J.D. (1953). Ultrastructure of Two Invertebrate Synapses. Proceedings of the
Society for Experimental Biology and Medicine 82(2), 219-223. doi:
10.3181/00379727-82-20071.
Salomon, D., Masgrau, E., Vischer, S., Ullrich, S., Dupont, E., Sappino, P., et al. (1994).
Topography of Mammalian Connexins in Human Skin. Journal of Investigative
Dermatology 103(2), 240-247. doi: https://doi.org/10.1111/1523-
1747.ep12393218.
127
Sato, H., Hagiwara, H., Ohde, Y., Senba, H., Virgona, N., and Yano, T. (2007).
Regulation of renal cell carcinoma cell proliferation, invasion and metastasis by
connexin 32 gene. J Membr Biol 216(1), 17-21. doi: 10.1007/s00232-007-9020-5.
Seidman, J.D., and Khedmati, F. (2008). Exploring the histogenesis of ovarian mucinous
and transitional cell (Brenner) neoplasms and their relationship with Walthard cell
nests: a study of 120 tumors. Arch Pathol Lab Med 132, 1753-1760.
Shareef, M.M., Cui, N., Burikhanov, R., Gupta, S., Satishkumar, S., Shajahan, S., et al.
(2007). Role of Tumor Necrosis Factor-α and TRAIL in High-Dose Radiation–
Induced Bystander Signaling in Lung Adenocarcinoma. Cancer Research 67(24),
11811-11820. doi: 10.1158/0008-5472.can-07-0722.
Shen, Y., Li, Y., Ma, X., Wan, Q., Jiang, Z., Liu, Y., et al. (2018). Connexin 43
SUMOylation improves gap junction functions between liver cancer stem cells
and enhances their sensitivity to HSVtk/GCV. Int J Oncol 52(3), 872-880. doi:
10.3892/ijo.2018.4263.
Shih Ie, M., and Kurman, R.J. (2004). Ovarian tumorigenesis: a proposed model based on
morphological and molecular genetic analysis. Am J Pathol 164, 1511-1518.
Shin, J.H., Park, S.J., Jo, Y.K., Kim, E.S., Kang, H., Park, J.H., Lee, E.H., and Cho, D.H.
(2012). Suppression of autophagy exacerbates Mefloquine-mediated cell death.
Neurosci Lett 515, 162-167.
Singer, G., Oldt, R., 3rd, Cohen, Y., Wang, B.G., Sidransky, D., Kurman, R.J., and Shih
Ie, M. (2003). Mutations in BRAF and KRAS characterize the development of
low-grade ovarian serous carcinoma. J Natl Cancer Inst 95, 484-486.
Sinyuk, M., Alvarado, A.G., Nesmiyanov, P., Shaw, J., Mulkearns-Hubert, E.E., Eurich,
128
J.T., Hale, J.S., Bogdanova, A., Hitomi, M., Maciejewski, J., Huang, A.Y.,
Saunthararajah, Y., and Lathia, J.D. (2015). Cx25 contributes to leukemia cell
communication and chemosensitivity. Oncotarget 6, 31508-31521.
Sokolov, M.V., Smilenov, L.B., Hall, E.J., Panyutin, I.G., Bonner, W.M., and
Sedelnikova, O.A. (2005). Ionizing radiation induces DNA double-strand breaks
in bystander primary human fibroblasts. Oncogene 24(49), 7257-7265. doi:
10.1038/sj.onc.1208886.
Söhl, G., and Willecke, K. (2004). Gap junctions and the connexin protein family.
Cardiovascular Research 62, 228-232.
Song, D., Liu, X., Liu, R., Yang, L., Zuo, J., and Liu, W. (2010). Connexin 43
hemichannel regulates H9c2 cell proliferation by modulating intracellular ATP
and [Ca2+]. Acta Biochim Biophys Sin (Shanghai) 42(7), 472-482. doi:
10.1093/abbs/gmq047.
Spagnol, G., Trease, A.J., Zheng, L., Gutierrez, M., Basu, I., Sarmiento, C., et al. (2018).
Connexin43 Carboxyl-Terminal Domain Directly Interacts with beta-Catenin. Int
J Mol Sci 19(6). doi: 10.3390/ijms19061562.
Srinivas, M., Calderon, D.P., Kronengold, J., and Verselis, V.K. (2006). Regulation of
Connexin Hemichannels by Monovalent Cations. The Journal of General
Physiology 127(1), 67-75. doi: 10.1085/jgp.200509397.
Stains, J.P., and Civitelli, R. (2005). Gap junctions in skeletal development and function.
Biochimica et Biophysica Acta (BBA) - Biomembranes 1719, 69-81.
Stains, J.P., Lecanda, F., Screen, J., Towler, D.A., and Civitelli, R. (2003). Gap
Junctional Communication Modulates Gene Transcription by Altering the
129
Recruitment of Sp1 and Sp3 to Connexin-response Elements in Osteoblast
Promoters. Journal of Biological Chemistry 278(27), 24377-24387. doi:
10.1074/jbc.M212554200.
Stoletov, K., Strnadel, J., Zardouzian, E., Momiyama, M., Park, F.D., Kelber, J.A., et al.
(2013). Role of connexins in metastatic breast cancer and melanoma brain
colonization. Journal of Cell Science 126(4), 904-913. doi: 10.1242/jcs.112748.
Subak-Sharpe, H., Burk, R.R., and Pitts, J.D. (1969). Metabolic Co-Operation Between
Biochemically Marked Mammalian Cells in Tissue Culture. Journal of Cell
Science 4(2), 353-367.
Taniguchi Ishikawa, E., Gonzalez-Nieto, D., Ghiaur, G., Dunn, S.K., Ficker, A.M.,
Murali, B., Madhu, M., Gutstein, D.E., Fishman, G.I., Barrio, L.C., and Cancelas,
J.A. (2012). Connexin-43 prevents hematopoietic stem cell senescence through
transfer of reactive oxygen species to bone marrow stromal cells. Proceedings of
the National Academy of Sciences 109, 9071-9076.
Temme, A., Buchmann, A., Gabriel, H.-D., Nelles, E., Schwarz, M., and Willecke, K.
(1997). High incidence of spontaneous and chemically induced liver tumors in
mice deficient for connexin32. Current Biology 7(9), 713-716. doi:
https://doi.org/10.1016/S0960-9822(06)00302-2.
Thiagarajan, P.S., Sinyuk, M., Turaga, S.M., Mulkearns-Hubert, E.E., Hale, J.S., Rao, V.,
et al. (2018). Cx26 drives self-renewal in triple-negative breast cancer via
interaction with NANOG and focal adhesion kinase. Nature Communications
9(1), 578. doi: 10.1038/s41467-018-02938-1.
Tomao, F., Marchetti, C., Romito, A., Di Pinto, A., Di Donato, V., Capri, O., Palaia, I.,
130
Monti, M., Muzii, L., and Benedetti Panici, P. (2017). Overcoming platinum
resistance in ovarian cancer treatment: from clinical practice to emerging
chemical therapies. Expert Opin Pharmacother 18, 1443-1455.
Torre, L.A., Trabert, B., and Desantis, C.E. (2018). Ovarian cancer statistics, 2018. 68,
284-296.
Trosko, J.E., and Ruch, R.J. (2002). Gap junctions as targets for cancer chemoprevention
and chemotherapy. Curr Drug Targets 3, 465-482.
Uhlén, M., Fagerberg, L., Hallström, B.M., Lindskog, C., Oksvold, P., Mardinoglu, A., et
al. (2015). Tissue-based map of the human proteome. Science 347(6220). doi:
10.1126/science.1260419.
Umezu, T., Ohyashiki, K., Kuroda, M., and Ohyashiki, J.H. (2013). Leukemia cell to
endothelial cell communication via exosomal miRNAs. Oncogene 32, 2747-2755.
Upham, B.L., and Trosko, J.E. (2009). Oxidative-Dependent Integration of Signal
Transduction with Intercellular Gap Junctional Communication in the Control of
Gene Expression. Antioxidants & Redox Signaling 11, 297-307.
Uzu, M., Sato, H., Yamada, R., Kashiba, T., Shibata, Y., Yamaura, K., et al. (2015).
Effect of enhanced expression of connexin 43 on sunitinib-induced cytotoxicity in
mesothelioma cells. J Pharmacol Sci 128(1), 17-26. doi:
10.1016/j.jphs.2015.04.002.
Van Zyl, B., Tang, D., and Bowden, N.A. (2018). Biomarkers of platinum resistance in
ovarian cancer: what can we use to improve treatment. Endocr Relat Cancer 25,
R303-r318.
Vang, R., Shih Ie, M., and Kurman, R.J. (2013). Fallopian tube precursors of ovarian
131
low- and high-grade serous neoplasms. Histopathology 62, 44-58.
Veeraraghavan, R., Hoeker, G.S., Alvarez-Laviada, A., Hoagland, D., Wan, X., King,
D.R., et al. (2018). The adhesion function of the sodium channel beta subunit (β1)
contributes to cardiac action potential propagation. eLife 7, e37610. doi:
10.7554/eLife.37610.
Véliz, L.P., González, F.G., Duling, B.R., Sáez, J.C., and Boric, M.P. (2008). Functional
role of gap junctions in cytokine-induced leukocyte adhesion to endothelium in
vivo. American Journal of Physiology - Heart and Circulatory Physiology 295,
H1056-H1066.
Veras, E., Mao, T.L., Ayhan, A., Ueda, S., Lai, H., Hayran, M., Shih Ie, M., and Kurman,
R.J. (2009). Cystic and adenofibromatous clear cell carcinomas of the ovary:
distinctive tumors that differ in their pathogenesis and behavior: a
clinicopathologic analysis of 122 cases. Am J Surg Pathol 33, 844-853.
Vergara, L., Bao, X., Cooper, M., Bello-Reuss, E., and Reuss, L. (2003). Gap-junctional
hemichannels are activated by ATP depletion in human renal proximal tubule
cells. J Membr Biol 196(3), 173-184. doi: 10.1007/s00232-003-0636-9.
Verma, N., and Tiku, A.B. (2017). Significance and nature of bystander responses
induced by various agents. Mutation Research/Reviews in Mutation Research
773, 104-121. doi: https://doi.org/10.1016/j.mrrev.2017.05.003.
Vinken, M., Vanhaecke, T., Papeleu, P., Snykers, S., Henkens, T., and Rogiers, V.
(2006). Connexins and their channels in cell growth and cell death. Cellular
Signalling 18(5), 592-600. doi: https://doi.org/10.1016/j.cellsig.2005.08.012.
132
W., W.T., R., D.M., John, O.B., R., A.-U.M., A., G.D., Harris, R., et al. (1999).
Functional characteristics of skate connexin35, a member of the γ subfamily of
connexins expressed in the vertebrate retina. European Journal of Neuroscience
11(6), 1883-1890. doi: doi:10.1046/j.1460-9568.1999.00607.x.
Wang, N., De Bock, M., Decrock, E., Bol, M., Gadicherla, A., Bultynck, G., et al. (2013).
Connexin targeting peptides as inhibitors of voltage- and intracellular Ca2+-
triggered Cx43 hemichannel opening. Neuropharmacology 75, 506-516. doi:
10.1016/j.neuropharm.2013.08.021.
Ward, J.F. (1988). "DNA Damage Produced by Ionizing Radiation in Mammalian Cells:
Identities, Mechanisms of Formation, and Reparability," in Progress in Nucleic
Acid Research and Molecular Biology, eds. W.E. Cohn & K. Moldave. Academic
Press), 95-125.
Warner, A.E., Guthrie, S.C., and Gilula, N.B. (1984). Antibodies to gap-junctional
protein selectively disrupt junctional communication in the early amphibian
embryo. Nature 311, 127-131.
Wei, C.-J., Xu, X., and Lo, C.W. (2004). Connexins and Cell signaling in Development
and Disease. Annual Review of Cell and Developmental Biology 20(1), 811-838.
doi: 10.1146/annurev.cellbio.19.111301.144309.
Weidmann, S. (1952). The electrical constants of Purkinje fibres. The Journal of
Physiology 118(3), 348-360.
Wigglesworth, K., Lee, K.B., O'brien, M.J., Peng, J., Matzuk, M.M., and Eppig, J.J.
(2013). Bidirectional communication between oocytes and ovarian follicular
somatic cells is required for meiotic arrest of mammalian oocytes. Proc Natl Acad
133
Sci U S A 110, E3723-3729.
Wingard, J.C., and Zhao, H.-B. (2015). Cellular and Deafness Mechanisms Underlying
Connexin Mutation-Induced Hearing Loss – A Common Hereditary Deafness.
Frontiers in Cellular Neuroscience 9(202). doi: 10.3389/fncel.2015.00202.
Winterhager, E., and Kidder, G.M. (2015). Gap junction connexins in female
reproductive organs: implications for women's reproductive health. Human
Reproduction Update 21, 340-352.
Wiszniewski, L., Limat, A., Saurat, J.-H., Meda, P., and Salomon, D. (2000). Differential
Expression of Connexins During Stratification of Human Keratinocytes. Journal
of Investigative Dermatology 115(2), 278-285. doi:
https://doi.org/10.1046/j.1523-1747.2000.00043.x.
Wu, R., Zhai, Y., Fearon, E.R., and Cho, K.R. (2001). Diverse mechanisms of beta-
catenin deregulation in ovarian endometrioid adenocarcinomas. Cancer Res 61,
8247-8255.
Wu, W., Fan, L., Bao, Z., Zhang, Y., Peng, Y., Shao, M., Xiang, Y., Zhang, X., Wang,
Q., and Tao, L. (2017). The cytoplasmic translocation of Cx32 mediates cisplatin
resistance in ovarian cancer cells. Biochem Biophys Res Commun 487, 292-299.
Xie, Y., Davies, S.M., Xiang, Y., Robison, L.L., and Ross, J.A. (2003). Trends in
leukemia incidence and survival in the United States (1973–1998). Cancer 97,
2229-2235.
Xu, J., and Nicholson, B.J. (2013). The role of connexins in ear and skin physiology —
Functional insights from disease-associated mutations. Biochimica et Biophysica
134
Acta (BBA) - Biomembranes 1828(1), 167-178. doi:
https://doi.org/10.1016/j.bbamem.2012.06.024.
Xu, X., Francis, R., Wei, C.J., Linask, K.L., and Lo, C.W. (2006). Connexin 43-mediated
modulation of polarized cell movement and the directional migration of cardiac
neural crest cells. Development 133(18), 3629-3639. doi: 10.1242/dev.02543.
Yang-Hartwich, Y., Gurrea-Soteras, M., Sumi, N., Joo, W.D., Holmberg, J.C., Craveiro,
V., Alvero, A.B., and Mor, G. (2014). Ovulation and extra-ovarian origin of
ovarian cancer. Sci Rep 4, 6116.
Yu, M., Zhang, C., Li, L., Dong, S., Zhang, N., and Tong, X. (2014). Cx43 reverses the
resistance of A549 lung adenocarcinoma cells to cisplatin by inhibiting EMT.
Oncol Rep 31(6), 2751-2758. doi: 10.3892/or.2014.3163.
Zhang, A., Hitomi, M., Bar-Shain, N., Dalimov, Z., Ellis, L., Velpula, K.K., et al.
(2015a). Connexin 43 expression is associated with increased malignancy in
prostate cancer cell lines and functions to promote migration. Oncotarget 6(13),
11640-11651.
Zhang, A., Hitomi, M., Bar-Shain, N., Dalimov, Z., Ellis, L., Velpula, K.K., et al.
(2015b). Connexin 43 expression is associated with increased malignancy in
prostate cancer cell lines and functions to promote migration.
Zhang, J.-T., and Nicholson, B.J. (1994). The topological structure of connexin 26 and its
distribution compared to connexin 32 in hepatic gap junctions. The Journal of
Membrane Biology 139(1), 15-29. doi: 10.1007/bf00232671.
Zhang, A., Hitomi, M., Bar-Shain, N., Dalimov, Z., Ellis, L., Velpula, K.K., Fraizer,
G.C., Gourdie, R.G., and Lathia, J.D. (2015). Connexin 43 expression is
135
associated with increased malignancy in prostate cancer cell lines and functions
to promote migration.
Zhang, Y.-W., Kaneda, M., and Morita, I. (2003). The Gap Junction-independent Tumor-
suppressing Effect of Connexin 43. Journal of Biological Chemistry 278(45),
44852-44856. doi: 10.1074/jbc.M305072200.
Zhang, X., Liu, Y., Si, Y.-J., Chen, X.-H., Li, Z.-J., Gao, L., Gao, L., and Zhang, C.
(2012). Effect of Cx43 gene-modified leukemic bone marrow stromal cells on the
regulation of Jurkat cell line in vitro. Leukemia Research 36, 198-204.
Zhou, J.Z., and Jiang, J.X. (2014). Gap junction and hemichannel-independent actions of
connexins on cell and tissue functions – An update. FEBS letters 588(8), 1186-
1192. doi: 10.1016/j.febslet.2014.01.001.
Zhu, D., Caveney, S., Kidder, G.M., and Naus, C.C. (1991). Transfection of C6 glioma
cells with connexin 43 cDNA: analysis of expression, intercellular coupling, and
cell proliferation. Proceedings of the National Academy of Sciences 88(5), 1883-
1887. doi: 10.1073/pnas.88.5.1883.
Zoidl, G., Bruzzone, R., Weickert, S., Kremer, M., Zoidl, C., Mitropoulou, G., et al.
(2004). Molecular Cloning and Functional Expression of Cx52.6: A Novel
Connexin with Hemichannel-forming Properties Expressed in Horizontal Cells of
the Zebrafish Retina. Journal of Biological Chemistry 279(4), 2913-2921. doi:
10.1074/jbc.M304850200.
136