| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
EMBRYONIC STEM CELLS |
aDepartment of Cell Biology and Physiology,
bDepartment of Neurology,
cDepartment of Neurological Surgery,
dSpinal Cord Injury Restorative Treatment and Research Program,
eDepartment of Anatomy and Neurobiology, Washington University School of Medicine, St. Louis, Missouri, USA
Key Words. Connexin • Dye coupling • Electron microscopy • Patch clamp • Reverse transcription-polymerase chain reaction
Correspondence: James E. Huettner, Ph.D., Dept. of Cell Biology & Physiology, Washington University Medical School, 660 South Euclid Avenue (Campus Box 8228), St. Louis, Missouri 63110, USA. Telephone: 314-362-6624; Fax: 314-362-7463; e-mail: huettner{at}cellbio.wustl.edu or John W. McDonald III, M.D., Ph.D., Kennedy Krieger Institute, International Center for Spinal Cord Injury, 707 North Broadway, Room 518, Baltimore, Maryland 21205, USA. Telephone: 443-923-9210; Fax: 443-923-9215; e-mail: mcdonaldj{at}kennedykrieger.org
Received January 4, 2005;
accepted for publication March 22, 2006.
First published online in STEM CELLS EXPRESS March 30, 2006.
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresAcknowledgmentsReferences |
|---|
-glycyrrhetinic acid, and arylaminobenzoates inhibited transjunctional currents. Dye uptake studies on human ES cell monolayers and recordings from solitary human ES cells gave evidence for the surface expression of connexon hemichannels. Human ES cells provide a unique system for the study of human connexin proteins and their potential functions in cellular differentiation and the maintenance of pluripotency.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresAcknowledgmentsReferences |
|---|
More than 20 different connexin (Cx) subunits that contribute to vertebrate gap junction channels have been cloned and characterized [5]. Formation of a gap junction involves the end-to-end coupling of two connexon hemichannels, each composed of six connexin subunits, within the surface membranes of adjacent cells. It was long believed that hemichannels outside gap junctions remained closed. However, more recent work indicates that currents or small molecules passing through isolated surface hemichannels may have important biological functions [6].
In rodents, expression of zygotic connexin genes, including cx31, cx40, and cx43, begins as early as the two- to four-cell stage [7, 8]. Widespread coupling throughout the embryo is prominent from compaction at the eight-cell stage through to implantation [3]. Following implantation, cells of the inner cell mass remain coupled to each other, but they lose gap junction-mediated transmission of dye to cells of the surrounding trophectoderm [4]. Further restrictions in coupling arise as development proceeds. Coupling diminishes across the boundaries between germ layers, although cells within each germ layer remain well coupled to each other [9]. These patterns of communication are thought to be important in the regulation of embryonic development, but identifying the salient molecules exchanged through gap junctions early in development has proven difficult.
Pluripotent mammalian embryonic stem (ES) cells, which are derived from the inner cell mass of preimplantation blastocysts, have the capacity to differentiate into cells of all three germ layers [10]. Under suitable conditions, ES cells remain pluripotent through repeated rounds of cell division in culture [11]. The recent isolation of human ES cells [12] has spurred great interest in their potential use for therapeutic tissue repair because appropriate manipulation of the culture environment can induce both mouse and human ES cells to differentiate in vitro into specific somatic cell types [13]. Although brief reports of connexin expression by mouse [14] and human [15–17] ES cells have appeared, the physiological properties of human ES cell gap junctions have not been characterized. Moreover, the roles that gap junctions play in the process of differentiation remain poorly understood.
In the present study, we demonstrate the formation of gap junction channels between human ES cells maintained in vitro. Our results show that human ES cells express RNA encoding most of the known human connexin genes. Electrical and dye coupling is prominent in human ES cell colonies and monolayers and between individual pairs of cells. In addition, human ES cells exhibit currents and dye loading from calcium-free solutions that are characteristic of connexon hemichannels present on the surface membrane. Thus, human ES cells provide a useful system for studying the role of human gap junctional communication during development and cellular differentiation.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresAcknowledgmentsReferences |
|---|
Human ES Cell Culture
Two lines of human ES cells were used in completing these studies: one donated by BresaGen (NIH code, BG01; provider’s code, human ESBGN.01; passage 42; BresaGen, Athens, GA) and the other purchased from WiCell Research Institute (NIH code, WA01; provider’s code, H1; passage 24; WiCell Research Institute, Madison, WI, http://www.wicell.org) and maintained according to the providers’ protocols [12, 15, 20–22]. Chromosome counts confirmed normal complements of 46XY (BG01/ESBGN.01) [21] and 46XX (WA01/H1) [12, 20] and chromosome number stability within each cell line. Studies were carried out on human ES cells with passage numbers ranging from 50 to 72 (BG01) and from 35 to 50 (WA01). Similar data were obtained with both cell lines, but for simplicity, all data presented in this report represents the BresaGen cell line (BG01). Undifferentiated human ES cells were maintained on a pre-existing feeder layer of MEFs (passage 2). The human ES cell culture medium consisted of 80% DMEM/Ham’s F-12 medium (Gibco), 15% FBS (Hyclone), 5% knockout serum replacement (Gibco), 2.0 mM L-glutamine (Gibco), 0.1 mM nonessential amino acids (Gibco), 4 ng ml–1 basic fibroblast growth factor (Sigma-Aldrich), and 0.1 mM ß-mercaptoethanol (Sigma-Aldrich). Cell cultures were incubated at 37°C in 5% CO2 in air and 95% humidity. After 3–4 days, phosphate-buffered saline (PBS)-based cell dissociation buffer (Gibco) was used to lift visually identified colonies of undifferentiated human ES cells off the feeder layer for passage to a new dish of feeder cells. Feeder-free cultures of undifferentiated human ES cells were prepared by transferring human ES cells isolated from feeder layers into feeder-free 35-mm dishes or glass-covered 24-well plates containing human ES cell-conditioned medium. The density of seeding human ES cells was estimated to be 6.0 x 104 ml–1 to 1.0 x 105 ml–1, based on counts of individual cells and small-cell clumps obtained after gentle trituration of isolated human ES cell colonies. For most experiments, contamination by MEF feeder cells was minimized by passaging feeder-free cultures two to four additional times, at 2–3-day intervals, before use. Human ES cell-conditioned medium was collected daily from cultures that contained feeder cells by removing all of the medium from each culture and replacing it with fresh medium. Plates or dishes were incubated with human ES cell-conditioned medium for 30 minutes before use [23].
Chromosome Counts
Chromosome counts of cell line BG01 were performed between passage 50 and passage 72. The cells were grown to 70% confluence in a 25-cm flask and fed with medium containing colcemide (final concentration, 0.02 µg ml–1) 1 hour prior to analysis. Cells were washed with PBS, trypsinized for 5 minutes, and collected by centrifugation (70g, 3 minutes) in a conical tube. Cells were incubated for 6 minutes at room temperature in 5 ml of hypotonic (0.56%) KCl and then fixed by five changes of methanol/acetic acid (3:1; total volume, 1 ml). To create cell spreads, the cell suspension was dropped from at least 1 foot above the surface of the slide and allowed to air dry. Slides were stained with Giemsa for at least 15 minutes, washed with running water for 1 minute, and air-dried. Chromosome spreads were photographed, and representative photos were used to provide chromosomal counts.
Reverse Transcription-Polymerase Chain Reaction
Total RNA was prepared from human ES cells and from MEF cultures using the RNeasy mini kit (Qiagen, Hilden, Germany, http://www1.qiagen.com). Small amounts of contaminating DNA were removed by treatment of the RNA samples with RNase-free DNase I (Promega, Madison, WI, http://www.promega.com) according to the manufacturer’s instructions. Two micrograms of total RNA was used to prepare single-strand cDNA. Omniscript reverse transcriptase (Qiagen) was used for reverse transcription following the manufacturer’s instructions. Polymerase chain reactions (PCRs) were performed according to Di et al. [24], with some modification using HotStarTaq DNA polymerase (Qiagen). Primers for PCR, which are listed in Table 1, were chosen using the PrimerSelect module of Lasergene (DNAStar, Madison, WI, http://www.dnastar.com) based on their specificity among homologous human connexins and their selectivity for human versus mouse connexin sequences. The PCR program was 94°C for 15 minutes to activate the dormant polymerase; followed by 32–36 cycles of 94°C for 1 minute, 60°C for 1 minute (55°C for Cx59 and Cx62), and 72°C for 1 minute; and completed with a final extension of 72°C for 10 minutes. As all paired primers lie within a single exon (most connexins are encoded in a single exon), a negative control for each reverse transcription-polymerase chain reaction (RT-PCR) was run using the same conditions but without reverse transcriptase to eliminate the possibility of genomic DNA contamination. In a 50-µl reaction mixture, 2 µl of 1 ng/µl human placenta DNA (Sigma-Aldrich) was used as a positive control. All primer pairs were also tested in PCRs with mouse genomic DNA (not shown) and MEF cell cDNA to rule out spurious amplification of mouse connexin sequences. RT-PCR products were separated on 1.5% agarose gels. The gels were stained with ethidium bromide and visualized by UV transillumination (Spectroline, Westbury, NY, http://www.spectroline.com). PCR products amplified from human ES cell cDNA were cut from the gel and partially sequenced to verify their identity.
|
|
For double immunofluorescence of rabbit anti-Oct3/4 and mouse anti-human nuclear antigen (Table 2), cells were fixed with 4% paraformaldehyde in 0.12 M sodium phosphate, pH 7.4, rinsed with PBS, permeabilized with 0.1% Triton X-100, and blocked with 10% normal goat serum. Cultures were incubated simultaneously with anti-Oct3/4 (1:200) and anti-human nuclear antigen (1:100) for 1 hour at room temperature. Cells were then rinsed and incubated for 1 hour at room temperature in a mixture of Cy2-conjugated goat anti-mouse IgG and Cy3-conjugated goat anti-rabbit immunoglobulin, both at 1:200 final dilution. Preliminary single immunofluorescence control experiments confirmed the specificity of the secondary antibodies and the adequacy of our filter set to prevent spillover between the Cy2 and Cy3 channels (data not shown).
Electron Microscopy
Feeder-free human ES cells, prepared as described above, were fixed at 5 days in vitro with 2.5% glutaraldehyde in 0.1 M PBS, postfixed in 1% OsO4, block-stained with 1% uranyl acetate, dehydrated in ethanol, and then embedded in polybed 812. Thin sections were cut and stained with uranyl acetate and lead citrate and then examined using a JEOL 100CX transmission electron microscope.
Electrophysiology
Whole-cell recordings were obtained with borosilicate glass pipettes. Passive properties and resting membrane potentials were evaluated using electrodes filled with (in mM) 140 potassium glucuronate, 5 KCl, 10 EGTA, 5 magnesium-ATP, 1 Tris-GTP, and 10 HEPES, pH adjusted to 7.4 with KOH. The open-tip resistance ranged from 1 to 3 MOhm. The bath was perfused at 1–2 ml/minutes with Tyrode’s solution (in mM) 150 NaCl, 4 KCl, 2 MgCl2, 2 CaCl2, 10 D-glucose, and 10 HEPES adjusted to pH 7.4 with NaOH. After seal formation in the Tyrode’s bath, a wide-bore (300 µm) multichannel delivery pipette was positioned to provide local perfusion of various external solutions. Currents were recorded with Axoclamp 200B and/or 200A amplifiers (Axon Instruments/Molecular Devices Corp., Union City, CA, http://www.moleculardevices.com), filtered at 1–5 KHz with the internal Bessell filter of the clamp, and digitized at 10–50 KHz. In-house software controlled the voltage commands and data acquisition to disc. Recordings were accepted for analysis if the seal resistance before breakthrough exceeded 1 GOhm. Immediately after breaking through to the whole-cell configuration the input resistance, series resistance and membrane capacitance were determined from the average of currents recorded during six hyperpolarizing voltage steps. In addition to the potassium glucuronate intracellular and Tyrode’s extracellular solutions described above, permeation through surface hemichannels was evaluated using internal and external solutions of the following compositions.
Internal Solutions. Cesium glucuronate: 140 mM cesium glucuronate, 5 mM CsCl, 5 mM magnesium-ATP, 1 mM Tris-GTP, 10 mM EGTA, 10 mM HEPES; pH adjusted to 7.4 with CsOH.
N-Methyl-D-glucamine chloride: 140 mM N-methyl-D-glucamine (NMDG) chloride, 5 mM magnesium-ATP, 1 mM Tris-GTP, 0.1 mM EGTA, 10 mM HEPES; pH adjusted to 7.4 with N-methyl-D-glucamine.
Tetraethylammonium chloride: 120 mM tetraethylammonium (TEA) chloride, 5 mM MgCl2, 5 mM magnesium-ATP, 1 mM Tris-GTP, 10 mM EGTA, 10 mM HEPES; pH adjusted to 7.4 with TEA hydroxide.
External Solutions. NaCl: 160 mM NaCl, 1 mM calcium glucuronate, 10 mM HEPES; pH adjusted to 7.4 with NaOH.
Sodium glucuronate: 160 mM sodium glucuronate, 1 mM calcium glucuronate, 10 mM HEPES; pH adjusted to 7.4 with NaOH.
NMDG-chloride: 160 mM N-methyl-D-glucamine chloride, 1 mM Ca-glucuronate, 10 mM HEPES; pH adjusted to 7.4 with N-methyl-D-glucamine.
NMDG-glucuronate: 160 mM N-methyl-D-glucamine d-glucuronate, 1 mM Ca-glucuronate, 10 mM HEPES; pH adjusted to 7.4 with N-methyl-D-glucamine.
Calcium-free: 160 mM NaCl, 0.5 mM EGTA, 10 mM HEPES; pH adjusted to 7.4 with NaOH.
In these solutions, small inorganic cations (sodium and cesium) or anions (chloride) are replaced by larger organic cations (TEA, NMDG) or anions (glucuronate), which should pass through the large diameter connexon pore but be impermeable through most conventional voltage-gated ion channels [27]. Mean ± SEM values are reported for all measured parameters. Statistical significance was assigned for p < .05.
Dye Injections
Whole-cell electrodes were used to inject individual cells with fluorescent dyes dissolved in the same internal solution used for electrophysiology experiments. Initially sulforhodamine B (555 Da; Sigma-Aldrich) or Lucifer Yellow CH (443 Da; Sigma-Aldrich) were injected at 0.4%; however, most experiments used aldehyde-fixable AlexaFluor hydrazides. AlexaFluor 488 (570 Da) and 568 (731 Da) hydrazides were purchased as 10 mM solutions in 200 mM KCl (Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com) and were diluted 1:100 into internal solution. For some experiments, AlexaFluor 568 hydrazide was co-injected with Lucifer Yellow coupled to 10 kDa anionic dextran, which is too large to pass through gap junction channels and therefore remains within the injected cell. Cells injected with fluorescent hydrazides were imaged live and in some cases were also fixed for subsequent immunofluorescence. Cultures were fixed in 0.12 M sodium phosphate buffer, pH 7.4, containing 4% paraformaldehyde alone to preserve antigenicity for Oct3/4 and the human nuclear antigen, although our preliminary experiments revealed superior retention of fluorescent AlexaFluor hydrazides by cells that were fixed with a mixture of 4% paraformaldehyde and 0.1% glutaraldehyde.
Pharmacological Agents
Compounds previously demonstrated to inhibit electrical or dye coupling between cells [1] or to block hemichannel-mediated currents [6] were dissolved in external solution and delivered to cells by local perfusion from a mutibarrelled pipette. Octanol [1] (Sigma-Aldrich) was diluted directly into the extracellular solution. Cobalt and lanthanum (Sigma-Aldrich), which inhibit currents mediated by hemichannels [28, 29], were prepared as 0.1 M stocks in water. The arylaminobenzoates niflumic acid and flufenamic acid [6, 30] and the aglycone from licorice (Glycyrrhiza glabra) root, 18--glycyrrhetinic acid [31], and its synthetic analog, carbenoxylone [31], were purchased from Sigma-Aldrich and prepared as 50 mM stocks in DMSO. Addition of 0.2% DMSO alone was found to have no effect on currents through gap junctions or hemichannels.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresAcknowledgmentsReferences |
|---|
|
|
|
|
Recordings from Coupled Cell Pairs
To evaluate cell-to-cell coupling directly [39, 40], we passaged human ES cells at lower density and used two electrodes to record simultaneously from both cells in isolated pairs that were in contact (Fig. 5). Both of the cells were initially clamped to 0 mV. Currents through gap junctions connecting the two cells were recorded by stepping the potential in one cell to values positive and negative to 0 mV, while maintaining the second cell clamped at 0 mV. This procedure was then repeated with the first cell held at 0 mV and the second cells stepped to different potentials. Such experiments yielded transjunctional conductance values that ranged from 2.7 to 79 nS, with a mean of 20 ± 5 nS (n = 15 cell pairs). In all cases, the junctional current reversed polarity at zero transjunctional potential, and there was good agreement in the junctional conductance as determined for steps applied to either of the two cells (Fig. 5E). Linear regression yielded a correlation coefficient of r = .996. Current voltage relations for whole cell current and for transjunctional current were relatively linear for transjunctional voltages that ranged from –30 to +30 MV.
|
-glycyrrhetinic acid, long-chain alcohols (e.g., octanol and heptanol), and arylaminobenzoates (e.g., flufenamic acid and niflumic acid), have been shown to reduce cell-to-cell coupling mediated by gap junctions. Although some of these drugs are not entirely specific for gap junction channels [1, 6, 30], they are widely used to identify coupling that is mediated by gap junctions and to probe the functional consequences of pharmacologically induced uncoupling on cell phenotype or survival. We initially tested the effect of 100 µM 18--glycyrrhetinic acid [31], flufenamic acid, and niflumic acid [30], as well as 1 mM octanol [1], 2 mM cobalt chloride, and 100 µM lanthanum chloride [1] on the transjunctional conductance between coupled cell pairs (Fig. 6). Lanthanum and cobalt caused no significant change in conductance, whereas all of the organic inhibitors reduced conductance. Onset and recovery of block by flufenamic acid was rapid compared with the slower kinetics observed for octanol, 18--glycyrrhetinic, and niflumic acid. These results are broadly consistent with previous pharmacological characterization of native and recombinant gap junction channels [1]. Inhibition of current flow by these compounds, together with the selective passage of low molecular weight dyes described above (Fig. 4A), provides the strongest evidence that the low resistance connections between adjacent cells are formed by gap junctions, rather than by incomplete mitosis or some other type of cytoplasmic continuity.
|
To test for the presence of functional hemichannels, we recorded from isolated human ES cells. In contrast to human ES cells growing in monolayers, which display low input resistance owing to the strong coupling to neighboring cells, the isolated human ES cells (n = 31) had significantly higher input resistance (717 ± 111 MOhm, p < .01) and lower membrane capacitance (60 ± 5.6 pF, p = .014), as well as a less negative mean zero current potential (–16 ± 1.3 mV, p < .01). As shown in Figure 7, depolarizing voltage steps evoked slowly activating outward currents with properties similar to those previously described for native and recombinant hemichannels.
|
Brief voltage steps from holding potentials of –80 or +60 mV (Fig. 8A) were used to evaluate the instantaneous current-voltage (I-V) relation for hemichannel-mediated whole-cell currents. As shown in Figure 8B, the instantaneous I-V of open hemichannels was linear and reversed near 0 mV. With external Tyrode’s solution, the whole-cell conductance through surface hemichannels in isolated human ES cells ranged from 3.6 to 90.5 nS, with a mean of 40.2 ± 5.6 nS (n = 20 cells). To evaluate the ionic selectivity of channels underlying these currents we substituted large organic cations (NMDG and TEA) or anions (glucuronate) for the small diameter inorganic ions (potassium, cesium, sodium, and chloride) present in our standard intracellular and extracellular solutions. Consistent with the nonselective permeation properties of connexons, we observed relatively linear I-V relations, which reversed between –10 and +10 mV, using a variety of different internal and external solutions (Fig. 8D). These included internal solutions containing potassium glucuronate, cesium glucuronate, NMDG-chloride, or TEA-chloride as the major intracellular salts, and paired with NaCl, NMDG-chloride, sodium glucuronate, or NMDG-glucuronate as the major extracellular salts (Materials and Methods). The linear instantaneous I-V relations, and the lack of any dramatic change in potential at which current reversed polarity, indicate that the large organic ions pass through the channels as readily as the smaller inorganic ions. This stands in contrast to most conventional voltage-gated channels, which typically are selective for smaller inorganic ions [27].
|
Evaluation of Junctional Versus Nonjunctional Connexin Distribution in Cell Pairs
It is of interest to ask whether the hemichannel-mediated currents recorded in isolated cells simply reflect the absence of a suitable partner for cell-to-cell junction formation or whether cells that have formed junctions with their neighbors might retain a significant proportion of unpaired connexons on their surface. To evaluate this question, we again recorded from pairs of cells, but both cells were stepped simultaneously to the same positive potential, relative to the bath. This protocol will maintain the transjunctional voltage at 0 mV and hence eliminate junctional current. Out of 22 cells (11 pairs) tested in this way, six cells displayed slowly activating outward currents characteristic of connexon hemichannels, and all six of these cells were paired with a cell that lacked hemichannel-mediated current. In these six cells, the ratio of surface hemichannel conductance to transjunctional conductance ranged from 0.36 to 2.1, with a mean of 1.14 ± 0.32 (n = 6 cells). Linear regression for hemichannel conductance and transjunctional conductance yielded a correlation coefficient of r = .742. These results suggest that connexons preferentially combine to form gap junctions, but hemichannels will be present in cells that express an excess of connexons, outnumbering the available partners on neighboring cells.
Dye Loading Through Surface Membrane Hemichannels
In addition to depolarizing voltage steps, opening of cell surface hemichannels is favored by removal of extracellular calcium [28, 41]. As shown in Figure 9A and 9B, local perfusion of isolated human ES cells with calcium-free external solution evoked a reversible increase in conductance. Previous studies have shown that cells expressing hemichannels on their surface membrane can be loaded with dye dissolved in calcium-free medium [49]. To determine whether hemichannels were present on the surface of human ES cells in high-density monolayers, we exposed cells to sulforhodamine B or propidium iodide in calcium-free Tyrode’s solution. As demonstrated in Figure 9C–9F, a subset of cells contained the dye after a brief incubation in calcium-free, but not in calcium-containing, solutions. As mentioned above, our recordings from cell pairs suggest that hemichannels are present on cells that express a higher number of connexons than their nearest neighbors. Thus, the uneven distribution of dye uptake is likely to reflect the location of cells with a higher relative level of connexin expression than surrounding cells.
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresAcknowledgmentsReferences |
|---|
Mammalian embryos express a large number of different connexins. Studies in rodents have reported zygotic expression of transcripts for Cx30, Cx31, Cx31.1, Cx36, Cx40, Cx43, Cx45, and Cx57 [7, 53–55]. Zygotic transcription of Cx26 and Cx32 cannot be detected in mouse embryos [53, 54], although Cx32 protein of maternal origin persists in mice up to implantation [56]. In contrast, rat embryos transcribe Cx26 [55], raising the possibility of species variability in the profile of connexin expression. Our results in human ES cells are broadly consistent with these studies on the embryonic pattern of connexin expression and with earlier brief reports of evidence for Cx43 and Cx45 expression by mouse [14] and human [15–17] ES cells. Transcripts for Cx40.1 and Cx50 were not observed in human ES cells by RT-PCR, whereas PCR products for Cx25, Cx26, Cx30.3, Cx31, Cx31.1, Cx31.9, Cx37, Cx40, Cx43, Cx45, and Cx46 were detected for all human ES cultures sampled, and somewhat weaker or more variable expression was detected for Cx30, Cx30.2, Cx32, Cx36, Cx47, Cx59, and Cx62. Consistent with prior immunofluorescent visualization of connexins in human embryos [51, 52], antibodies to Cx40, Cx43, and Cx45 labeled points of contact between adjacent human ES cells. Thus, undifferentiated human ES cells express most of the known connexins at the level of RNA, and at least some of these are translated into protein and delivered to the surface membrane at points of intercellular contact.
Cell-to-cell coupling via gap junctions is a common feature of embryonic development, although the role of embryonic cell junctions remains elusive. Direct evidence for electrical coupling among embryonic cells was first described in squid embryos [57] and was subsequently demonstrated in a variety of other species. In both mouse [58–60] and human [51] embryos, organization of connexins into gap junctions can be visualized morphologically, beginning at the four- to eight-cell stage. In mouse embryos, evidence for functional coupling can first be obtained at the eight-cell stage [3], whereas in human embryos, extensive dye coupling arises somewhat later [50]. Initially, all cells within the embryo are coupled electrically, and dye injected into one cell quickly spreads to all others. As differentiation proceeds, there is a progressive loss of communication between different tissues [3, 9], whereas coupling within a given compartment remains strong. For example, following implantation, cells in the inner cell mass remain well coupled, but spread of fluorescent dye to the surrounding trophoblast diminishes [3]. Similarly, cells within each of the early germ layers remain coupled to each other, whereas dye fails to spread across the boundaries between germ layers. Electrical coupling may persist after dye coupling has been lost [3, 9], but there is clearly a change in communication as tissues differentiate in vivo. In human ES cell monolayer cultures, dye injected into individual cells spreads quickly and uniformly to all surrounding cells. Therefore, cultured human ES cells remain extensively coupled through repeated passages, preserving the communication pathways that exist within the inner cell mass. In mixed cultures that included both human ES cells and mouse embryonic fibroblast feeder cells, we observed dye coupling within each population; however, coupling between human ES and MEF cells was extremely rare. These results suggest that preservation of human ES cell pluripotency does not require direct transmission of signals from MEF cells to human ES cells via gap junction channels, but they leave open the possibility that coupling between adjacent human ES cells may play a role in coordinating cellular responses to external signals that maintain pluripotency or that stimulate differentiation.
Indeed, the function of embryonic cell-to-cell communication has not been determined [61]. Single and double knockout mice, which lack one or more individual connexins, exhibit relatively normal early development [62–65], although compensation by other connexins expressed in the embryo might explain the lack of an obvious phenotype prior to implantation. Several early reports provided evidence for developmental abnormalities following acute blockade of embryonic gap junctions by injection of blocking antibodies [66] or connexin antisense RNA [67]. More recent pharmacological work has tested the effect of chronic exposure to the gap junction inhibitor 18- alpha glycyrrhetinic acid. Studies of intact or chimeric mouse inner cell mass reported essentially normal differentiation up to the blastocyst stage during continuous exposure to the drug [55, 68]. On the other hand, 18- glycyrrhetinic acid was found to inhibit in vitro differentiation of human NT2/D1 cells [69], which are multipotent teratocarcinoma cells that share a number of properties with pluripotent ES cells. In addition, recent studies of transformed cell lines have raised the possibility that Cx43 may have growth regulatory effects that are independent of its ability to form functional gap junctions [70, 71]. Clearly, the functional implications of connexin expression on human ES cell growth and differentiation merit further investigation.
Previous studies in both native and recombinant systems have documented a number of differences in the properties of gap junctions or hemichannels formed by different connexins, including differences in kinetics, unitary conductance, and permeability for specific dyes [45, 72, 73]. The large number of connexin RNAs expressed in embryos and human ES cells suggests that, in both cases, the cells may produce heteromeric connexons containing two or more distinct connexin subunits. Further work will be needed to quantify the relative connexin protein levels and to determine the stoichiometry of connexins that combine to form connexon hexamers in these cells.
In general, strong depolarization or reduction in extracellular calcium promotes the opening of surface membrane hemichannels [28, 41]. Prolonged depolarization of human ES cells to membrane potentials of +30 mV or greater elicited slowly activating outward currents typical of connexon hemichannels. The nonselective nature of this conductance was verified by comparing instantaneous I-V relations for internal and external solutions of diverse composition. For all solutions tested, the current evoked by holding at positive potentials was essentially linear and reversed between –10 and +10 mV, indicating similar permeability to both small inorganic and larger organic ions. Removal of extracellular calcium produced a significant increase in membrane conductance and allowed for dye entry from the extracellular medium, consistent with opening of surface hemichannels in calcium-free solutions, as observed in other cell types [49]. In medium that contained 1–2 mM calcium, currents through human ES cell hemichannels were minimal when cells were maintained at their normal zero current potential between –15 and –30 mV, suggesting that most hemichannels remain closed in cells at rest.
Investigation of human ES cells is still at an early stage. Identifying characteristics of human ES cells are currently limited to the expression of a few relatively specific markers and the ability to differentiate into cells from all three germ layers. The gold standard for proof of an undifferentiated nonhuman embryonic stem cell phenotype is the ability to generate a chimeric animal following injection into the blastocyst inner cell mass. This of course is not a testable option for human ES cells. Our data provide additional ultrastructural, immunocytochemical, and physiological properties that may aid in proper confirmation of undifferentiated human ES cells. In addition to expression of the Oct3/4, SSEA-3, SSEA-4, TRA-1-60, and TRA-1-81 antigens, the expression profile for connexins (positive for Cx25, Cx26, Cx30.3, Cx31, Cx31.1, Cx31.9, Cx37, Cx40, Cx43, Cx45, and Cx46; negative for Cx40.1 and Cx50) and physiological dye coupling can be added to the growing list of human ES cell characteristics in the undifferentiated state. Collectively, our results establish human ES cells as a good model system for the investigation of direct intercellular communication channels during early stages of differentiation.
|
DISCLOSURES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresAcknowledgmentsReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresAcknowledgmentsReferences |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresAcknowledgmentsReferences |
|---|
