動情素在大鼠心肌細胞間隙接合功能調控之角色探討

Abstract

動情素能減少缺血再灌流時心肌組織心律不整的發生，為測試動情素影響間隙接合以維持心肌細胞的同步收縮與代謝平衡的假說，我們在培養的新生大鼠心肌細胞利用氰化物抑制粒線體有氧呼吸，並以碘乙酸鈉抑制無氧醣解作用，以細胞內代謝性抑制的模式模擬系統缺血的情形，探討動情素對心肌細胞間隙接合的影響。以微量注射Lucifer yellow的方式探討間隙接合之功能，發現代謝性抑制可以顯著地降低心肌細胞間染劑偶合為控制組的8.5% ± 0.6%，若預先以17β雌二醇處理，則心肌細胞以劑量依賴的方式增加細胞間染劑偶合至控制組的76% ± 15% (EC50 = 0.41 µM)。為了探討動情素受體是否參與動情素對細胞間隙接合功能之影響，我們發現以無效空間異構物17α雌二醇並無法取代17β雌二醇對代謝性抑制所引起的細胞間染劑偶合的效應。此外，以動情素受體的拮抗物tamoxifen合併處理，則可以抵消17β雌二醇的效應。我們也發現17β雌二醇-牛血清蛋白-FITC可以、但牛血清蛋白-FITC則無法標示心肌細胞表面，顯示心肌細胞表面可能具有膜動情素受體。雙重免疫螢光染色顯示代謝性抑制會造成細胞間隙接合上堆積大量的去磷酸化Cx43，而預先以17β雌二醇處理則可以避免代謝性抑制引起細胞間隙接合上堆積去磷酸化的Cx43。以Lucifer yellow微量注射後的心肌細胞進行雙重免疫螢光染色的結果顯示，17β雌二醇會改善心肌細胞因代謝性抑制造成染劑偶合降低的情形，同時也使細胞間隙接合上去磷酸化的Cx43減少，而絲氨酸368磷酸化的Cx43 (Ser368Cx43)則增加。代謝性抑制會造成心肌細胞的Ser368Cx43下降，若以17β雌二醇預先處理代謝性抑制的心肌細胞則可以顯著地增加Ser368Cx43的量。若以蛋白質激酶C的抑制劑chelerythrine處理，則可以阻止前述17β雌二醇造成代謝性抑制下心肌細胞Ser368Cx43的增加，暗示17β雌二醇可能使蛋白質激酶C活化。由於內皮細胞脂筏蛋白Cav1在膜動情素受體所引發之快速訊息作用中扮演重要的角色，我們進一步收集心肌細胞均質液經高速離心與電泳分離後，以Cx43、ERα與肌肉型脂筏蛋白Cav3的抗體進行免疫轉漬反應。結果發現在清潔劑(Tx)不可溶分離液內的Cx43以Cx43-P2與Cx43-P1為主，在Tx可溶分離液則以Cx43-P1與Cx43-P0為主。而ERα與Cav3在Tx不可溶或Tx可溶分離液內均有反應。以蔗糖梯密度離心法進行上述蛋白質之分布分析，結果發現Cav3主要分布在5%至35%非連續蔗糖梯密度溶液之脂筏界面所在的第三、四、五分離管，Cx43的分布則涵蓋第三至第七分離管，而ERα則平均分布在所有分離管，這顯示部份的ERαCav3共同出現於脂筏界面分離管中。不具膜透性之17β雌二醇-牛血清蛋白、ERα及Cav3抗體的免疫螢光染色，進一步證實心肌細胞具有特定之膜動情素受體。代謝性抑制造成細胞內p-Src活性增加、使Cx43的酪氨酸265 (Cx43Y265)與Cav3酪氨酸磷酸化並使Cav3從脂筏中脫落。免疫沉澱分析顯示，代謝性抑制會減弱ERα與Cav3彼此間的連結。雙重免疫螢光染色的共軛焦顯微鏡觀察顯示，有部分的ERα與Cav3同位分布於細胞膜的表面；以Mβcd處理會瓦解ERα與Cav3在細胞膜表面的同位分布，顯示二者相連並共存於脂筏。除了有部份的Cx43與PKCε在心肌細胞的細胞相接處同位分布外，也有少部分的Cx43和Cav3在細胞相接處同位分布，而Mβcd處理卻不影響Cx43與Cav3在細胞相接處的同位分布。免疫沉澱分析則進一步顯示17β雌二醇會防止代謝性抑制所引起的p-Src與Cx43之間的連結，並促使p-PKCε與Cx43的連結。綜合以上的結果，我們推測心肌細胞ERα與Cav3相連共存於脂筏，代謝性抑制會使Cx43去磷酸化、Cx43Y265磷酸化及p-PKCε與Cx43的連結減少，並抑制細胞間隙接合的功能。17β雌二醇降低代謝性抑制所引起細胞間隙接合功能下降之可能的作用機制是在代謝性抑制之下Cav3會從脂筏中脫落，Cav3一旦分離之後ERα就不再受Cav3抑制，此時17β雌二醇能快速活化ERα，並活化鄰近的蛋白質激酶Cε，進而使Ser368Cx43增加以維持Cx43的磷酸化狀態，另一方面活化的ERα也可能阻礙代謝性抑制所引起的p-Src與Cx43的連結，進而降低代謝性抑制所造成細胞間隙接合功能的抑制效果。

This study was set to test the hypothesis that 17β-estradiol (E2) regulates the coordination of cell-to-cell events in cardiomyocytes by affecting gap junction intercellular communication (GJIC). The effects of E2 on GJIC were assessed by Lucifer yellow dye coupling in cultured neonatal rat cardiomyocytes after metabolic inhibition (MI) using potassium cyanide and sodium iodoacetate. MI significantly reduced dye coupling of cardiomyocytes to 8.5% ± 0.6% of control levels, and pretreatment with E2, but not its inactive isomer 17β-estradiol, dose-dependently (EC50 = 0.41 µM) increased the dye coupling up to 76% ± 15% of control levels. The effect of E2 on MI-induced dye uncoupling was abolished by tamoxifen, a potent estrogen receptor (ER) antagonist. The ligand, E2-BSA-FITC, labeled the cardiomyocyte surface, whereas BSA-FITC did not, suggesting the presence of membrane-associated E2 receptors. Double immunofluorescence microscopy showed that MI-induced the accumulation of non-phosphorylated Cx43 at the gap junction and that this was prevented by E2 pretreatment. Labeling of Lucifer yellow-microinjected cardiomyocytes with antibodies specific for serine368-phosphorylated Cx43 (Ser368Cx43) or non-phosphorylated Cx43 (Cx43-NP) confirmed that E2 reduced the MI-induced inhibition of dye coupling and accumulation of Cx43-NP concomitant with the reappearance of Ser368Cx43 at the gap junction. MI caused a decrease in Ser368Cx43 protein levels, and pretreatment with E2 significantly increased the levels of Ser368Cx43. Inhibition of protein kinase C (PKC) with chelerythrine blocked the E2-induced increase of Ser368Cx43 levels in MI-treated cardiomyocytes. There is evidence showing that caveolin-1 potentiates nongenomic effects of ERα in endothelial cell. Triton-X-100 (Tx) extraction and immunoblot analyses were used to determine the subcellular distribution of Cx43 isoforms, Cav3 (muscle specific caveolae marker), and ERα in cardiomyocytes. While Cx43-P2 and -P1 were detected in Tx-insoluble fraction and Cx43-P1 and -P0 were detected in Tx-soluble fraction, ERαand Cav3 were detected in both Tx-insoluble and -soluble fractions. Cardiomyocyte lysates were further fractionated in a 5% to 40% discontinuous sucrose gradient by centrifugation. Cav3 was detected at the interface of lipid raft in fractions 3 to 5 of the discontinuous sucrose gradient and Cx43 was detected in fractions 3 to 7. In contrast, ERα was recovered from all fractions of the discontinuous sucrose gradient, indicating a wide distribution of ERα in cardiomyocytes. Labeling of cardiomyocytes with membrane-impermeable E2-BSA ligands combined immunolabeling of Cav3 confirmed the presence of membrane binding sites of E2 in cardiomyocytes. Confocal microscopy further demonstrated a partial co-localization of ERα and Cav3 on the dorsal surface of cardiomyocytes and this co-distribution of ERα and Cav3 was abolished by treatment of 5 mM MβCD. MI induced an increase in the levels of p-Src and tyrosine phosphorylation of Cx43. Immunoprecipitation using anti-p-Src and anti-tyrosine265-phosphorylated Cx43 (Cx43Y265) antibodies showed that MI increased the association between p-Src and Cx43Y265. MI also increased the association between p-Src and Cav3 and tyrosine phosphorylation of Cav3, which resulted in the dissociation of Cav3 from caveolae in cardiomyocytes. Double immunofluorescence microscopy also showed that Cx43 was colocalized with PKCε at gap junction. Immunoprecipitation further demonstrated that MI decreased the association between Cx43 and p-PKCε and the MI-induced dissociation of Cx43 and p-PKCε was prevented by treatment of E2 before MI. In conclusion, E2 may elicit an ERα-mediated signaling pathway in MI-treated cardiomyocytes by changing the distribution of tyrosine-phosphorylated Cav3 and its association with ERα via a caveolae-associated signaling mechanism. The activated ERα in cardiomyocytes may activate PKCε which, in turn, attenuates the inhibitory effect of MI on GJIC by affecting the phosphorylation status of Cx43.