|Home | About | Journals | Submit | Contact Us | Français|
Endothelium-mediated vasodilation is specifically enhanced in uterine circulation during pregnancy, and production of nitric oxide (NO) is increased in response to a wide array of agonists. Uterine artery endothelial cells from nonpregnant (NP-UAECs) or pregnant (P-UAECs) ewes maintained in culture still show a pregnancy-enhanced difference in ATP-stimulated endothelial NO synthase (eNOS; official symbol NOS3) activation, even though NOS3 protein, purinergic receptors, and associated cell signaling proteins are expressed at equal levels. We have also shown that the pregnancy-enhanced endothelial cell NO response to ATP requires an enhanced and sustained capacitative entry phase that is likely mediated via canonical transient receptor potential protein/inositol 1,4,5-trisphosphate receptor type 2 interaction. In this study, we now show by simultaneous video imaging of individual Fura-2-loaded cells that the pregnancy-enhanced capacitative entry phase is not continuous and equal in all cells, but is in fact mediated as a series of periodic [Ca2+]i bursts within individual cells. Not only does pregnancy increase the number of bursts over a longer time period in individual cells, but also a greater proportion of cells exhibit this burst activity, and at high cell density this occurs in a synchronous manner. The mediator of cell synchronization is connexin 43 (Cx43) gap junctions because 1) Cx43 is readily detectable by Western blot analysis in UAECs, whereas Cx40 and Cx37 are weakly detected or absent, and 2) pregnancy-specific enhancement of [Ca2+]i bursts by ATP is blocked by inhibitory loop peptides selective to Cx43 ((43,37)GAP27) but not by a scrambled control peptide or (40)GAP27 or (40,37)GAP26 peptides, which are specific to Cx40 or Cx37. The relationship between Ca2+ bursts and NOS3 activation is further established by the finding that (43,37)GAP27 inhibits ATP-stimulated NOS3 activation but has no effect on cell mitogenesis. We conclude that it is pregnancy-enhanced gap junction communication between cells that underlies pregnancy enhancement of capacitative entry via TRPC3 and, in turn, NOS3 activation. Such improved gap junction function allows greater and more sustained [Ca2+]i responses to agents such as ATP within a single cell, as well as the additional recruitment of greater numbers of cells to the response in a coordinated and synchronous manner to support enhanced NO production.
Normal cardiovascular function and the age-related decline therein are major concerns in the general populace. One notable observation is that although all of us suffer from this decline to some degree with advancing age, women in general show a sex-specific delay in development of these symptoms, which is thought to relate to the protective effects of estrogen . Pregnancy is a high-estrogen state in which profound changes in vascular function also occur, with dramatic increases in blood flow to the uterus to meet the continually increasing needs of the growing fetus. Even after decades of study, the exact mechanisms underlying the rise in uteroplacental perfusion and accompanying decrease in vascular reactivity remain unknown. Although increased angiogenesis may, in part, underlie the response, pregnancy is also a time of enhanced vasodilator production by uterine artery (UA) endothelium in response to a number of different agonists [2–4]. Several investigations have confirmed that UA endothelium is more capable of agonist-dependent nitric oxide (NO) and prostacyclin production during pregnancy [2, 5, 6]. Our own studies have further revealed that this adaptive response of UA endothelium includes profound reprogramming at the level of postreceptor cell signaling, specifically at the level of both kinase activation and Ca2+ signaling, particularly during the sustained phase of Ca2+ signaling [3, 4, 7].
More recently, we have also reported that pregnancy-specific alterations in Ca2+ signaling in response to agonists such as ATP (mediated via heptahelical receptors/heterotrimeric G proteins coupled to phospholipase C-β [PLC-β]) include a sustained-phase response that is a form of capacitative entry  and involve agonist-specific and pregnancy-enhanced association of canonical transient receptor potential protein-3 (TRPC3) with inositol 1,4,5-trisphosphate receptor type 2 (ITPR2; previous symbol IP3R2) . The physiologic importance of Ca2+ elevation through IP3 production in these cells as a player in NOS3 activation is clearly illustrated by the inhibitory effect of the PLC inhibitor U73122 or the IP3R antagonist, 2-aminoethyl diphenyl borate (2-APB) on NOS3 activation . Nonetheless, pregnancy enhancement of capacitative entry and NOS3 activation occur in spite of the fact that the P2Y2 receptor, its associated G proteins, PLCB, as well as TRPC3 and ITPR2 themselves are expressed equally in nonpregnant or pregnant uterine artery endothelial cell (UAEC) preparations [8, 9]. We also reported recently that when individual pregnant UAECs (P-UAECs) are observed for prolonged periods (30 min), the greater sustained phase of capacitative entry is observed as a series of periodic [Ca2+]i bursts ; no comparative data from individual nonpregnant UAECs (NP-UAECs) have been published. The questions that remain, and the focus of the current study, are: What is the comparative nature of the [Ca2+]i response in individual NP-UAECs to that observed in P-UAECs, and how can pregnancy enhance sustained-phase Ca2+ signaling without altered expression of these key signaling molecules that otherwise collectively mediate cyclic capacitative entry? Furthermore, how does this change in sustained-phase [Ca2+]i “bursts” relate to pregnancy-enhanced NOS3 activation?
Fura-2 am was obtained from Molecular Probes (Eugene, OR), CaCl2 from Calbiochem (San Diego, CA), and ATP (disodium salt) and all other chemicals, unless noted otherwise, were from Sigma (St. Louis, MO). Also unless noted otherwise, minimal essential medium (MEM) d-Val and all other cell culture reagents were purchased from Invitrogen (Carlsbad, CA). For [Ca2+]i imaging studies, 35-mm dishes with glass coverslip windows were purchased from MatTek Corp. (Ashland, MA). Vascular endothelial growth factor (VEGF-165) was from R&D Systems Inc. (Minneapolis, MN). The (43,37)GAP27 peptide (sequence SRPTEKTIFII), (43,37 scramble)GAP27 (sequence REKIITFIPT), (40)GAP27 (sequence SRPTEKNVFIV), and (40,37)GAP26 (sequence VCYDQAFPISHIR) were custom synthesized by Biopeptide Co. Inc. (San Diego, CA) or were purchased from AnaSpec (San Jose, CA). The purity of these peptides was >95% by HPLC.
Uterine arteries were obtained from Polypay and mixed Western breed nonpregnant sheep (n = 4) and pregnant ewes (n = 6) at 120–130 days of gestation during nonsurvival surgery, as described previously [3, 4]. Procedures for animal handling and protocols for experimental procedures were approved by the University of Wisconsin-Madison Research Animal Care Committees of both the School of Medicine and Public Health and the College of Agriculture and Life Sciences and follow the recommended American Veterinary Medicine Association guidelines for humane treatment and euthanasia of laboratory farm animals.
Uterine artery endothelial cells plated to 35-mm dishes with glass coverslip windows (MatTek Corp.) were grown to the required density (20%–100%, as described) and then loaded  with 5 μM Fura-2 am in Krebs buffer (125 mM NaCl, 5 mM KCl, 1 mM MgSO4, 1 mM KH2PO4, 6 mM glucose, 2 mM CaCl2, and 25 mM HEPES, pH 7.4) for 45 min. After rinsing and incubating a further 30 min to complete ester hydrolysis, Fura-2 loading was verified by viewing at 380-nm ultraviolet excitation on a Nikon inverted microscope (TS100; Melville, NY). About 60 cells were then preselected and video recordings commenced using alternate excitation at 340 and 380 nm at 1-sec intervals and measuring emitted light using a digital camera. From the ratio of emission at 510 nm detected at the two excitation wavelengths, and by comparison to a standard curve established for the same settings using buffers of known free Ca2+ concentration, the [Ca2+]i was then calculated for individual cells in real time using InCyt Im2 software (Intracellular Imaging Inc., Cincinnati, OH).
To prepare protein lysate, cells were grown on T75 flasks to 70%–80% density. Subsequent lysis of cells, protein determination, and Western blotting were as described previously . Blots were probed with connexin 43 (Cx43) antibody (no. sc-9059; 1:1000; Santa Cruz Biotechnology Inc., Santa Cruz, CA; also no. c6219; 1:5000; Sigma) or heat shock protein 90 (Hsp90) antibody (1:5000; ABR-Affinity BioReagents, Golden, CO). After washing, membranes were subsequently incubated with goat anti-rabbit horseradish peroxidase-conjugated secondary antibody (for Cx43, 1:3000; for Hsp90, 1:5000; Cell Signaling Technology, Danvers, MA), followed by detection of signal using enhanced chemiluminescence (ECL) reagent system and HyperFilm ECL (GE Healthcare Bio-Sciences Corp., Piscataway, NJ); densitometry results were quantified using Molecular Analyst 1.4 (Bio-Rad Laboratories Inc., Hercules, CA). Optimum concentrations of primary and secondary antibodies were predetermined using the Protean II Multiscreen apparatus (Bio-Rad Laboratories).
Cells grown to confluence on 35-mm dishes with glass coverslip windows (MatTek Corp.) were scraped at several sites with a razor blade and incubated for 10 min in the dark at room temperature with 0.05% Lucifer Yellow and tetramethylrhodamine (TMR) dextran (Molecular Probes) dissolved in Krebs buffer (above) which was Ca2+ free and containing an additional 50 μM ethylene glycol tetraacetic acid (EGTA) . Cells were subsequently rinsed four times with Ca2+-free Krebs/50 μM EGTA buffer and viewed by fluorescent microscopy to determine the extent of yellow dye entry beyond cells with red dye loading in cells adjacent to the “wound site” vs. general loading of cells with yellow but not red dye throughout the monolayer.
Uterine artery endothelial cells were grown on a T75 flask to 100% confluence and then passaged to two 12-well plates. Cells were then maintained 24–48 h (until they reached >80% confluence). Prior to stimulation, cells were washed twice and then incubated for 1 h in Krebs buffer, which was then removed and replaced with fresh buffer or with 300 μM (43,37)GAP27 in Krebs buffer. Serum starvation continued for three more hours prior to addition of 1.5 μCi per well of l-[2,3,4,5-3H]arginine monohydrochloride (61 Ci/mmol; GE Healthcare Bio-Sciences), followed by addition of vehicle or ATP (30 min, final concentration, 100 μM) to the appropriate wells; the assay was then continued as described previously [10, 11].
Uterine artery endothelial cells were grown on a T75 flask to 70% confluence and then passaged to two 12-well plates. The cells were allowed to attach overnight before beginning the 4-h serum starvation in 900 μl of MEM containing 0.01% bovine serum albumin (BSA), 1.0 units/ml penicillin, 0.5 μg/ml streptomycin, and 4 μg/ml gentamicin. After 1 h of serum starvation, the MEM/0.01% BSA media were removed and replaced with fresh MEM/0.01% BSA or 300 μM (43,37)GAP27 in MEM/0.01% BSA, and incubation was continued for three more hours. After pretreatment, each well was further treated by addition of 100 μM ATP or 10 ng/ml VEGF (control) for 12 h. As described previously , 1 μCi of [methyl-3H]thymidine (GE Healthcare Bio-Sciences) was then added to each well and allowed to incorporate for a further 4 h. The reaction was stopped, cells were washed, and the remaining acid-insoluble material was solubilized and radioactivity determined as for NOS3 activity assay.
For each experiment, fluorescent signal was averaged from at least 20 responding cells. Imaging data are typically from 6 to 10 separate dishes, or for other assays at least four independent experiments, and are presented as means ± SEM. Data were analyzed by Student t-test, paired t-test, or ANOVA, as appropriate. A value of P < 0.05 was considered statistically significant.
We have previously shown in single UAECs from pregnant ewes (by photometry one cell at a time) that ATP can stimulate a more prolonged [Ca2+]i response than in cells from NP ewes, and that this can also include a period of superimposed rapid oscillations within the first 5 min . Imaging of multiple cells either freshly isolated  or on the luminal surface of intact vessels  confirmed that the more sustained nature of the [Ca2+]i response occurs in vivo and specifically defined its direct association with more sustained NO production during pregnancy. More recently, we have shown that when groups of NP-UAECs and P-UAECs are imaged en masse, the average response of cells from pregnant ewes shows enhanced capacitative entry during a 30-min period that is associated with greater TRPC3 channel interaction with ITPR2 . Here, we studied individual cells during a 30-min period and found that this enhanced phase of capacitative entry was sustained not as a constantly elevated plateau, but as a series of periodic [Ca2+]i bursts (Fig. 1). Furthermore, the effect was clearly both cell density and pregnancy dependent. Figure 1 shows representative images of cells at a density of 20%–100% and recordings of 100 μM ATP-induced [Ca2+]i responses in NP-UAECs and P-UAECs at these same cell densities. Of note, the initial response in the first few minutes (less than 5 min) agreed with that described previously —namely, that the initial rapid [Ca2+]i release peak was of similar size in NP-UAECs and P-UAECs—but repeated [Ca2+]i bursts occurred for a longer time in P-UAECs than in NP-UAECs. Some additional superimposed rapid oscillations (see inset) were also seen in P-UAECs but not NP-UAECs. At 100% cell density (Fig. 1A), these rapid [Ca2+]i oscillations usually occurred superimposed upon the first two bursts but sometimes also occurred even at the seventh burst. At 70% density (Fig. 1B), both the pregnancy-associated rapid oscillations were greatly reduced and the sustained periodic [Ca2+]i bursts were reduced in number, although they were still more commonly seen in P-UAECs than NP-UAECs. At 20% cell density (at which point cells rarely make contact) the initial [Ca2+]i responses remained, but periodic bursts were no longer observed in either NP-UAECs or P-UAECs.
Although visually there is a difference in the [Ca2+]i burst response relative to both cell density and pregnancy, there is still the question of how to quantify such a difference. Because burst activity is dramatic and periodic, and therefore readily identified, we counted the number of major [Ca2+]i peaks (initial peak or subsequent bursts), as indicated in Figure 1A. Criteria to define a burst were that it had a clear minima since the last burst and a clear maxima, that the change in fluorescence was appropriately positive or negative in the emission values obtained at 340- or 380-nm excitation, and that the peak maximum achieved at least twice basal [Ca2+]i. Using this approach, Figure 2 summarizes the overall likelihood of seeing each burst of the repeated burst pattern in NP-UAECs and P-UAECs. At the highest cell density (100%), ATP induced multiple [Ca2+]i bursts in the 30 min tested, and the numbers of [Ca2+]i bursts seen in P-UAECs were consistently and significantly greater than those seen in NP-UAECs. As the cell density decreased, both NP-UAECs and P-UAECs showed fewer bursts, and the difference between NP-UAEC and P-UAEC burst patterns diminished. At 70% density, the pregnancy-enhanced difference was only significantly different for the fourth burst. At the lowest cell density (20%), ATP induced the initial [Ca2+]i response and sometimes a second burst, but otherwise all apparent differences between NP-UAEC and P-UAEC behavior observed at higher cell density were lost.
Because digital video imaging was used on fields of cells, analyzing individual cell responses simultaneously was possible, and changes in [Ca2+]i could be compared in neighboring cells. This analysis revealed that rapid [Ca2+]i oscillations and subsequent bursts were synchronous among local subgroups of cells (Fig. 3). This was most obvious at 100% cell density and occurred even in cells not directly adjacent to one another. This synchronization of [Ca2+]i responses in local groups of cells was more pronounced (P < 0.005) in P-UAECs (mean, 57.4% ± 5.2%; n = 7 dishes) than in NP-UAECs (mean, 25.3% ± 5.5%; n = 5 dishes) and in both cases was not readily observed at cell densities below 70%. With respect to the exact degree of synchronization, our measurements were made with a time interval of 1 sec. In the vast majority of cases, the responses were indeed synchronous rather than sequential, and there was no consistent evidence for a lead or “pacemaker” cell responding ahead of the rest in that subgroup.
Because the number of cells responding to agonist was also cell density dependent, and the synchronization of responses between cells observed above is typical of that described elsewhere in cells expressing gap junctions [15–17], we examined the possible expression in UAECs of the most commonly expressed gap junction proteins observed in other endothelial cells—namely, Cx37, Cx40, and Cx43 . We found that in both NP-UAECs and P-UAECs, Cx43 was readily detectable by Western blotting (Fig. 4A); similar results were found using separate antibodies from Santa Cruz Biotechnology and Sigma. In contrast, Cx37 was not detected, and Cx40 was barely detected (data not shown). Of note, although consistency of loading was verified by detection of total Hsp90 protein, the expression of Cx43, although somewhat variable among cells from different animals, was equal overall in NP-UAECs vs. P-UAECs and was detected at the expected native as well as higher molecular weights, consistent with possible posttranslational modification, as reported in other cells [16, 18]. Further confirmation that UAECs form “true” gap junctions between cells at high density rather than operate via hemichannels  was also established using Lucifer Yellow and TMR dextran dye transfer in scrape-loaded cells. TMR dextran has too high a molecular weight to pass through gap junctions, and so it only enters cells damaged by the scrape loading. In contrast, Lucifer Yellow enters the damaged cells and then is sufficiently small to spread further to neighboring cells via intercellular gap junctions. Alternatively, if cells express uncoupled hemichannels on the cell surface, Lucifer Yellow dye can enter all cells equally, regardless of the “wound” site . Our results (Fig. 4B) revealed that after a 10-min incubation in a Ca2+-free Krebs buffer containing 50 μM EGTA, cells adjacent to the scrape-loading site clearly showed efficient Lucifer Yellow transfer several cell layers deep when they were confluent. In contrast, confluent areas well away from the cut site, and so free of any cell damage (i.e., no red dye entry), showed no comparable Lucifer Yellow dye entry above the background, even when the gain of the imaging camera was increased.
In light of our Western blot analysis data (Fig. 4A) and in order to investigate the effect of gap junctional communication on the periodic [Ca2+]i bursts, rapid oscillations, and synchronization observed above, (43,37)GAP27, a peptide homologous to the Gap 27 domain of the second extracellular loop of Cx43 , was used to specifically interrupt gap junction communication. Pretreatment with (43,37)GAP27 for 3 h had little effect on the maximal height (Fig. 5A) or incidence (Fig. 5B) of the initial peak but had a profound effect in preventing long-term [Ca2+]i bursts and rapid oscillation responses to ATP. The (43,37)GAP27 also inhibited the [Ca2+]i burst pattern of both NP-UAECs and P-UAECs to a common low level (Fig. 5B). Moreover, (43,37)GAP27 prevented cell synchronization (data not shown). It should be noted that although (43,37)GAP27 (sequence SRPTEKTIFII) is widely used to inhibit Cx43 function, this sequence also corresponds to loop sequences in Cx37. To establish the specificity of the inhibitory effects of the (43,37)GAP27 peptide, we also performed parallel inhibitory studies in P-UAECs of (40)GAP27, specific to Cx40 , and (40,37)GAP26, specific to Cx37 (not detectable in UAECs) as well as Cx40 (faintly detectable in UAECs) . A control scrambled peptide, (43,37 scramble)GAP27, was also evaluated. Although (43,37)GAP27 had a small (13%) effect on lowering the height of the first peak, this was not significant. However, (43,37)GAP27 was inhibitory to [Ca2+]i burst activity, reducing the number of bursts by 60% (P < 0.0001; Fig. 6). In contrast, none of the other peptides had any inhibitory effect on burst activity. Although two appeared to increase burst activity by approximately 10%, this was not significant in either case.
Previous studies have inferred that sustained capacitative Ca2+ entry in particular is important to NOS3 activation . In previous studies of UAECs, we have used 2-APB and U73122 to inhibit any elevation of [Ca2+]i in response to ATP and have shown a corresponding 60% drop in NOS3 activation . Nonetheless, there is a problem in that the UAEC initial peak and sustained phase both depend on IP3 receptor function, and such agents act to either prevent IP3 production (U73122) or block IP3 action (2-APB), and so inhibit both the initial [Ca2+]i peak as well as the sustained capacitative entry phase. Our data on the effects of (43,37)GAP27 selectively blocking the capacitative entry phase of [Ca2+]i bursts alone provided us with a chance to test the importance of capacitative entry alone more selectively. Our findings (Fig. 7A) showed that (43,37)GAP27 dramatically impaired NOS3 activation in response to ATP in both NP-UAECs and P-UAECs, but in contrast, (43,37)GAP27 did not alter the otherwise stimulatory effect of ATP on thymidine incorporation (Fig. 7B). Of note, the magnitude of the effect of ATP on thymidine incorporation in UAECs was similar to that of 10 ng/ml VEGF_165 (run as an internal control; data not shown). Thus, the action of (43,37)GAP27 was selective to a subset of phenotypic responses of UAECs known to depend on capacitative entry, rather than being due to a general alteration in cell function or overt toxicity.
Pregnancy is a time of enhanced endothelial vasodilation in the uterine vasculature, and our early studies revealed that the freshly isolated endothelium shows increased expression of hormone receptors, NOS3 protein and, indeed, cytosolic phospholipase A2 . Nonetheless, a number of studies in the UAEC primary culture model have shown that even when these protein levels normalize after three to four passages, enhanced NOS3 activation is still retained. Subsequent studies have revealed that differences in cell function are retained, due in large part to remapping of both Ca2+-mediated and kinase-mediated intracellular signaling pathways . Pregnancy-enhanced [Ca2+]i responses occur in UAECs to a number of agonists working through heptahelical receptors coupled to heterotrimeric G proteins and PLCB, including ATP acting through the P2Y2 receptor [4, 8].
We have used ATP as a stimulus in many of our previous UAEC studies to date because it can be used for repeated stimulation without inducing the homologous desensitization otherwise seen in response to agonists such as bradykinin . The maximally effective dose of ATP in UAECs is 100 μM , a concentration that is above basal levels in plasma  but well below the maximum that may occur locally given platelet degranulation . Recent attempts to explain the mechanism of [Ca2+]i responses in UAECs to agonists such as ATP have revealed that pregnancy enhancement of sustained capacitative entry occurs in a manner associated with pregnancy-enhanced and agonist-specific TRPC3/ITPR2 interaction [8, 9]. Because of the involvement of ITPR in both the initial release of Ca2+ from the endoplasmic reticulum and the association and presumed activation with TRPC3, it is then no surprise that we previously reported that U73122, a PLC inhibitor, or 2-APB, an ITPR inhibitor, fully blocks elevation of [Ca2+]i in UAECs. Using this approach, we have calculated that 60% of the overall activation of NOS3 in response to ATP in P-UAECs is due to Ca2+ mobilization . Previous studies of this enhanced [Ca2+]i response to ATP have identified P2Y2 receptors as the possible mediators coupled through heterotrimeric G proteins and PLCB (PLC-β3) [4, 8]. Nevertheless, even though TRPC3/ITPR2 coupling and associated capacitative entry both are agonist specific, completely IP3 dependent, and pregnancy enhanced, it is intriguing that neither TRPC3 or ITPR2, nor indeed any of the other proteins mediating the chain of events from receptor coupling to capacitative entry, is expressed at any different level in P-UAECs than in NP-UAECs. So how does altered function occur in pregnancy? Several mechanisms could be proposed that are based on altered phosphorylation or relative proximity of TRPC3 and/or ITPR2. Our findings here, however, suggest that another mechanism altogether underlies pregnancy-enhanced capacitative entry; namely, pregnancy enhancement of gap junction function.
In a recent study, we monitored individual P-UAECs loaded with Fura-2 using video imaging and so determined the responses in many cells simultaneously. Such studies showed that the sustained phase of capacitative entry in UAECs is, in fact, not a constant elevation of [Ca2+]i but a repeated periodic increase in [Ca2+]i ([Ca2+]i bursts) . In this study, we further report novel findings from extended simultaneous recordings of individual cells from both NP-UAECs and P-UAECs. We found that the repeated periodic [Ca2+]i bursts observed in P-UAECs exceed those observed in NP-UAECs but, surprisingly, the pregnancy-enhanced periodicity of this response was lost as the cell density was reduced. This raises some important questions regarding the molecular basis for pregnancy adaptation of uterine blood flow. The prolonged burst phenomenon we have observed has been reported in other nonendothelial cell types (such as hepatocytes) and may be synchronous when cell-cell communication occurs via gap junctions . Of the most commonly observed gap junction proteins in endothelial cells, Cx43 is most readily observed in UAECs from both nonpregnant and pregnant ewes. Given the presence of Cx43, there are two general patterns of cell-cell communication that can occur. The first is direct cell-cell communication via true gap junctions formed between adjacent cells. The second is paracrine communication by the release of substances such as ATP through a hemichannel on the luminal surface of one cell and stimulation of a receptor on a neighboring cell . Fortunately, these two different mechanisms can be distinguished readily. First, paracrine communication via hemichannels will be relatively less dependent on cell density than true effects mediated by gap junctions, where actual cell contact is a strict requirement. Second, paracrine signaling via hemichannels alone will be revealed by incubation in a solution containing Lucifer Yellow dye, a molecule of low enough molecular weight to pass through Cx hemichannels or through gap junctions. In the case of hemichannels, dye will enter all cells equally, whereas for true gap junctions, as observed here, dye will enter only from neighbors damaged by scrape loading or microinjection. Third, Cx mimetic peptides based on the interacting extracellular loop sequences will break down specific connections between cells , and thereby block true gap junction function but not hemichannel function . Because the studies here show that the [Ca2+]i burst phenomenon is cell density dependent, that Lucifer Yellow transfer is only detectable above basal levels in the cell layers adjacent to scrape-loading sites, and that the [Ca2+]i burst phenomenon is sensitive to the Cx43 loop (inhibitory) peptide (43,37)GAP27, but not the scrambled control peptide, we have definitively shown that UAECs communicate via true gap junctions. The GAP27 peptide that is effective ((43,37)GAP27) is derived from a loop sequence found in Cx43 that is also homologous to that in Cx37. Nonetheless, Cx37 was not clearly detected in UAECs by Western blot analysis, and other control GAP peptides selective for Cx37 or Cx40, but not Cx43, were unable to inhibit the [Ca2+]i burst response. Together, these observations make it clear that the [Ca2+]i burst phenomenon in UAECs is mediated by Cx43, an isoform which is known to form gap junctions selective not only to Ca2+ but also IP3 .
The existence of cell-cell communication via Cx43 is certainly sufficient to explain the cell synchronization phenomenon we observed, along with its disappearance at low cell density. The ability of cells to synchronize in this way has several physiologic ramifications. It immediately raises the possibility that one cell that successfully responds to an agonist by raising [Ca2+]i may then be able to recruit additional coupled neighbors to respond. Thus, our finding that pregnancy is a time of enhanced communication (with greater numbers of cells responding and twice as many cells synchronized in the pregnant as in the nonpregnant state) suggests that pregnancy-enhanced NO production by UAECs possibly may involve not only greater NOS3 activation per cell, but also recruitment of greater numbers of cells. Consistent with this, although (43,37)GAP27 had little effect on the initial Ca2+ peak magnitude, it effectively reduced subsequent [Ca2+]i bursts in NP-UAECs and P-UAECs down to a common low level (and eliminated synchronized responses by both NP-UAECs and P-UAECs), and it inhibited ATP-stimulated NOS3 activation. The degree of inhibition of NOS3 activation in P-UAECs and NP-UAECs was around 50% of that observed in the absence of (43,37)GAP27. This is not quite as great as the 60% inhibition observed with the ITPR antagonist 2-APB , but the latter compound also inhibits the initial intracellular Ca2+ release from the endoplasmic reticulum in addition to subsequent burst activity.
Western blot analysis of lysates from NP-UAECs and P-UAECs isolated from different ewes showed clearly detectable expression of Cx43 at multiple apparent molecular weights (Fig. 4A). Although some degree of interanimal variability exists in the expression of Cx43, as would be expected for studies of individual animals, there is no remarkable difference in average expression between NP-UAECs and P-UAECs, yet function as measured by [Ca2+]i bursts suggests P-UAECs still have greater Cx43-mediated cell-cell communication. The function of Cx proteins in general, and Cx43 in particular, is regulated by a number of posttranslational mechanisms, including cyclic nucleotide-dependent insertion in the plasma membrane, which can be counterbalanced by kinase-dependent phosphorylation and closure . The presence of multiple-molecular weight bands in our Western blots of UAEC lysates is clearly consistent with previously described posttranslational modification (possibly phosphorylation events). Indeed, the multiple Cx43 bands were detected at molecular masses similar in range to the previously named “P0, P1, and P2” isoforms of Cx43 (corresponding to 42- to 46-kDa apparent molecular mass), which are formed by phosphorylation-associated conformation changes ). Further studies will be necessary to determine both the nature and the extent to which Cx43 phosphorylation occurs in UAECs, how phosphorylation is mediated, and how this is possibly altered by pregnancy. Cyclic nucleotide-enhanced insertion of Cx43 into the plasma membrane also needs to be considered because during pregnancy, the uterus is known to release such large amounts of cAMP and cGMP that both are detectable in the ovine circulation [27, 28] and human myometrium . It is at least possible that some of the drop in systemic resistance observed during pregnancy is secondary to generalized cyclic nucleotide upregulation of gap junction insertion, and thus function, in the systemic endothelium itself. Of additional interest, the enhancement of hand vein endothelial cell [Ca2+]i responses seen in pregnancy is not observed in patients who develop preeclampsia . It remains to be seen whether these patients also fail to show enhanced cell-cell communication due to a lack of cyclic nucleotide-mediated insertion, or perhaps alternatively due to increased gap junction closure by the action of specific kinases.
In summary, our former studies in UAECs suggested that pregnancy-enhanced [Ca2+]i responses at the level of capacitative entry are due to pregnancy-specific enhancement of the association of ITPR2 with TRPC3 in response to ATP . The association of ITPR2 with TRPC3 is most likely dependent on IP3 because 2-APB inhibits the sustained [Ca2+]i response as well as the initial release of Ca2+ from intracellular stores . In the studies reported here, we have further shown that sustained capacitative entry in UAECs is due to periodic [Ca2+]i bursts in individual cells, and this burst response is enhanced by both pregnancy and cell density. Cx43 is abundantly expressed in UAECs, and the inhibitory GAP peptide (43,37)GAP27 can selectively block the [Ca2+]i burst phenomenon in NP-UAECs and P-UAECs at high cell density down to a lower common level otherwise only observed in cells at 20% confluence (i.e., not making contact with each other). In light of this and the previous reports that Cx43 gap junctions can indeed transport IP3 between cells , we propose that one possible model for the pregnancy-induced enhancement of ITPR2/TRPC3 association is not so much mediated by changes in the expression of TRPC3 or ITPR itself, or even direct posttranslational modification of those proteins, but may be due to the greater pregnancy-enhanced ability to shuttle IP3 between the cells via functional Cx43 gap junctions. In such a model, the higher levels of gap junction function in P-UAECs simply allow the rise in cytosolic concentration of IP3 to be coordinated across many cells at once. A recent publication, however, suggests that cells communicating by gap junctions can also synchronize at an electrical coupling level, even when not considered classically excitable cells . The hallmark of this type of electrical synchronization is that it would produce exact synchronization and also be more capable of acting over longer distances. This would be particularly attractive in allowing downstream resistance vessels to more readily communicate with larger distributing vessels in a retrograde fashion (i.e., against the direction of flow). Although the vast majority of our observations suggest the responses of UAECs are indeed synchronous rather than sequential, we would qualify this by acknowledging that we do not have sufficient temporal resolution in our current studies to clearly distinguish which mechanism applies at this time. Nonetheless, our results here strongly implicate regulation of cell-cell communication via Cx43 as the mechanism by which ITPR2/TRPC3 interaction is enhanced in pregnancy. Of further physiologic significance, pregnancy-enhanced NOS3 activation observed in UAECs is lost when Cx43 gap junction function is inhibited. Thus, regulation of Cx43 function is a specific and indeed crucial pregnancy-programmed event in UAECs that enhances capacitative entry and, in turn, NOS3 activation, but does not have an impact on other cell functions, such as mitogenesis.
Given our findings on the molecular basis of pregnancy-specific adaptation, we must also ask whether the aforementioned failure of vascular endothelium to enhance vasodilator (and particularly NO) production in disease states such as preeclampsia has its origins in a failure to increase cell-cell communication or, alternatively, whether there is a pathologic event that actively shuts down such an adaptation. Studies in hand vein endothelial cells show that pathologic pregnancy is associated with a failure to induce sustained [Ca2+]i responses otherwise seen in normal pregnancy , but they do not answer our question of failed adaptation vs. inhibited adaptation. To that end, it is interesting to note that two factors associated with preeclamptic pregnancy—namely, tumor necrosis factor-α and VEGF—both have a negative impact on Cx43-mediated cell-cell communication in other cell models. Tumor necrosis factor has been shown to promote gap junction closure and subsequent breakdown of Cx43 in corneal fibroblasts , and VEGF has been show to promote phosphorylation and closure of Cx43 in both human umbilical vein endothelial cells (HUVECs) and Ea.hy926 cells . Thus, it is distinctly possible that preeclampsia is an inhibition or even reversal of Cx43 functional adaptation in UAECs during pregnancy, but further studies are necessary to establish whether this is indeed the case. Nonetheless, at least we now have a molecular signaling model on which to base such studies on the failure of vascular adaptation during pregnancy, and more broadly perhaps to explore new ways to reverse the age-related decline in cardiovascular function in both men and women.
We would like to thank Terrance Phernetton for assistance with animal preparations.
1Supported by National Institutes of Health grant nos. HL64601, HD050578, HL079020, HL49210, and HD38843.