2-Aminoethoxydiphenyl borate (2-APB), an inositol 1,4,5-triphosphate receptor modulator, inhibits capacitive current transients measured in normal rat kidney and human embryonic kidney 293 cells, an indication of blocking gap junction channels between these cells. Here, we used the dual whole-cell patch-clamp method to study the actions of 2-APB on gap junction channels formed by selected connexins expressed in a communication-deficient neuroblastoma cell line (N2A). 2-APB dose-dependently and reversibly blocked junctional currents of connexin (Cx) 50 gap junction channels. The concentration-inhibition curve of 2-APB on the junctional current indicated an IC50 of 3.7 μM, lower than that of most gap junction inhibitors. At a concentration of 20 μM, 2-APB also significantly blocked junctional conductance in cell pairs coupled by Cx26, Cx30, Cx36, Cx40, and Cx45 but did not appreciably affect coupling in cell pairs expressing Cx32, Cx43, and Cx46. Although concentration inhibition curves of 2-APB on Cx36 channels were similar to Cx50 (Cx36; IC50, 3.0 μM), IC50 values were higher for Cx43 (51.6 μM), Cx45 (18.1 μM), and Cx46 (29.4 μM). The blocking action of 2-APB did not substantially alter transjunctional voltage-dependent gating of Cx50 gap junction channels, and recordings from poorly coupled pairs of Cx50-transfected N2A cells indicated that 2-APB reduced gap junction channel open probability without changing the main state single-channel conductance. The differential efficacy of block by 2-APB of gap junction channels formed by different connexins may provide a useful tool that could be exploited in gap junction research to selectively block certain gap junction channel subtypes.
Gap junction channels connect the cytoplasm of adjacent cells, allowing direct intercellular exchange of ions and small signaling molecules with molecular masss below approximately 1 kDa. The molecular constituents of gap junction channels in vertebrates are a family of approximately 20 homologous proteins called connexins (Willecke et al., 2002). Gap junction channels not only play important functional roles in numerous physiological processes, but they may also provide a pathway for spread of toxic molecules, so-called “bystander cell killing” (Lin et al., 1998; Cusato et al., 2003). The electrical and metabolic coupling that gap junctions provide may thus be detrimental under certain conditions, such as in the ischemic brain and myocardium (Lin et al., 1998; Cusato et al., 2003) and in conditions of hyper- or desynchronized excitability, such as epileptogenesis and certain arrhythmias (Nakase and Naus, 2004; Varro et al., 2004). Compounds that block gap junctions might thus be expected to be usefully cardio- or neuroprotective in these pathophysiological conditions.
Highly potent and selective channel-blocking molecules have played key roles in elucidating both structural and functional features of other membrane channels. However, only very few drugs have been identified that block gap junction channels, and most of these are nonspecific, acting on other channel types. Among the classes of molecules that have been shown to reduce gap junction conductance are long-chain alcohols (Johnston et al., 1980; Burt and Spray, 1989); volatile anesthetics (Burt and Spray, 1989); glycyrrhetinic acid derivatives (Davidson and Baumgarten, 1988); oleamide (Guan et al., 1997); aminosulfonates (Bevans and Harris, 1999); tetraalkylammonium ions (Musa et al., 2001); arylaminobenzoates (Harks et al., 2001; Srinivas and Spray, 2003); polyamines (Musa and Veenstra, 2003); weak acids, which act by intracellular acidification (Spray et al., 1984); and certain antimalarial compounds, including quinine and derivatives such as mefloquine (Srinivas et al., 2001; Cruikshank et al., 2004). With the exception of intracellular acidification (where apparent pKas for different connexins vary over a range of approximately 1 pH unit) (Stergiopoulos et al., 1999), polyamines (which blocks connexin (Cx) 40 but not Cx43 channels) (Musa and Veenstra, 2003), and quinine and its derivatives (which potently block Cx50 and Cx36 channels while sparing gap junctions formed of other connexins at even much higher concentrations; Srinivas et al., 2001; Cruikshank et al., 2004), uncoupling agents show poor selectivity for gap junctions formed by different types of connexins.
A recent study reported that 2-aminoethoxydiphenyl borate (2-APB), a membrane permeable modulator of inositol 1,4,5-triphosphate (IP3) receptors (Maruyama et al., 1997) that has been of widespread recent use in blockade or activation of transient receptor potential (TRP) channels (Hu et al., 2004; Xu et al., 2005), blocked capacitive current transients in normal rat kidney (NRK) cells (IC50: 5.7 μM) and in human embryonic kidney (HEK) 293 (tsA201) cells (IC50, 10 μM), an indication of blocking the gap junction channels (Harks et al., 2003). Although 2-APB (at 100 μM) was recently used as a “specific” gap junction channel blocker in critical control experiments demonstrating peptide permeation through gap junctions (Neijssen et al., 2005) and at lower concentrations (10 μM) to disrupt intercellular communication in vascular wall (Griffith et al., 2005), direct evaluation of its action on junctional conductance and the selectivity of 2-APB for gap junction channels formed by distinct connexin subtypes has not been investigated previously. Here, we employed dual whole-cell patch-clamp recording methods to study the actions of 2-APB on junctional conductance in pairs of communication-deficient neuroblastoma cells (N2A) expressing homomeric gap junction channels formed of Cx26, Cx30, Cx32, Cx36, Cx40, Cx43, Cx45, Cx46, and Cx50. Our data demonstrate that 2-APB is a potent uncoupling agent for gap junction channels formed by certain connexins (in particular Cx36 and Cx50), whereas this drug is much less effective when applied to gap junctions formed by other connexins.
Materials and Methods
Neuroblastoma (N2A) cells were stably transfected with Cx50 or transiently transfected with Cx26, Cx30, Cx32, Cx36, Cx40, Cx43, Cx45, and Cx46, as described previously (Srinivas et al., 1999; Srinivas and Spray, 2003). Transfected N2A cells were plated at low density onto 1-cm-diameter glass coverslips. Coverslips were placed in a recording chamber on an inverted phase-contrast microscope (Nikon Diaphot; Nikon, Tokyo, Japan). The cells were constantly perfused (approximately 1.5 ml/min) with an external solution containing 140 mM NaCl, 5 mM KCl, 2 mM CsCl, 2 mM CaCl2, 1 mM MgCl2, 5 mM HEPES, 5 mM d-glucose, 2 mM pyruvate, and 1 mM BaCl2, pH 7.4. 2-APB was dissolved in a stock solution of 10 mM in equal molar NaOH solution or at 100 mM in dimethyl sulfoxide and diluted into the final concentrations on the day of experiment, with pH adjusted as necessary. Drug was delivered through a gravity-fed perfusion system with a solution exchange time of approximately 20 to 30 s. Junctional current was measured between cell pairs using the dual whole-cell voltage clamp technique with Axopatch 1D patch-clamp amplifiers (Axon Instruments Inc., Union City, CA) at room temperature (22–25°C). Patch pipettes prepared with a puller (P97 Flaming/Brown micropipette puller; Sutter Instrument Co., Novato, CA) had resistances of 3 to 5 MΩ when filled with internal solution containing 130 mM CsCl, 10 mM EGTA, 0.5 mM CaCl2, 3 mM MgATP, 2 mM Na2ATP, and 10 mM HEPES, pH titrated to 7.2 with CsOH. Macroscopic and single-channel recordings were filtered at 0.2 to 0.5 kHz. Data were acquired using pClamp6 software (Axon Instruments Inc.) and digitized at 1 to 2 kHz sampling rate. Each cell of a pair was initially held at a holding potential of 0 mV. To evaluate junctional coupling, 300-ms hyperpolarizing pulses from the holding potential of 0 mV were applied to one cell to establish a transjunctional voltage gradient (Vj), and the junctional current was measured in the second cell. Macroscopic junctional conductance normalized to initial values at the beginning of the experiment (Gj) was calculated as Gj = Ij/Vj, where Ij is the measured junctional current, and Vj is transjunctional voltage. One to several concentrations of 2-APB were used for each pair of cells to determine the concentration-inhibition curve. The IC50 and Hill coefficient were determined using Origin software (Microcal Software, Inc., Northampton, MA) as described previously (Srinivas et al., 1999; see Fig. 5 legend). To evaluate the voltage dependence of the junctional conductance, 7.5-s hyperpolarizing and depolarizing pulses (–100 to 100 mV, 20-mV increments) were applied every 30 s from the holding potential of 0 mV. Only those pairs where Gj was between 2 and 8 nS were used to measure macroscopic voltage sensitivity to minimize the impact of uncompensated series resistance (Moreno et al., 1991). The instantaneous and steady-state levels of junctional current were measured at the beginning and at the end of each Vj pulse. Steady-state junctional current at each voltage was normalized relative to the instantaneous current, and these Gj,ss values were plotted as a function of the transjunctional voltage. The relationship between Gj,ss and Vj was fitted with a two-state Boltzmann equation: where V0 is the voltage at which the conductance is half-maximal, Gmax is the maximal normalized conductance, Gmin is the normalized voltage-insensitive residual conductance, and A is a parameter defining the steepness of voltage sensitivity (Spray et al., 1984).
Weakly coupled pairs of Cx50-transfected N2A cells, where unitary gap junction channel-mediated currents were readily identifiable, were analyzed further with Fetchan and Pstat software for display of all-points amplitude histograms and calculation of single-channel conductances (Axon Instruments Inc.). Vj was held at –30 mV to measure unitary currents. Unitary conductance was also measured by fitting a linear function to the single-channel current-voltage relationship, determined by applying slow voltage ramps from –100 to + 100 mV in 15 s to one cell of the pair and recording the current response from the other cell. Repeating the ramp several times was necessary to obtain an adequate number of points to determine the conductance of the main open state and the subconductance state.
Data are presented as means ± S.E.M. Paired Student's and nonpaired t tests were used to compare the significance of differences between paired and nonpaired groups of data, respectively.
2-APB Dose-Dependently and Reversibly Blocks Cx50 Gap Junction Channels Expressed in N2A Cells. Dual whole-cell voltage-clamp was used to directly measure transjunctional current between pairs of Cx50-transfected N2A cells. Under control conditions, repeated –30 mV, 300-ms voltage pulses in one cell of the pair-induced transjunctional currents that were relatively stable, whereas cells were perfused with external solution (Fig. 1A, initial 1-min recording segment). Bath application of 100 μM 2-APB (molecular structure and duration of exposure indicated in Fig. 1A) resulted in rapid elimination of the transjunctional current as shown in Fig. 1A. Complete recovery of transjunctional current was observed within a few minutes after washout of 2-APB (break in time axis corresponds to 4 min), whereas lower concentrations of 2-APB (1–10 μM) produced a partial block of junctional currents. No significant block of Cx50 gap junctional current was observed with concentrations of 0.1 or 0.3 μM 2-APB. The dose-response plot to seven distinct 2-APB concentrations was constructed, and data points were fit with a Hill equation (Fig. 1B). The estimated IC50 was 3.4 μM, indicating that 2-APB is a quite potent uncoupling agent for Cx50 gap junction channels. The Hill coefficient was 0.91, suggesting little cooperativity in the process of channel blockade.
2-APB Failed to Change Voltage-Dependent Gating of Cx50 Channels. Junctional conductance of channels formed by Cx50 shows a strong dependence on transjunctional voltage, where steady-state junctional conductance is well described by a two-state Boltzmann equation (Srinivas et al., 1999). To evaluate whether this voltage-dependent gating of Cx50 gap junction channels is substantially altered during the block by 2-APB, we applied doses of 2-APB (3–10 μM) that blocked 50 to 80% of junctional conductance and compared steady-state voltage sensitivity with that before partial blockade. Our data indicate that although a substantial number of channels were blocked by 2-APB, the remaining gap junction channels demonstrated similar voltage dependence of junctional voltage (Fig. 2A). The best fit for the Boltzmann equation for the average Gj,ss was obtained for both positive and negative polarities. The following parameters were obtained: for negative voltages, Gmin, 0.27; V0, –38 mV; and A, 0.13; and for positive voltages, Gmin, 0.29; V0, 40 mV; and A, 0.13. These values are very similar to those observed in Cx50 gap junction channels in control conditions (Fig. 2B, dotted line; Srinivas et al., 1999). In addition, kinetics of channel closure were calculated in the presence of 2-APB. Time constants were at +60 (0.85 ± 0.20 s), +80 (0.45 ± 0.15 s), –60 (0.69 ± 0.08 s), and –80 (0.59 ± 0.24 s) mV, and these time constants are not statistically different from those reported previously under control conditions (Srinivas et al., 1999).
2-APB Blocked Cx50 Channels Mainly by Reducing Popen, Not Single-Channel Conductance. To evaluate the blocking actions of 2-APB at the level of individual gap junction channels, we recorded unitary current events in weakly coupled pairs of Cx50 transfectants. As shown in Fig. 3A, single and multiple unitary current events were readily identifiable with nearly identical single-channel conductance estimated from the mean variance difference at each level of open or the closed state (Fig. 3B). Application of 2-APB caused a rapid reduction in the number of channels remaining in the open state.
During progressive blockade by 2-APB and in the early phase of recovery, we did not detect any significant change in the amplitude of unitary currents, indicating that 2-APB did not appreciably alter the single-channel conductance of Cx50 gap junction channels. In addition to the main conductance state, a subconductance state of single Cx50 gap junction channels was also observed that was prominent in unitary currents recorded in response to Vj ramps ± 100 mV (Fig. 3C; see also Srinivas et al., 1999). Both main state and substate conductances of Cx50 gap junction channels were completely blocked by 2-APB.
2-APB Selectively Blocks Gap Junction Channels Formed of Different Connexins. We tested whether 2-APB could also affect gap junction channels formed of other connexins. Typical results of exposures of Cx36, Cx45, Cx46, and Cx43 gap junction channels to 2-APB concentrations ranging from 5 to 100 μM are illustrated in Fig. 4. 2-APB concentrations as low as 5 μM substantially blocked Cx36 channels; both the extent and the speed of onset of blockade increased with dosage, such that 20 μM 2-APB decreased junctional conductance (Gj) by >80% within 3 min of application (Fig. 4, top). Gap junctional conductance between Cx45-transfected N2A cells was less sensitive to 2-APB (Fig. 4, middle); 20 μM 2-APB reduced Gj by 50% or less, whereas 50 to 100 μM 2-APB led to an approximately 75 to 80% decrease. Cx43 channels were even less sensitive to 2-APB (Fig. 4, bottom), with even 100 μM not achieving complete inhibition. As illustrated in the response of Cx43 channels to 50 and 100 μM 2-APB, recovery of the less sensitive connexins to the highest concentrations of 2-APB was slow and in many cases incomplete.
A comparison of the degree of inhibition of gap junction channels formed in N2A cells transfected with Cx26, Cx30, Cx32, Cx36, Cx40, Cx43, Cx45, Cx46, and Cx50 to 20 μM 2-APB is shown in Fig. 5A. At this concentration, 2-APB inhibited Gj in Cx36, Cx40, and Cx50 transfectants by >95%; Cx26, Cx30, and Cx45 by approximately 50%; and Cx32, Cx43, and Cx46 by 20% or less. To characterize the responses of 2-APB-sensitive and -insensitive connexins in more detail, dose-inhibition curves were constructed by treating Cx36-, Cx43-, Cx45-, and Cx46-transfected cell pairs with 2-APB concentrations ranging from 0.1 to 100 μM (Fig. 5B). IC50 and Hill coefficient values obtained by fitting the concentration-inhibition curves to Hill equations are shown in Table 1. Hill coefficients (nH) for all fits were in the range between 0.9 and 1.4 for each of the connexins, indicating a low degree of cooperativity in channel closure by this drug for the tested gap junction channels.
2-APB was initially shown to be an antagonist for IP3 receptors, blocking IP3-induced calcium release in cerebellar microsomes with an IC50 of 42 μM (Maruyama et al., 1997) and blocking the IP3 receptor in intact cells at 75 μM (Ma et al., 2000). More recent studies have shown that 2-APB is an even more potent blocker for store-operated calcium entry than for IP3-induced calcium release (for review, see Bootman et al., 2002). Moreover, 2-APB at 100 μM concentrations has been found to block volume-regulated anion channels (Lemonnier et al., 2004), to cause mitochondrial swelling (Peppiatt et al., 2003), and to activate calcium permeable cation channels (Braun et al., 2003). In addition, 2-APB was recently reported to exert opposite effects on members of the TRP gene family, decreasing activity of TRP5 and TRP6 at 20 μM (Xu et al., 2005) and activating TRPV1, TRPV2, and TRPV3 at higher concentrations (Hu et al., 2004).
A recent study revealed an unexpected inhibitory action of 2-APB on gap junction intercellular communication. 2-APB effectively blocked transient capacitive currents in confluent monolayer HEK293 and NRK cells, an indication of blockade of gap junction channels (Harks et al., 2003). Although the connexins expressed in these cell lines have not been extensively studied, HEK293 cells have been shown to express Cx45 (Butterweck et al., 1994) and NRK cells to express Cx43 (Musil and Goodenough, 1991). The rather similar IC50 values for blockade of HEK293 cells (10 μM) and NRK cells (5.7 μM) reported by Harks et al. (2003) are not as expected from our measurements of differential sensitivity of these connexins to 2-APB. The IC50 values we obtained for both Cx45 and Cx43 were considerably higher (18.1 and 51.8 μM, respectively). Note that Hashitani et al. (2004) reported blockade of bladder smooth muscle coupling at 100 μM and that Niejssen et al. (2005) used the same concentration for their experiments, demonstrating peptide permeation through gap junctions in Cx43 transfectants; however, Griffith et al. (2005) used 10 μM for their studies on vascular smooth muscle. There are several possible explanations for the disparity between IC50 values in the Harks et al. (2003) and values for the individual connexins presented here. First, coupling strength in the previous study was measured as change in whole-cell capacitance, which would be expected to be affected both by coupling to other cells and by changes in nonjunctional properties; however, in the Harks et al. study junctional conductance measured in NRK cells disappeared completely in response to 50 μM 2-APB, whereas changes in nonjunctional conductances were minimal. A second possibility for the higher sensitivity of the cell lines studied by Harks et al. is that cytoplasmic environment or lipid composition of the membrane between cells used for the evaluation may lead to differences in potency of 2-APB. Finally, the connexin composition of the NRK and HEK293 cells may be more complex than has been reported, with additional connexins altering the behavior observed in this study for Cx43 or Cx45 homomeric channels.
Our data demonstrate for the first time that 2-APB is a potent gap junction blocker of Cx50 and Cx36 gap junction channels expressed in N2A cell pairs. This blocking action of 2-APB is dose-dependent and reversible. In addition, as demonstrated for Cx50, 2-APB action is independent of voltage gating and does not affect single-channel conductance. The Hill coefficient for blockade of all connexins studied in detail was found to be approximately 1, suggesting little cooperativity in channel closing. For the gap junction channel blockers quinine, mefloquine, and flufenamic acid, we have reported Hill coefficients of between 2 and 3 (Srinivas et al., 2001; Srinivas and Spray, 2003; Cruikshank et al., 2004). Whether the different Hill coefficients and potencies of these compounds imply separate binding sites and gating mechanisms remains to be carefully explored.
The sensitivity of gap junctions formed by certain connexins to blockade by 2-APB is higher than to any previously reported gap junction inhibitor except mefloquine (Cruikshank et al., 2004) and is also equal or higher than that of the IP3 receptor or store-operated channels. The high potency of 2-APB for certain connexins suggests that this compound might usefully be modified in the pursuit of even more potent and highly specific gap junction-blocking molecules.
Efficacies of channel blockade have been compared previously for several gap junction inhibitors applied to several gap junction channel types. Although it is difficult to compare the efficacies of low-affinity ligands such as the alcohols, volatile anesthetics, and glycyrrhetinic acid derivatives that have been reported to block gap junction channels (see Spray et al., 2002), both quinine and mefloquine have been shown to produce connexin-specific blockade (Stergiopoulos et al., 1999; Srinivas et al., 2001; Musa and Veenstra, 2003; Cruikshank et al., 2004). For intracellular acidification, the order of sensitivity determined in the most complete study thus far by a single laboratory (Stergiopoulos et al., 1999) is Cx50 (pK, 7.2; n = 8) > Cx46 (pK, 7; n = 2.5) ∼ Cx45 (pK, 7; n = 7) ∼ Cx26 (pK,7; n = 4) > Cx37 (pK, 6.9; n = 3.2) > Cx43 (pK, 6.7; n = 5.3) ∼ Cx40 (pK, 6.7; n = 3.3) > Cx32 (pK, 6.5; n = 6). It is possibly noteworthy with regard to similar sensitivity of Cx36 and Cx50 to several other blocking agents that the sensitivity of Cx36 to intracellular acidification is lower than that of these other connexins (M. Srinivas and D.C. Spray, unpublished data). For quinine, the order of sensitivity for the connexins examined was Cx36 > Cx50 ≫ Cx45 > Cx43 > Cx26 ∼ Cx40 ∼ Cx32, with n = 1.6 and 1.9 for the two most sensitive connexins (Srinivas et al., 2001); for mefloquine, the order of sensitivity was Cx36 > Cx50 > Cx43 > Cx26 ∼ Cx32 > Cx46 (9). The data presented in this manuscript indicate that for all connexins tested, the order of sensitivity to a dose of 2-APB (20 μM) was Cx36 ∼ Cx40 ∼ Cx50 > Cx45 ∼ Cx26 ∼ Cx30 > Cx46 ∼ Cx43 > Cx32. Thus, 2-APB differs most strongly from proton-induced junctional channel closure with regard to Hill coefficients and the major differences in sensitivity of Cx36 and Cx50 to acidification. Compared with the antimalarials, 2-APB differs both with regard to Hill coefficient and also to the effects on Cx43. Although Cx43 is moderately sensitive to the antimalarials, it is the least sensitive connexin tested in response to 2-APB.
2-APB has been shown to block slow calcium waves in myometrial (Ascher-Landsberg et al., 1999), intestinal (Hirst et al., 2002; Varro et al., 2004), and urinary bladder (Imai et al., 2002) smooth muscle, to decrease spontaneous calcium elevations in astrocytes in slice preparations (Parri and Crunelli, 2003), and to reduce spinal cord damage in response to injury (Thorell et al., 2002). Although each of these findings has been interpreted in the context of blockade of the IP3 receptor, it should be noted that each result might be expected as a consequence of gap junction channel blockade. For example, a recent thorough study (Hashitani et al., 2004) demonstrated blockade of urinary bladder smooth muscle Ca2+ waves by 2-APB and other gap junction blockers, but no effect was observed when intracellular Ca2+ release was selectively blocked by thapsigargin and xestospongin.
This work was supported by National Institutes of Health Grants MH65495, NS34931, and DK41918 (to D.C.S) and EY13869 (to M.S.), by a Canadian Institutes of Health Research grant (to D.B.), by the Canada Research Chair program (to D.B.), and by a Mini Fellowship, Faculty of Medicine and Dentistry, University of Western Ontario (to D.B.).
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
ABBREVIATIONS: Cx, connexin; 2-APB, 2-aminoethoxydiphenyl borate; IP3, inositol 1,4,5-triphosphate; TRP, transient receptor potential; NRK, normal rat kidney; HEK, human embryonic kidney; N2A, neuro-2A cells (neuroblastoma cell line); TRPV, TRP vanilloid.
↵1 These authors contributed equally to this work.
↵2 Current affiliation: Department of Optometry, State University of New York, New York, New York.
- Received August 4, 2006.
- Accepted September 15, 2006.
- The American Society for Pharmacology and Experimental Therapeutics