Shear stress immediately triggers nonselective cation currents in atrial myocytes
Transient inward current was recorded in rat atrial myocytes at resting potentials during the application of shear stress (IS) following a series of conditioning pulses (Fig. 1A). The IS were dramatically increased and prolonged when external Ca2+ was removed from puffing solutions, while it was slightly reduced at 10 mM external Ca2+ concentrations (Fig. 1B). Next, we examined what types of ions contribute to IS by substitutions of major cations in the external solutions with choline. Results presented in Fig. 1C and 1D show that Na+ and K+ are the major ions to carry IS at -70 mV. The results in Fig. 1E and 1F obtained using high Ca2+- and low Na+-external solutions represent that Ca2+ also can contribute to IS.
Because shear-activated currents were nonselective cation currents we used symmetrical CsCl solutions and zero-Ca2+ puffing solutions to better isolate and maximize the current for further mechanism study. Transient inward Cs+ currents were observed under shear application (IS,Cs), and they were consistently enhanced with zero-Ca2+ solutions (2–4-fold; Fig.1G). Measurements of IS,Cs at varying shear forces (1.4–16 dyn/cm2) revealed that the IS,Cs were activated faster and more strongly as the strength of shear stress increased, reaching maximum at ~10 dyn/cm2 (Fig. 1H, I).
Nonselective cation currents are induced by Cx43 hemichannel activity under shear stress
Our previous report suggested that ATP releases from atrial myocytes under shear stress are mediated by Cx43 hemichannels [14, 15]. Therefore, a possible role of Cx43 hemichannels in the activation of IS,Cs were further examined. Pretreatment of La3+ (2 mM) that non-selectively suppresses gap junction channels almost completely suppressed IS,Cs (Fig. 2A). IS,Cs was eliminated by the application of high concentrations (50 µM) of carbenoxolone (“CBX”) that suppresses gap junction channels, Cx and pannexin (Px) (Fig. 2B). Blockade of Px using probenecid (800 µM) did not alter IS,Cs (Fig. 2C), which also excludes the possibility that Cx-mediated intracellular Ca2+ increase secondarily activates Px . Lack of the effect of external Ca2+ on Px  is also inconsistent with the property of IS,Cs. Low external Ca2+ solutions can enhance the activity of calcium homeostasis modulator 1 (CALHM1). However, CALHM1 is known to insensitive to CBX , which excludes its contribution to IS,Cs. TRPM4 is activated by prolonged shear stress via intracellular Ca2+ releases . The TRPM4 inhibition using 9-phenanthrol (9-PT) decreased IS,Cs by ~25% (Fig. 2D).
The role of Cx43 hemichannel in the generation of IS,Cs was further examined by the inhibition of Cx43 using an anti-Cx43 antibody of Cx43. IS,Cs was almost completely blocked by dialysis with internal solutions containing the Cx43 antibodies (400 ng/ml) for 15 min (Fig. 2E, F; “Cx43 Ab”). When the same experiments were performed in cells dialyzed with scramble as a control, IS,Cs was not significantly reduced up to 15-min-dialysis (Fig. 2E, F). Note that IS,Cs measured in control internal solutions were slightly larger than those detected in cells dialyzed with scramble-containing internal solutions (Fig. 2E, F).
Next, we used Cx43 cKO mice to further confirm the specific role of Cx43 in the activation of IS,Cs. Cx43 expression was reduced to ~20% in Cx34 cKO mice using the present TAM injection protocol (Fig. S1). The peak magnitudes of IS,Cs in atrial myocytes from Cx43 cKO mice were much smaller than those measured in control mouse atrial cells (Fig. 2G). Shear-induced cation currents were similarly eliminated by Cx43-specific siRNA in HL-1 atrial cells (Fig. S2). To confirm the role of Cx43 hemichannel in the activation of IS,Cs, we used TAT-Gap19 that inhibits Cx43 hemichannels without blocking Cx40 or Px1 hemichannel [25, 26]. External application of TAT- Gap19 suppressed IS,Cs in a concentration dependent manner with an almost complete inhibition at 60 µM (Fig. 2H). Taken together, these data indicate a key role of Cx43 hemichannels in the activation of IS,Cs.
Shear stress induces a dye flux via Cx43 hemichannels in atrial myocytes
To address the activation of Cx43 hemichannels by shear stress in atrial myocytes we performed live-cell imaging for Cx-permeable fluorescent dye (calcein-red/orange). In the dye-loaded cells cellular fluorescence was immediately decayed by the application of shear stress with 2 mM Ca2+-containing Tyrode’s solutions (Control in Fig. 3A, B). To compare the effects of several interventions the time courses of averaged calcein fluorescence, measured in multiple cells, were normalized to control levels and presented. In addition, dye efflux rate was estimated as %DF/F0 per min to compare the peak shear effects. Results in Fig. 3B and 3C show that shear-induced dye efflux was accelerated to ~300% of the control by the removal of extracellular Ca2+ (“Ca2+-free”) and it was almost completely suppressed by the application of La3+, but not by Px blockade (“probenecid”).
The next series of experiments were carried out using puffing solutions containing zero extracellular Ca2+ as a control under repetitive applications of 1-min-long shear stress in the same myocytes to further investigate the pathway of dye efflux under shear stress. Calcein signal decay in the absence of shear stress (“no shear”) was much lower than that in the presence of shear stress (Fig. 3D, E). After the myocytes were treated with 10 µM CBX, that blocks Px more selectively than Cx , shear-mediated dye efflux was not significantly altered (Fig. 3D, E). This result confirms that Px does not contribute to shear-mediated calcein efflux in these myocytes.
In atrial myocytes preincubated with Gap19, there was no significant shear-dependent calcein efflux, in that the calcein signals measured without (“no shear, Gap19”) and with shear stimuli (“1st and 2nd shear, Gap19”) were similar (Fig. 3F, G), suggesting a role of Cx43 hemichannels on shear-induced calcein efflux. Using Cx43 cKO mice, we confirmed that shear-stress-induced calcein efflux was negligible in atrial myocytes lacking Cx43 (Fig. 3H, I). In contrast, in atrial cells from control mice, shear stress continued to induce significant dye efflux, which was not altered by probenecid (“2nd shear, prob") (Fig. 3H, I). These results suggest the activation of Cx43 hemichannels by shear stress in atrial myocytes.
Autocrine activation of P2 receptors mediates IS,Cs: a major role of P2X4R
Since Cx43 hemichannels mediate ATP release from atrial myocytes under shear stress , P2 receptors can be activated in an autocrine mode. Therefore, we tested whether P2X receptors that carry cations are activated in this way and contribute to IS,Cs. When ATP was removed by external application of apyrase IS,Cs were completely suppressed (Fig. 4A). When P2 receptors were suppressed by suramin (10-30 mM), IS,Cs were reduced by ~75% (Fig. 4B, G). Application of the P2XR-selective antagonist iso-PPADS at the concentrations of 10-100 mM reduced IS,Cs by 55%-60% (Fig. 4C, G). Inhibition of the P2Y1 receptor (P2Y1R) using MRS2179 (400 nM) reduced the current by 20-30% (Fig. 4D, G). Additional application of 9-PT did not further suppress the current, suggesting that TRPM4 may be a downstream effector molecule under the activation of P2Y1R. These results suggest that autocrine activation of P2XR and P2Y1R may be responsible for most of cation influx at diastole under shear stress, with a major contribution by P2XRs.
Moreover, to identify the P2XR subtype to carry the IS,Cs pharmacological interventions were further used. Since P2X4R and P2X5R are abundant in rat atrial myocytes , their contributions to IS,Cs were examined with dialyzing their antibodies (400 ng/ml) into the myocytes through the patch pipettes. Results in Figure 4E show that introduction of anti-P2X4R antibodies into the cells suppressed IS,Cs by ~50% within 15 min of dialysis. In contrast, dialysis of cells with the P2X4R antibodies together with their antigenic peptides as a control did not significantly alter the currents (Fig. 4E). When the same experiments were carried out using anti-P2X5R antibodies, IS,Cs were not significantly reduced by introduction of the antibodies (Fig. 4G; Fig. S3).
Furthermore, selective inhibition of P2X4R with its antagonist 5-BDBD reduced IS,Cs by 50%-60% (Fig. 4F, G) at its maximally effective concentration (10 mM) [27, 28]. Consistently, PSB-12054, another P2X4R antagonist , inhibited IS,Cs in a concentration-dependent manner with a maximal suppression of ~40% at 3 µM (Fig. S4, Fig. 4G). Altogether, these results suggest that P2X4R is a major subtype responsible for P2XR component (~50%) of IS,Cs.
Atrial hypertrophy and dilation induced by short- and long-term TAC
PO such as hypertension is one of leading causes of atrial hypertrophy that leads to depressed atrial contractility in patients and represents an important risk factor for atrial fibrillation, thromboembolic stroke, and HF. Under these pathologic conditions, atria are subjected to higher shear stress. Next, we investigated whether IS,Cs and Cx43-P2X4R signaling under shear stress are altered as atrium undergoes remodeling with gradual increase in afterload using a TAC rat model. Transthoracic echocardiography results represent that TAC for 5 weeks produced typical LV hypertrophy, in which ejection fraction (EF; +20%, P < 0.01), fractional shortening (FS; +40%, P < 0.01), and wall thickness (+40%, P < 0.001) were increased (Fig. S5). At 5-week-TAC, the LA chamber was slightly, but significantly, dilated compared with the sham (+30%, P < 0.05; Fig. 5A, B). The FS of LA in the 5-week-TAC rats were not altered (Fig. 5B). Measurement of membrane capacitance using whole-cell patch clamp showed ~two-fold increase in atrial cell surface area after TAC for 5 weeks (+100%, P < 0.001; Fig. 5B).
At ~20 weeks after TAC surgery, left ventricle showed HF characteristics, such as lower ventricular EF and wall thickness, and larger LV inner diameter (Fig. S5). After 20-week-TAC, LA dilation became more severe (+70%, P < 0.01 vs. sham; Fig. 5A, B) and contractility of LA was lower than that in the age-matched sham group (-50%, P < 0.05; Fig. 5B). The capacitances of LA cells were not further changed after 20-week-TAC (Fig. 5B).
IS,Cs and its P2X4R component increase during compensatory hypertrophy but decrease with depressed atrial contractility under chronic TAC
Comparison of LA IS,Cs revealed approximately two-fold increase in IS,Cs density during compensatory hypertrophy (upper panels of Fig. 5C; Fig. 5D, “Control”). Inhibition of IS,Cs by 5-BDBD (10 µM) (by ~50%) was not altered in these hypertrophied LA cells (Fig. 5C, D, upper panels), which results larger 5-BDBD-sensitive IS,Cs in these LA cells (sham, 0.81 ± 0.21 pA/pF, n = 12; TAC, 1.4 ± 0.33 pA/pF, n = 11, P < 0.05) (Fig. 5C, D). During HF IS,Cs of LA cells were decreased again (lower panels in Fig. 5C, D), resulting in no difference in the IS,Cs from LA cells from sham rats (Fig. 5C, D). Furthermore, IS,Cs in 20-week-sham LA cells were larger than those from LA cells of 5-week-sham rats. At this stage, ~50% of IS,Cs in LA cells were still sensitive to 5-BDBD in sham group while smaller (~40%) fraction of IS,Cs was suppressed by 5-BDBD in TAC group (Fig. 5C, 5), resulting in a reduction of 5-BDBD-sensitive IS,Cs (P2X4R current) in failed LA cells (sham, 0.97 ± 0.18 pA/pF, n = 15; TAC, 0.71 ± 0.22 pA/pF, n = 10, P < 0.05). Altogether, these results suggest that IS,Cs may be a compensatory mechanism that is enhanced during LA remodeling under PO, and that this shear response is ameliorated with severe LA dilation during HF.
Next, we investigated whether P2X4R protein expression in LA muscle is changed in short- and long-term-TAC rats. Results on Figure 5E and 5F show that the expression of P2X4Rs in RA and LA muscles significantly increased during hypertrophy, consistent with larger P2X4R-sensitive currents at this stage. Note that the P2X4R expression is significantly higher in old (20 week) sham atrial muscle compared to that of younger sham (5 week). In contrast, the levels of P2X4R proteins in LA- and RA-muscles during HF were significantly lower than those from the age-matched control group (Fig. 5E, F, lower panels), consistent with the reversal of IS,Cs and decrease of 5-BDBD-sensitive P2X4R current in failed LA myocytes. We observed a similar tendency in the changes of magnitudes of IS,Cs and 5-BDBD-sensitive currents in RA myocytes by short- and long-term TAC, although the fraction of 5-BDBD-senstivive current was significantly larger in RA myocytes compared to LA myocytes (Fig. S6). Taken together, the results suggest that short-term PO increases P2X4R-mediated cation currents, at least in part, via upregulation of P2X4R protein expression, and that long-term PO attenuates IS,Cs and P2X4R function. In addition, expression of P2X4Rs and their functions may increase with aging, which may cause a reduction in internal P2X4R pool to be immediately regulated by shear stress.
Attenuation of shear-activated Cx43 hemichannel activity in atrial myocytes during HF
The changes of IS,Cs in LA myocytes by PO can be caused by alterations in the activity of Cx43 hemichannels. Next, we examined whether Cx43 hemichannel-mediated calcein efflux under shear stress is altered during compensatory hypertrophy and HF. The results shown in Figure 6A and 6B present calcein signal-averages measured in multiple RA- and LA- myocytes from 5- and 20-week-TAC rats and from their age-matched sham rats. During PO-induced hypertrophy shear-induced calcein efflux was not altered in RA and LA myocytes.
During HF, LA myocytes showed significantly lower shear-induced calcein efflux than the age-matched sham group (“20-wk, shear”, Fig. 6B, D). The RA cells from 20-week-TAC rats also showed a similar tendency, with less reduction in the calcein efflux under shear stress (“20-wk, shear”, Fig. 6A, C). Notice that the shear-induced calcein efflux rate was generally 20~30% higher in RA cells compared with LA myocytes (Fig. 6). There were no age-dependent changes in the magnitude of calcein efflux in LA and RA myocytes from sham groups (compare 5-week and 20-week sham in Fig. 6). The results indicate that shear-induced Cx43 hemichannel activity may be maintained during compensatory phase, and that this response is ameliorated with severe atrial dilation under chronic PO.
Alteration in expression and phosphorylation of Cx43 by short- and long-term TAC
Next, we investigated whether the alterations of shear-induced Cx43 hemichannel function by PO are caused by remodeling in Cx43 protein expression using immunoblotting method. Results in Figure 6E and 6F show that total Cx43 expressions were not changed in RA and LA muscles during hypertrophy, while they were significantly decreased under HF. This result was consistent with the functional changes in Cx43 hemichannels in atrial cells under hypertrophy and HF (Fig. 6 A-D). Note that the levels of ANP expression were gradually increased in both RA- and LA-tissues by PO (Fig. 6E, F; “ANP”). Overall ANP protein expressions were higher in LA tissue than RA tissue.
We further assessed the phosphorylated form of Cx43 at Ser368 (pS368-Cx43), a known protein kinase C (PKC) site [30-32], because PKC is a downstream signaling molecule activated by shear-P2Y1R signaling in atrial myocytes . This phosphorylation is also thought to contribute to the localization and remodeling of Cx43 in cardiac muscle . The levels of pS368-Cx43 were increased only in the LA muscles by PO (Fig. 6E, F) and much lower in older LA muscles (Fig. 6E, F). The results on the alteration of pS368-Cx43 indicates that this phosphorylation is PO-dependently regulated in LA muscle.