|Home | About | Journals | Submit | Contact Us | Français|
All cells possess long-term, steady-state voltage gradients across the plasma membrane. These transmembrane potentials arise from the combined activity of numerous ion channels, pumps and gap junction complexes. Increasing data from molecular physiology now reveal that the role of changes in membrane voltage controls, and is in turn controlled by, progression through the cell cycle. We review recent functional data on the regulation of mitosis by bioelectric signals, and the function of membrane voltage and specific potassium, sodium and chloride ion channels in the proliferation of embryonic, somatic and neoplastic cells. Its unique properties place this powerful, well-conserved, but still poorly-understood signaling system at the center of the coordinated cellular interactions required for complex pattern formation. Moreover, disregulation of ion channel expression and function is increasingly observed to be not only a useful marker but likely a functional element in oncogenesis. New advances in genomics and the development of in vivo biophysical techniques suggest exciting opportunities for molecular medicine, bioengineering and regenerative approaches to human health.
Regulation of the cell cycle is of significant importance to many areas of biology. During development stem cells must maintain their proliferative potential, while some terminally differentiated cells such as neurons no longer divide once specified. In wound healing and regeneration, cells must initially proliferate in order to fill damaged areas or replace lost structures, and then downregulate growth once the proper pattern has been restored. Disruption of cell cycle checkpoints can lead to uncontrolled division of cells and is highly relevant to cancer biology and human health. While tremendous progress has been made on the molecular details of biochemical checkpoint machinery (e.g., kinase cascades),1 an important area of mitotic regulation still offers much opportunity for new discoveries: bioelectrical events controlling transmembrane voltage potential.2
The first correlations between membrane potential (the voltage gradient across the plasma membrane, Vmem) and proliferative ability came from observations that cell types with a very high resting potential such as muscle cells and neurons show little if any mitotic activity (reviewed in ref. 3). Though the early 1950’s it remained unclear whether there was a causal relationship between Vmem and proliferation, or if both were simply characteristics related to the specialized function of these cells. This question invoked more intense research in the late 1950’s and early 1960’s, following a number of studies in which multiple groups reported a decrease in the membrane potential of cells following malignant transformation.4–6 These results, in addition to observations that cultured cells under high growth conditions show a decrease in Vmem, were among the first to suggest a causal relationship between ion flow and the cell cycle.
These ideas were formally tested in a series of groundbreaking experiments by Clarence D. Cone, Jr. throughout the late 1960’s. He first observed that Vmem varied through the cell cycle and postulated that the variations were directly related to progression through G1/S and G2/M transitions in proliferating cells.7 In a follow up study to explicitly test causation, he altered the intracellular ionic concentration of cells and was able to induce a reversible mitotic block by mimicking Vmem to levels observed in neurons.8 Even more impressively, it was shown that sustained depolarization was able to induce DNA synthesis and mitosis in mature neurons.9,10
Cone would later synthesize these studies, as well as the pioneering prior work, into a theory on the basic mechanism of mitotic control and oncogenesis.11 This paper eloquently argued for a direct relationship between the cycle and electrical transmembrane potential and discussed possible mechanisms, both molecular and physiological. His commentary was notable given the limited information and methods available at the time, and Cone’s innovative work and ideas provided the foundation for formal inquiry into the relationship between membrane potential and the cell cycle.
Since then, molecular, physiological and pharmacological tools have become much more sophisticated, clarifying the molecular details of biochemical pathways involving cyclins, c-myc, c-fos, numerous tumor suppressors and oncogenes. While considerable modern work underscores the link between membrane potential and the cell cycle, this fascinating bioelectric control mechanism is still not well known in the field.
Here, we briefly summarize some of the recent studies examining the relationship between Vmem and cell proliferation in differentiated cells, and the modulation of expression and regulation of ion channels throughout development as cells progress from a stem cell state to terminal differentiation (Table 1). We also discuss data on the role of membrane potential in neoplastic cells, and suggest that this mechanism is an attractive target for modulation in regenerative medicine and developmental biology contexts.
While rapid changes in membrane potential (Vmem) are best known in neurons and muscle, steady-state Vmem levels are associated with all cells, and exhibit cyclic fluctuations on much longer timescales than the familiar millisecond action potentials.12,13 Transmembrane potential arises from the combined actions of numerous channels and pumps, which segregate ions across the cell surface under constraints of concentration gradient and charge. Thus, modulation of specific ion channels and transporters is both an endogenous method for controlling cellular Vmem and a tractable technique for functional experiments. It should be noted that a number of channels and pumps are now known to have additional regulatory roles independent of their current-passing functions.14–17 In this review, we focus on functions specifically associated with charge transfer and voltage control.
Pharmacological blockade of ion channels has been a popular method of disrupting membrane potential; while not as specific as molecular approaches (knockout, RNAi or morpholinos), it has two advantages. First, it allows a more precise control of the timing of current inhibition. Second, it often produces more informative results than gene-specific loss of function because of the extensive compensation among multiple channel types: whereas phenotypes may be masked by redundancy in single-gene knockdowns, a pharmacological approach can reveal the function of membrane potential per se, which is not necessarily dependent on any one particular gene product.
Membrane potential has been examined as a key regulator of proliferation in a number of cell types, suggesting that modulation of Vmem is required for both G1/S phase and G2/M phase transitions. Depolarization of membrane through changes in extracellular ion concentration inhibits G1/S progression of lymphocytes, astrocytes, fibroblasts and Schwann cells suggesting hyper-polarization is a required step for S phase initation.18–21 In B cell lymphocytes, inhibition of channels induces a reversible cell cycle arrest;22 similar results have been noted in other cell types.22–27 Analysis of downstream targets has revealed that inhibition of potassium channels in these cells resulted in expression changes of a number of proteins, including IL-1/2 and transferrin, both of which are implicated in cell cycle control.18,22,23,26,28,29
Conversely, depolarization of the plasma membrane seems to be essential for the G2/M transition. Current models outline a rhythmic oscillation of membrane potential throughout the cell cycle, with a spike in hyperpolaration occurring before DNA synthesis followed by a prolonged period of depolarization necessary for mitosis, and appears to be conserved mechanism from early cell divisions in embryos through the normal division of differentiated tissues.30 The exact threshold of Vmem necessary to drive cells though proliferative stages is not known, and is likely to vary between cell type, and in development, the stage of differentiation.
In human endothelial cells, modulation of Vmem through applied electric fields revealed that hyper-polarizing currents arrest cell division.31 This arrest is characterized by downregulation of cyclin E and concurrent upregulation of the cyclin inhibitor p27, forging a direct link between transmembrane potential and known regulators of cell cycle. Conversely, depolarization of PUS-1.8 mouse macrophages results in DNA synthesis and progression through the cell cycle, and is marked by a subsequent upregulation of the transcription factors c-myc and c-fos.32 Both of these results fit the current Vmem oscillatory model of cell cycle, and suggest an intimate relationship between membrane potential and the well know cyclin-dependent pathways.
While a detailed consideration of the roles of bioelectric signals with differentiation is beyond the scope of this review, differentiation is often closely associated with changes in proliferative capacity. A number of cell types have shown a strong correlation between membrane potential and differentiation, with Vmem becoming more hyperpolarized as cells are specified. For example, the neural crest cells of quail embryos exhibit a −35 mV resting potential early in development, but as development progresses the resting potential shifts to −55 mV.33 These changes have been directly correlated to a turnover in potassium channels, with the early membrane voltage attributed to the expression of a K+ channel, ERG, while at later stages ERG expression is lost and replaced with inward-rectifying K+ channels associated with many differentiated tissues. Similarly, the normal course of human mesenchymal stem cell differentiation is accompanied by a progressive hyperpolarization; it has recently been shown that this is an instructive parameter, as artificial depolarization keeps the MSCs in the stem state, while hyperpolarization accelerates their differentiation.34
Cell cycle is a key parameter of cellular behavior that must be tightly regulated and coordinated during morphogenesis associated with development and regeneration. For example, in addition to normal turnover, changes in membrane potential have been linked to the proliferation of cells during wound healing. In cultured cells, modulation of Vmem using K+ channel-inhibiting drugs increases wound healing in cell monolayers;35 this is a related but distinct mechanism to the modulation of the voltage gradient across an epithelium,36 which is also a crucial component of wound healing and normal development.37 Regulation of gap junctional communication is known to control the proliferation of adult stem cells during regeneration in planaria,38 although the specific ions involved have not been identified. Whereas activation of currents is necessary for proliferation, inhibition of specific currents has been shown to disrupt normal development via changes of proliferative capacity. Knockdown or inhibition of two K+ channels (KV 1.3 and KV 1.5) resulted in cell cycle arrest at G1 in rat oligodendrocyte precursors;39,40 this effect was characterized by accumulation of p27 and p21. Blockade of other K+ channels results in similar signaling cascades, suggesting convergent mechanisms downstream of the activity of many diverse channels.40
Alongside the functional control of proliferation by ion channel function, observations from a wide array of cells have also shown the converse—control of physiology as a function of the cell cycle, suggesting bi-directional regulatory circuits. An increase in the expression or activity of potassium channels was observed after exposure to mitogens,19,41–48 demonstrating channel expression downstream of proliferation checkpoints. It remains unclear which mitogen activated pathways are responsible for the regulation of potassium channels, although the involvement of p21 and its downstream targets, the protein kinase Raf and GTPase Ras, has been shown.49 It is probable the multiple downstream signaling cascades exist, as the regulation of ion channels by mitogens is likely to be dependent on both cell type and the specific mitogenic signals used.
A number of ion channels show variation of expression or activity across stages of the cell cycle. During the G1/S transition, multiple families of K+ currents become active, including ATP-sensitive K+ channels, outward rectifying currents (KIR), and Ca2+ activated K+ channels.50–53 During the G2/M transition, increases in potassium channel EAG currents have been noted.54 In addition, progression to M phase is also characterized by an increase in chloride flux. Voltage activated chloride currents show a strong increase during G2; the NCC27 channel is activated, and ClC-2 shows M phase-specific expression and phosporylation.55–57
Alongside characterization of individual cells in vitro, ion channel and gap junction function changes across the cell cycle have also been observed in embryonic development.58–60 The complex bidirectional relationship between ion transporter function and cell cycle suggests this set of mechanisms as a powerful and versatile physiological network which can be used during pattern formation in a flexible and highly dynamic control mechanism to synchronize cell division.
Neoplasia has long been associated with aberrant changes in cell cycle.61,62 Alterations in membrane potential and ion channel expression/function have been observed in a wide array of cancers.63 Likewise, alongside ion-independent roles of GJs in neoplasm,64 it is clear that gap junctions are an important component of Vmem, regulating its fluctuations,65 establishing iso-potential cell fields, and modulating cellular responses to external electric fields.66 For brevity, the well-known role of gap junctions in cancer is not discussed here (reviewed in refs. 67 and 68).
The proliferation of some tumor cells is dependent on voltage-gated potassium channels.69,70 hERG channels are particularly implicated,71–76 as are 2-pore channels such as KCNK9.77 In the case of KCNK9, it is known that its oncogenic potential depends on K+ transport function, not some other role of the protein,78 and in human colorectal tissue KCNK9 K+ channel expression was shown to be significantly elevated.79 A screen of several cervical cancers found the K+ channel EAG expressed in 100% of the biopsies analyzed, and overexpression of EAG in human cells resulted in more quickly dividing progeny in culture.80,81 This result was replicated in vivo using mice implanted with human EAG (hEAG) expressing CHO cells. All of the mice receiving CHO hEAG injections formed tumors while none of CHO controls formed a growth greater than 1 mm in size.81 hEAG-1 is a true oncogene since its overexpression drives mammalian cells into uncontrolled proliferation and favors tumor progression in cells injected into immune-suppressed mice.81 Likewise the an EAG relative, hERG, is not normally present in most differentiated cells besides the heart but has been observed in a number of human cancers and neoplastic transformation in prostate epithelium.73,82 In these cells, hERG appears to recruit tumor necrosis factor receptor (TNFR) to the plasma membrane and cause a subsequent increase in NFκB, a known proliferation control gene. In addition to modulations of single channels, some cancers are characterized by the activation of multiple potassium currents, such as in the case human melanoma lines which express of both hEAG1 and Ca2+-activated K+ channels.83 Indeed, complex interactions by multiple channels likely exist, and currents driven by diverse families of potassium channels (including calcium activated, inward rectifying Kir, EAG and ERG) have all been correlated with cancerous tissue.
Manipulation of membrane H+ flux can confer a neoplastic phenotype upon cells,84 and voltage-gated sodium channels potentiate breast cancer metastasis.85 Studies in glioma lines have revealed a role of the ClC-3 chloride efflux channel in cellular division.86 Following a chloride buildup, Cl− efflux is a known driving force in cytoplasmic condensation and is required for mitosis to progress. Expression of ClC-3 is localized to both the plasma membrane and mitotic spindle of D54-MG and U251-MG glioma cells, and inhibition of the channel though hairpin RNA constructs resulted in the loss of premitotic condensation and arrest of the cell cyle. These results have been supported in studies of human prostate cancer lines, and support the role of chloride channels as key regulators of proliferation through cell size regulation.87,88
Sodium channels have been implicated in mouse cancers in vivo. The knockout APCmin/+ line share a mutation found in many human colorectal cancers, and subsequently develop multiple intestinal neoplasias.89 In vivo transepithelial voltage recordings in this line revealed an increase in Na+ compared to wild type mice that was the result of an increase in expression of the ENaC Na+ channel. The downstream targets of Na+ signaling are not well known, but it was noted that neither Cl− nor Ca2+ absorption were not altered in the APCmin/+ line.
Unique bioelectrical properties of tumor tissue have been recognized for a long time.94–109 How relevant the changes of ion flux are to neoplasm in general will require much future work. The majority of studies has examined the presence of ion channels in cancerous tissues but have not explicitly examined the physiological role of Vmem changes in the cell. Nor do we know in many cases at which stage of neoplasm development the bioelectric signals are relevant. Cancer cells have been observed to be depolarized with respect to normal healthy epithelial tissue,110–113 and hyperpolarization therapy (molecular-genetic or pharmacological) remains to be tested as a therapeutic approach.
Characterization of ion channels involved in cancer will undoubtedly be of high interest in human health. First, expression of ion channels may prove to be highly relevant markers when screening tissue if unique among the transformed tissue, or if expression levels are significantly elevated.80,114–116 Voltage sensitive dyes have been used successfully in vivo to image the action potentials of the visual cortex, and modification of these techniques has been suggested for cancer screens. Double dye systems have more recently been developed to examine the membrane potential of non-nervous tissue in vivo.117 Application and analysis of these dyes could potentially reduce cost, invasiveness and time involved with diagnosis compared to more traditional biopsies, and in the case of epidermal malignancies such as melanoma may involve no invasiveness whatsoever if they could be applied and examined in intact skin.
Second, inhibition of ion channels through pharmacological treatment has been proposed as a potential cancer treatment.118 Drugs that target membrane voltage-generating transporters have shown clinical promise in cancer.119 Indeed, evidence in cancer lines supports this theory. In human prostate cells inhibition of Ca2+ dependent K+ channels lead to a decrease in cell proliferation,120 and knockdown of EAG with antisense oligonucloetides reduced division rates in somatic cancer lines.81 Growth suppression of pancreatic tumor cells occurs after selective blockade of IK-type channels.121 Control of tumor growth though pharmacology is especially exciting in cancers which display ion channels that are non-existent throughout the rest of the body as there would be little chance of the drug interacting with healthy tissue in the donor. For ion channels which are present in both tumors and surrounding tissues, targeted delivery remains an active area for investigation.
Modulation of the membrane voltage as a novel parameter controlling cell proliferation offers unique opportunities for guiding morphogenesis in vitro (tissue engineering) and in vivo (regenerative medicine).2,122,123 For example, V-ATPase proton pump activity in the zebrafish eye is needed for retinoblast proliferation and survival,124 while induction of proton efflux from wound tissue has been shown to induce complete regeneration of the tail in Xenopus tadpoles.125 The molecular details of epigenetic bioelectrical pathways must be considered in developing strategies for rational modulation of cell behavior based on bioelectric controls.
How are changes in Vmem transduced into alter-ations of mitotic behaviors (Fig. 1)? One likely mechanism is regulation of Ca2+ entry into the cell, and the positive feedback loop that would occur between Ca2+ entry and Ca2+ dependent potassium channels.126 For example, increasing intracellular calcium concentration with the ionomycin restored normal division in cells which were cultured in high potassium media, a condition normally inhibitory to division.28 In addition, Na+ influx is required for the uptake of metabolic substrates and subsequent progression through G1; hyper-polarization of the plasma membrane through potassium channels has been suggested to increase the rate influx and intracellular concentration of Na+.47 However, these links remain poorly understood and further research will be necessary to determine the direct and indirect downstream signal cascades resulting from Vmem flux.
Additional mechanisms for sensing changes in plasma membrane polarization levels include: proteins that change conformation upon Vmem changes and activate integrin-dependent127 or PTEN phosphatase-dependent cascades,128,129 depolarization-dependent nuclear translocation of the NRF-2 transcription factor,130 induction of specific kinases such as KID-1,131,132 and the influx of mitogens such as serotonin,133–137 which is controlled by Vmem through the voltage-powered serotonin membrane transporter SERT and through gap-junctional paths via electrophoresis.138
The control of cell functions by membrane voltage is highly non-linear because the existence of multiple feedback loops (Fig. 2). For example, many of the channels, pumps and gap junctions that determine transmembrane potential are themselves pH- and voltage-sensitive, leading to complex recursion of effects. This is a very powerful mechanism for buffering evolutionary control mechanisms, sometimes resulting in positive feedback loops (such as NFκB, which is turned on by K+ loss, downregulating transcription of the potassium importer HK-ATPase139) as well as negative feedback loops (e.g., depolarization can activate the V-ATPase hyperpolarizing pump). Thus, quantitative mathematical modelling will be necessary to integrate the temporal dynamics of multiple ion fluxes and the resulting Vmem changes and thus develop strategies for placing cells into specific Vmem states in biomedical applications.
The non-linear and non-local aspects of bioelectric signals result in some very interesting features of Vmem control over proliferation during morphogenesis.140–143 For example, mitotic upregulation induced by the V-ATPase is limited to the regenerating region (not the rest of the tadpole) when Xenopus tadpoles regenerate their tails by a voltage-dependent mechanism.125 Alongside this spatial control, temporal control can also be used with high resolution: in depolarization-induced overproliferation of melanocytes, these cells undergo no more than 1 extra cell cycle, despite the continued presence of the depolarizing potassium channel mutant protein.117
In our lab, gain-of-function studies altering bioelectric signals during embryogenesis by misexpression of specific ion transporters and their mutants have revealed a wide variety of subtle phenotypes related to proliferation and differentiation (Fig. 3). The ability to integrate growth control with morphogenesis via bioelectric parameters is only beginning to be understood, and further molecular investigation will surely reveal details of significant importance to biomedicine. Interestingly, we have observed that not only long-term, but also transient, modulation of Vmem has the ability to activate proliferation. Electroporation, a technique in which exogenous DNA is introduced into cells by square electric pulses on a millisecond timeframe, is widely used in cell biology, developmental biology and regenerative medicine.144–148 Surprisingly, the process of electroporation itself, using no DNA, is sufficient to activate not only de-differentiation149 but also hyperproliferation (Fig. 3E–E”’). This is surprising, given the rapid nature of electroporation compared to the time-scale of cell division, and indicates that significant care must be used using any methods that disrupt transmembrane potential.
Membrane voltage is a well-conserved and probably ancient control system, functioning in phyla ranging from plants150 to higher vertebrates. Several areas of this field suggest exciting future advances.
First, physiological parameters such as membrane voltage and specific ion content may be used as in vivo markers to identify special subpopulations of cell types. For example, human mammary tumor cells fall 4 Gaussian distributions of voltage with means of −9, −17, −24 and −2 mV.151 The functional significance and the value of this as a marker remain to be tested, but given the important regulatory roles of Vmem, it is likely that these data are informing us of important distinctions among subtypes of the population.
Second, it is abundantly clear that the original hypothesis of depolarization inducing growth3 is too simple. It is much more likely that types of cells (e.g., embryonic, committed, neoplastic, etc.,) can be sorted into distinct regions in a multi-dimensional state space with axes corresponding to physiological parameters (of which membrane potential is just one). Moreover, because of the Vmem oscillations occurring during the cell cycle, it is clear that temporal variation must be added to models of this signalling system.
Third, it should be noted that assigning a single Vmem value to a cell is also a significant oversimplification. In fact, a number of embryonic blastomeres (Fig. 4A) and mammalian cells in culture (Fig. 4B) exhibit distinct domains of membrane voltage around the cell periphery. While the physiological literature often reports one particular mV reading for a cell this neglects the considerable complexity of microdomains of Vmem on cells, presumably established by distinct population of channel/pump proteins on lipid rafts152,153 and fence functions performed by plasma membrane and cortical/cytoskeletal components. Thus, since different Vmem values can be present in domains as small as 2 µ (reviewed in ref 154), each cell potentially contains a 2-dimensional surface which encodes a tremendous amount of information that can be transmitted to distinct intracellular second-messenger pathways as well as neighboring cells.155
Finally, it must be observed that transmembrane potential is only the best-known and most tractable of the cellular bioelectric parameters. Subcellular organelles such as mitochondria, endo-somes, phagosomes, ER and nucleus all possess a transmembrane potential due to specifically-localized ion channels.156–164 Future efforts must develop subcellularly-targeted voltage reporter proteins165 and mutant channels that can be used to study and functionally alter the bioelectric signalling in distinct intracellular locales.
The detailed understanding of the contribution of transmembrane potential to cell cycle control, and especially the integration of this mechanism into biochemical and genetic mechanisms occurring during complex morphogenesis, will reveal fascinating aspects of interdisciplinary biophysics and will offer significant opportunities for the biomedicine of birth defects, cancer and regenerative medicine.
This paper is dedicated to C.D. Cone, who was one of the first to study in detail the profound role of membrane voltage in proliferation control. We are grateful to the members of the Levin lab and the bioelectricity community for many useful discussions. We thank Dany Adams for the photos in Figure 4, and Ai-Sun Tseng for her comments on the manuscript. M.L. is supported by grants R01-GM077425, EY018168 and GM078484 from the NIH. D.B. is supported by Forsyth Institutes T32 grant 5T32DE007327-07. Kelly McLaughlin is supported by NSF grant IOS-0843355.