Ca2+ Waves in Lung Capillary Endothelium
Although cytosolic Ca2+ importantly regulates organ function, lung microvascular [Ca2+]i regulation remains poorly understood because of the lack of direct in situ quantification. In the present study, we report the first endothelial [Ca2+]i quantification by the fura 2 method in microscopically imaged venular capillaries of the isolated blood-perfused rat lung. Sequential images indicated the presence of intercellular Ca2+ waves that spontaneously originated from pacemaker endothelial cells and then spread for short distances along the capillary wall, inducing synchronous endothelial [Ca2+]i oscillations. Fast Fourier analyses of the oscillations revealed a dominant wave component with an amplitude of 37 nmol/L, frequency of 0.4 min−1, and velocity of 5 μm/s. The intracellular Ca2+ wave was unaffected by blood flow stoppage or by infusions of Ca2+-containing or Ca2+-free dextran. Inhibition of the wave by thapsigargin in Ca2+-free dextran and by the gap junction uncoupler, heptanol, indicated that it was generated by endosomal Ca2+ release in the pacemaker cell and was propagated by gap junctional communication. In the presence of histamine, enhancement of the wave accounted for a significant component of the coordinated [Ca2+]i increase in the capillary segment. No intercellular Ca2+ waves were evident in adjoining alveolar epithelial cells. Our findings indicate a novel mechanism of [Ca2+]i regulation in the lung capillary under both resting and stimulated conditions. Pacemaker-induced Ca2+ waves, generated intracellularly by unknown initiating mechanisms, communicated to adjoining cells to determine [Ca2+]i profiles in short interbranch segments of capillary walls.
The cytosolic [Ca2+] ([Ca2+]i) is a second messenger of wide-ranging importance in the regulation of cell and organ function. Increases of cytosolic Ca2+ not only regulate intracellular processes, but when intercellularly propagated as Ca2+ waves, they may be instrumental in organizing multicellular function.1 2 Intercellularly propagated intracellular Ca2+ waves have been induced by agonists and by mechanical or electrical stimuli in several cell types, including tracheal epithelium, hepatocytes, pancreatic acinar cells, and astrocytes,3 4 5 6 7 8 9 where they are likely to be important for the coordination of ciliary beat frequency, bile flow, enzyme secretion, and neuronal excitation, respectively. Intercellular transmission of cytosolic Ca2+ waves also occurs in mechanically stimulated endothelial cells in culture,1 indicating that endothelial cells use intracellular Ca2+ to intercommunicate mechanosensitive signals.
In lung microvessels, endothelial cells serve a variety of functions, such as maintenance of the vascular permeability barrier, establishment of a nonthrombogenic luminal surface, and secretion of vasodilatory factors such as NO. Under inflammatory conditions, the cells secrete cytokines and process leukocyte migration. Many of these functions are [Ca2+]i dependent, and their signaling mechanisms involve critical increases in [Ca2+]i.2 Endothelial cells of lung microvessels are likely to be syncytially interconnected, as in airway microvessels,10 and may be linked through intercellular channels, as in hamster cheek pouch microvessels.11 Hence, it is possible that intercommunication of Ca2+–regulatory signals among these cells provides the basis of coordinated lung endothelial function.
The understanding of cytosolic Ca2+ intercommunication in lung is hampered by the lack of [Ca2+]i quantification in lung tissues in situ. Hence, our primary aim was to develop the fura 2–ratioing method12 for the first direct [Ca2+]i quantification in intact lung microvessels. For these determinations, we selected the subpleural capillary bed, which we have used previously for microvascular and interstitial imaging.13 14 15 While developing these approaches, we unexpectedly determined the existence of spontaneously generated low-amplitude Ca2+ waves that propagated intercellularly along the capillary wall. These waves, which have not been previously reported, provide a potential mechanism for the coordination of endothelial [Ca2+]i and, possibly, [Ca2+]i-dependent function in lung microvessels.
Materials and Methods
Fluorescent Probes and Drugs
The dyes fura 2, fura 2-AM, Hoechst 33342 (Molecular Probes), diI-AcLDL (Biotechnology Research Institute), Cy3 (Cy3 labeling kit, Sigma Chemical Co), and the agents thapsigargin, digitonin, ionomycin (Calbiochem-Navabiochem), and histamine dihydrochloride (Sigma) were used in solutions of delipidated albumin (BSA, fraction V, Sigma) or dextran (70 kD, Pharmacia) in Ringer's solution containing (mmol/L) Na+ 144, K+ 4, Ca2+ 1.5, and lactate 28, at pH 7.4, and osmolarity of 300 mOsm (The Advanced Micro-Osmometer, model 3M0, Advanced Instruments).
Lung preparation methods have been reported several times16 17 and are described here briefly. Lungs removed from anesthetized rats (450 to 550 g, Harlan Sprague-Dawley Inc; 2% halothane inhalation followed by 35 mg/kg IP sodium pentobarbital) were pump-perfused with autologous rat blood through cannulas inserted in the pulmonary artery and left atrium. A tracheal cannula was used for lung inflation with a gas (30% O2/6% CO2/balance N2) that maintained blood Po2, Pco2, and pH at 140 mm Hg, 35 mm Hg, and 7.4 (178 pH/Blood Gas Analyzer, Corning), respectively. A heat exchanger (44TD, Yellow Springs Instrument Co) maintained perfusate temperature at 37°C. Recordings of lung vascular and airway pressures (pressure transducer, P23 ID, Gould Statham), referred to the micropuncture level, were displayed on a multichannel recorder (RS 3400, Gould). Lung vascular pressures and blood flow were varied by adjusting pump rate and height of the venous outflow. Lung blood flow of 14 mL/min was maintained at pulmonary artery pressure of 10 cm H2O. Based on our previous data,18 venular capillary pressure was assumed to be 1 cm H2O greater than left atrial pressure, which we adjusted as indicated. Airway pressure was held constant at 5 cm H2O during experimental periods but cyclically varied at other times to induce ventilation at regular intervals.
Lungs were positioned on a vibration-free air table (Micro-G, Technical Manufacturing Corp) and viewed by intravital microscopy (Vanox, Olympus Corp). To prevent drying, the experimental lung surface was layered with a drip of silicone oil (Dow Corning 200) warmed to 37°C. Venular capillaries (diameter, 20 to 25 μm), viewed through a custom-built intravital fluorescence microscopy system, were identified by their convergent flow patterns.18 Except where stated, all determinations were made in the presence of capillary blood flow.
Membrane-permeant fura 2-AM (10 μmol/L), which deesterifies intracellularly to impermeant fura 2,19 was infused for 20 minutes into capillaries through a microcatheter (PE-10, Clay-Adams) wedged in the venous system. To prevent dye leakage from the capillary during injections of fura 2-AM, we established absorptive conditions (lumen-directed liquid flux) in the capillaries13 by maintaining a capillary pressure of 1 cm H2O.
For fluorescence excitation, mercury lamp (Mercury-100, Olympus; USH-I02D, Ushio) illumination was passed through a filter wheel and shutter (Lambda-10, Sutter Instrument Co) mounted on a microscope head (BH2-RFCA, Olympus) and equipped with filters appropriate for excitation at 340, 360, and 380 nm (340HT15, 360DF10, and 380HT15, Omega Optical). Excitation time of exposure and excitation wavelength were controlled by our computer-based imaging system (Image-640 RTP, Matrox; MCID-M4, Imaging Research; 466/ME, Dell Computers). The resulting lung capillary fluorescence was collected through UV-compatible objectives (×40 Fluor LWD, Nikon), dichroic and emission filters (400 DCLPO2 and 510WB40, Omega Optical), and a projection lens (×3.3, Nikon) into an image-intensifier (KS 1381, Videoscope) and video camera (CCD-72, Dage) or a cooled CCD camera (C4880, Hamamatsu). Neutral density filters (Omega Optical) placed in front of the lamp housing (LH100, Olympus) and the 380-nm filter reduced excitation intensities.
Each ratiometric determination of [Ca2+]i consisted of four frame-averaged sets of exposures that alternated every 33 milliseconds between the 340- and 380-nm excitations. Thus, the 340/380 fluorescence ratio was generated with a total exposure time of 264 milliseconds. The images were background-corrected, with background determined in images captured before fura 2 loading or after quenching fura 2 fluorescence by a 5-minute infusion of 1 mmol/L Mn2+ in Ringer's solution. To exclude errors attributable to nonlinearity of camera dynamic range, fluorescence of <15 or >245 Gy was rejected. To exclude errors attributable to image shading,20 data were analyzed within the central 80% of the image.
The 340/380 ratio was converted to nanomolar [Ca2+]i by the method of Grynkiewicz et al,12 using the reported fura-Ca2+ Kd of 224 nmol/L and our calibrations for the maximum 340/380 ratio, the minimum 340/380 ratio, and the Sf2/Sb2 ratio, where Sf2 and Sb2 are values for fluorescence resulting from excitation at 380 nm under Ca2+-free and Ca2+-rich conditions, respectively. Calibrations involved capillary infusions of Ca2+-rich (Ca2+=1.5 mmol/L) and Ca2+-free Ringer's solution, each containing 1 mmol/L Mg2+ and the membrane pore–forming agent ionomycin (10 μmol/L). Ca2+-free Ringer's solution was prepared without the addition of Ca2+ and contained EGTA (0.5 mmol/L). [Ca2+] of each solution was confirmed by Ca2+ electrometry (M93-20, Orion).
Fluorescence Colocalization Studies
The fluorescence of a second fluorophore was determined by quenching fluorescence in fura 2–loaded cells by exposing the capillary to unfiltered mercury lamp illumination and then reloading the capillary with the second fluorophore.
To load alveolar epithelial cells with fura 2-AM, in three lungs single alveoli were micropunctured using previously described methods16 17 and microinfused with fura 2-AM for 20 minutes. Residual fura 2-AM in the alveolar lumen was washed out by a microinfusion of buffer, and the alveoli were imaged for [Ca2+]i determinations as described above.
Using FFT for time to frequency domain conversion,21 we determined amplitude- and phase-frequency relationships of the ratiometric signal (MathCAD V5.0, MathSoft Inc; Origin V 4.0, Microcal Software Inc) in images obtained in 5- to 10-minute runs at 10-second intervals. FFT analyses of imaging noise in our recording system, determined by recording fluorescence of fura 2 in Ringer's solution held in a 50-μm-diameter glass capillary, revealed component waves with amplitudes of <10 nmol/L. Hence, to reject contamination from noise data, in the experimental FFT analyses all wave components with an amplitude of <10 nmol/L were rejected. Phase differences between paired oscillations were determined at identical frequency for the wave component that had the highest amplitude.
Data are mean±SE. Experiments are reported from 35 capillaries from 19 lungs (n indicates the number of capillaries). Differences were tested by paired and unpaired t tests as well as by the Wilcoxon matched-pair test, as appropriate. Differences among more than two groups were tested by the ANOVA Newman-Keuls test. Significance for these tests and of correlation coefficients obtained for pairs of data was accepted at P<.05.
Imaging Endothelial [Ca2+]i in the Lung Capillary
Endothelial cells lining subpleural capillaries of the isolated blood-perfused rat lung were loaded with fura 2, as described in “Materials and Methods,” and then the capillaries were imaged by intravital fluorescence microscopy for endothelial [Ca2+]i determination by the ratiometric method. All images were obtained between 10 minutes and 2 hours after fura 2 loading. Capillary images were taken at different excitations to obtain the 340/380 ratiometric image for [Ca2+]i quantification as shown for a single experiment in Fig 1A⇓ through 1C. The plane of best focus was set at maximum capillary diameter by adjusting the microscope objective. We determined that in the experimental period, the Ca2+-independent fluorescence of fura 2 attributable to excitation at 360 nm remained constant (not shown). Therefore, progressive deesterification of fura 2-AM to fura 2 was not detectable.
Identification of Individual Endothelial Cells
The capillary margin was marked by a series of fluorescence spots (Fig 1A and 1B⇑⇑) that in double-labeling experiments colocalized with fluorescence of the nuclear staining dye Hoechst 33324 (Fig 1D⇑) (n=5). This colocalization confirmed that each fluorescence spot delineated the nuclear-perinuclear region of a single endothelial cell. [Ca2+]i was spatially determined on the 340/380 image with respect to cell location determined on the 380-nm image (Fig 1B and 1C⇑⇑). Since [Ca2+]i at nuclear-perinuclear and cytoplasmic regions revealed no differences, we conclude that a nucleus–cytoplasmic Ca2+ gradient was absent.
To identify the phenotype of cells lining the capillary wall, we infused capillaries with the fluorescent endothelial marker diI-AcLDL (n=4). In every case, fluorescence was detectable in capillary wall cells, confirming the endothelial phenotype of the cells. As shown for one capillary in Fig 2A⇓, double-labeling experiments indicated that fura 2 fluorescent cells were always positive for diI-AcLDL fluorescence. Therefore, nonendothelial cell types, including pericytes22 and smooth muscle cells, which take up fura 2 but not diI-AcLDL, were not present. Moreover, in saponin-permeabilized capillaries exposed to anti–smooth muscle actin antibody fluorescently labeled with Cy3, we recorded only dark images (not shown), which are consistent with the absence of smooth muscle actin in these vessels. The absence of smooth muscle cells was also confirmed by immunohistochemistry using anti–smooth muscle actin antibody, which stained negatively in subpleural microvessels but positively in larger vessels that lay deeper in the parenchyma (Fig 2B⇓).
Effect of Digitonin
In cultured cells, time-dependent sequestration of anionic dyes, such as fura 2, occurs into intracellular organelles and is revealed by the presence of residual intracellular fluorescence after treatment of the cells with low concentrations of detergents (eg, 10 μmol/L digitonin).23 24 However, no residual cell fluorescence was detectable after 5-minute infusions of digitonin (5 to 20 μmol/L) for up to 2 hours after fura 2 loading (n=7, not shown). Hence, (1) fura 2 sequestration into intracellular organelles was not evident, and (2) fura 2 fluorescence was intracellular, because cell fluorescence decreased only when the cell was permeabilized by digitonin.
The Pacemaker-Generated Intracellular Ca2+ Wave
In 16 of 19 capillaries imaged under baseline conditions, periodic [Ca2+]i increases occurred spontaneously and sequentially from cell to cell along capillary wall segments. We imaged these sequential [Ca2+]i increases in 12 capillaries as intercellular Ca2+ waves, which originated from a single cell that may be termed the “pacemaker.” In the example in Fig 3A⇓, note that comparison of frames 1 and 2 reveals the beginning of the Ca2+ wave as an increase of [Ca2+]i solely in a pacemaker cell located at the capillary branch point (Fig 3A⇓, lower arrow). [Ca2+]i increases followed sequentially in other cells and reached maximum in frame 5, at which time the increases extended to a distance of 47 μm from the pacemaker (upper arrow in Fig 3A⇓). The subsequent recession of the wave followed a symmetrical pattern of sequential decreases of [Ca2+]i. Throughout, [Ca2+]i remained higher in the pacemaker than in other cells of the segment.
Comparisons of [Ca2+]i values averaged over a 5-minute period among endothelial cells of an unbranched capillary segment revealed higher mean [Ca2+]i (120±8 nmol/L) in pacemaker than in nonpacemaker (92±6 nmol/L) cells (n=15, P<.01). Pacemakers were detected in every capillary and, in 12 of 15 capillaries, were located at vessel branch points. Although we identified only one pacemaker cell per segment, we cannot discount the possibility that others may have existed or that the Ca2+ wave evident in a capillary segment was attributable to pacemaker cells that lay outside the observed focal plane.
By contrast with untreated capillaries, images of capillary walls obtained during intracapillary infusion of heptanol, the gap junctional uncoupler,11 were markedly lacking in Ca2+ waves (Fig 3B⇑) (n=5). Although this result indicates that intercellular Ca2+ wave propagation in the lung capillary is attributable to gap junctional communication between endothelial cells,11 [Ca2+]i continued to oscillate in some pacemaker cells despite heptanol treatment. An example of such a pacemaker cell, identified as such in control recordings (not shown) as cells that originated the Ca2+ wave and in which mean [Ca2+]i was higher than in surrounding cells, is shown in the capillary in Fig 3B⇑ (arrow). Hence, even when not propagated, [Ca2+]i increases were generated in pacemaker cells by unidentified intrinsic mechanisms.
The spread of intracellular Ca2+ from the pacemaker was also evident as synchronous [Ca2+]i oscillations in all cells of the capillary segment. Fig 4A⇓ depicts [Ca2+]i oscillations in cells a and d of the capillary shown in Fig 1A⇑, which, despite their separation, showed simultaneous [Ca2+]i peaks at the 1-, 2-, and 3-minute time points. Not shown are oscillations in cells b and c (Fig 1A⇑), which were also synchronous with those of a and d.
Effect of Blood Flow
To determine the possible effect of blood flow, at constant capillary pressure we stopped blood flow for 10 minutes (n=4). However, blood flow stoppage had no effect on the [Ca2+]i oscillations (Fig 4B⇑). Therefore, the [Ca2+]i oscillations were not attributable to motion of blood in the capillary. Moreover, since the lungs were held at constant inflation, the oscillations were not attributable to lung volume fluctuations.
Cultured Endothelial Cells
By contrast with intact endothelium, [Ca2+]i oscillations were absent in unstimulated cultured bovine pulmonary artery endothelial cells (Fig 4C⇑, n=5). This is consistent with previous reports25 26 and reveals an unexplained difference in [Ca2+]i behavior between cultured and in situ endothelial cells.
Alveolar Epithelial [Ca2+]i
For comparison with endothelial [Ca2+]i, we quantified epithelial [Ca2+]i in alveoli adjacent to capillaries by loading them with fura 2 by micropuncture (see “Materials and Methods”). To rule out imaging noise as a possible cause of the observed capillary [Ca2+]i oscillations, we determined that no significant [Ca2+]i oscillations were evident in alveolar epithelial cells imaged under identical conditions (Fig 4D⇑, n=5). To rule out other artifacts, we also determined that the oscillations were not evident (1) in the Ca2+-insensitive fluorescence of fura 2 excited at 360 nm (not shown), which rules out fluctuations in cell fura 2 content as a factor in the oscillations, and (2) after intracapillary injections of dextran-saline at 4°C (Fig 4E⇑) (n=4, P<.01), which confirmed that the oscillations resulted from temperature-sensitive cellular processes.
Effect of Dextran
To rule out receptor-mediated effects that may have induced the intercellular Ca2+ waves,26 27 we infused neutral dextran (70 kD) to replace capillary blood with a solution lacking any known ligands for endothelial receptors. In different experiments, dextran-Ringer was given either with 1.5 mmol/L Ca2+ (n=6) or in a Ca2+-free solution (n=4) as 5- or 30-minute infusions. Ca2+-free dextran-Ringer was prepared by replacing Ca2+ with Mg2+ and by inclusion of the Ca2+ chelator EGTA (0.5 mmol/L). In both sets of experiments, dextran infusions caused no detectable effect on [Ca2+]i oscillations or in mean [Ca2+]i of any cell. Results from one experiment, shown in Fig 5⇓, exemplify these findings, which indicate that [Ca2+]i oscillations occurred independently of the binding of soluble ligands to endothelial receptors or to the influx of extracellular Ca2+. However, as exemplified by the experiment shown in Fig 5⇓, the endosomal Ca2+-ATPase pump inhibitor thapsigargin27 28 effectively inhibited the [Ca2+]i oscillations in every case, when given as an intracapillary infusion in a Ca2+-free solution (n=4, P<.01). The thapsigargin-sensitive inhibition in the Ca2+-free condition indicates that the oscillations were entirely attributable to endosomal Ca2+ release.
Frequency Analyses by FFT
As shown for one experiment in Fig 6A⇓, in every case FFT analyses of the [Ca2+]i oscillations revealed wave components of which the dominant (wave with highest amplitude) occurred at the low-frequency end of the amplitude-frequency spectrum. We restricted further consideration to this dominant component, which had an amplitude of 37±5 nmol/L at a frequency of 0.4±0.1 min−1 (n=16). Not shown are amplitude-frequency plots for the stopped-flow condition and the dextran infusion experiments (n=12), all of which gave profiles similar to baseline as expected.
Pacemaker [Ca2+]i oscillations were phase-advanced 21±4° over oscillations of adjoining nonpacemaker cells (Fig 6B⇑, solid bar on left) (P<.01). However, the phase locking deteriorated with intercellular distance along the capillary. Phase difference was markedly higher between remote pairs of cells selected from two different capillary segments that were separated by a branch (Fig 6B⇑, open bar on left); hence, intercommunication of cytosolic Ca2+.deteriorated with intercellular distance.
We calculated a propagation velocity (v) of the intracellular Ca2+ wave of 5±2 μm/s (n=9) from the wave-velocity equation v=2πfd/Δθ, using values for intercellular phase delay (Δθ) of the dominant wave component at frequency f and internuclear distance d, between adjoining cells.
FFT analyses of the heptanol data (n=5) indicated a heptanol-induced threefold increase of phase difference between pacemaker and nonpacemaker cells (Fig 6B⇑) (P<.01). Therefore, heptanol deteriorated the phase locking of [Ca2+]i oscillations between cells. Under control conditions, the amplitude of the dominant wave component was not significantly different between pacemaker and nonpacemaker cells (Fig 6C⇑). After heptanol, mean [Ca2+]i for pacemaker and nonpacemaker cells did not significantly differ from baseline. However, heptanol decreased the amplitude in both cell types, although in pacemaker cells the decrease did not achieve statistical significance (Fig 6C⇑). The greater decrease of the amplitude in nonpacemaker than in pacemaker cells signifies the cell uncoupling effect of heptanol, which blocked communication of an identical component of the Ca2+ wave between the two cell types.
Effect of Histamine
To determine the interaction of pacemaker and nonpacemaker cells during agonist-induced [Ca2+]i increases, in separate groups we obtained baseline [Ca2+]i measurements during capillary infusions of Ca2+-containing dextran. Then we followed with continuous infusions of 5 μmol/L histamine in Ca2+-containing or Ca2+-free dextran. To determine the effects of gap junction uncoupling, in one group we included heptanol (3 mmol/L) with histamine infusion in Ca2+-free dextran.
The control histamine-induced endothelial [Ca2+]i response in the presence of external Ca2+ characteristically consisted of an initial transient and then a sustained [Ca2+]i increase (Fig 7A⇓). This typical biphasic response, previously reported in cultured endothelial cells, is attributable to the sequential effects of endosomal Ca2+ release followed by external Ca2+ entry (reviewed by Hosoki and Iijima29 ). The mean [Ca2+]i and the mean amplitudes of [Ca2+]i oscillations were averaged over a 4-minute period following the initial transient and are shown in Tables 1⇓ and 2.
Control histamine-induced [Ca2+]i increases were similar in pacemaker and nonpacemaker cells. Although mean [Ca2+]i was higher in pacemakers than in nonpacemakers as before, histamine-induced [Ca2+]i increases were synchronous (Fig 7A⇑) and not significantly different between the two cell types (Table⇑s 1 and 2, data column 1). Histamine markedly increased mean [Ca2+]i and markedly augmented [Ca2+]i oscillations, of which the amplitudes increased, on average, to more than four times the baseline. Oscillation frequency, which was identical in the two cell types, increased from 0.5±0.1/min at baseline to 0.8±0.2/min after histamine (P<.05). The magnitude of the frequency increase was significantly less than that of the amplitude increase (P<.05).
Similar to reported data for cultured endothelial cells,30 removal of external Ca2+ markedly damped these histamine-induced [Ca2+]i responses in the capillary. The increases of mean [Ca2+]i were inhibited by 54% and 41% of control in pacemakers and nonpacemakers, respectively (Table 1⇑, column 2). The corresponding inhibitions of the amplitude increases amounted to 57% and 46% (Table 2⇓, column 2). However the increases of oscillation frequency were not significantly different from control (not shown).
A marked difference between the pacemaker and nonpacemaker [Ca2+]i responses to histamine became evident in the presence of heptanol. In pacemakers, increases of mean [Ca2+]i and oscillation amplitude, which were already blunted in external Ca2+-free conditions, were not further inhibited in the presence of heptanol (Fig 7C⇑; Tables 1 and 2⇑⇑, column 3). Therefore, heptanol had no independent Ca2+-inhibiting effect in pacemakers; hence, a toxic Ca2+-depressing effect attributable to heptanol may be ruled out. By contrast, in nonpacemakers, heptanol given in external Ca2+-free conditions completely inhibited [Ca2+]i responses to histamine in that we detected no significant increases in either mean [Ca2+]i or oscillation amplitude (Fig 7C⇑; Tables 1 and 2⇑⇑, column 3). We conclude that intercellular communication of the augmented pacemaker Ca2+ wave accounted for a significant part of the nonpacemaker [Ca2+]i response to histamine.
The substantial [Ca2+]i increase of the pacemaker cell in external Ca2+-free conditions (Table 1⇑, column 2) suggested histamine-induced Ca2+ release from endosomal stores. To test this possibility, we depleted endosomal Ca2+ stores using thapsigargin infusions in Ca2+-free dextran. As shown for one experiment in Fig 8⇓, thapsigargin caused the usual cytosolic Ca2+ transient attributable to endosomal Ca2+ release. However, following thapsigargin administration, histamine failed to increase [Ca2+]i (n=4). These results, which confirm similar findings in cultured endothelial cells,27 29 indicate that Ca2+ release from thapsigargin-sensitive endosomal stores constituted an important mechanism in the pacemaker [Ca2+]i response to histamine.
In the present study, we show for the first time that spontaneous low-amplitude Ca2+ waves exist under resting conditions in lung capillary endothelium. The waves were evident as periodic sequential increases and decreases of [Ca2+]i along rows of cells in capillary wall segments. The waves propagated short distances of ≈50 μm along the capillary wall. A major feature was that the waves appeared to originate from specific endothelial cells that were located mostly at vessel branch points. We have called these the “pacemaker” cells, although we recognize that the waves may have originated in other cells located outside our visual field. The waves were inhibited by heptanol, the gap junction uncoupler; hence, the presence of these waves indicates that endothelial cells of the lung capillary intercommunicate with intracellular Ca2+ signals under resting conditions and that the communication probably occurs across gap junctions.
We selected the 20- to 25-μm-diameter lung venular capillary to facilitate measurements in a vascular wall consisting purely of endothelial cells. Smooth muscle cells and pericytes are unlikely to have existed, because lung vessels of diameter <30 μm lack smooth muscle, especially in the venous system,31 and lung pericytes are rare, particularly in small mammals.32 In any case, the endothelial phenotype of the experimental cells was confirmed by their positive staining with diI-AcLDL, an endothelial cell marker, but not with anti–smooth muscle actin, a pericyte and smooth cell marker.22
The postulated pacemaker cells were clearly distinguished from other cells of the capillary segment, because they were phase-advanced in [Ca2+]i oscillations. In addition, their mean [Ca2+]i was higher, and in the majority of capillaries, they originated directly imaged intercellular Ca2+ waves (Fig 3A⇑). Despite inhibition of wave propagation by heptanol, pacemaker [Ca2+]i continued to oscillate with a dominant amplitude that was not significantly different from control. By contrast, in other cells heptanol markedly inhibited [Ca2+]i oscillations. These findings dispel the notion that [Ca2+]i regulation is homogeneous in lung endothelial cells and lead to the important conclusion that in lung capillary endothelium, a subclass of endothelial cells, the pacemakers, generates [Ca2+]i oscillations that are communicated to adjoining cells.
Although we imaged only linear propagation of the waves along the capillary margin in our focal plane, it is likely that the waves also spread to regions outside our focal plane. Typically, in the wave propagation path we detected two to four endothelial cells in which we determined synchronous [Ca2+]i oscillations. The absence of [Ca2+]i oscillations and waves in in situ alveolar epithelial cells and cultured endothelial cells imaged under identical conditions rules out technical artifacts as a cause of the capillary [Ca2+]i oscillations. Two potential difficulties of the fura 2 technique are attributable to dye sequestration into organelles by organic anionic transporters24 and progressive intracellular deesterification of fura 2-AM. However, these were not significant problems in the present study, because within the experimental period, neither dye sequestration (which is evident as residual cell fluorescence following digitonin permeabilization23 24 ) nor progressive deesterification (which should increase the Ca2+-independent fluorescence of fura 2) was detectable.
We considered the possibility that the Ca2+ waves may be shear-induced, because in cultured endothelial cells exposed to flow, [Ca2+]i oscillations directly correlate with applied shear stress; cessation of perfusion immediately abolishes the oscillations.25 By contrast, 10-minute stoppages of blood flow at constant capillary pressure had no effect on the present [Ca2+]i oscillations; hence, they were not shear-induced. However, an important result was obtained when we infused capillaries with dextran-Ringer, which either contained Ca2+ or was Ca2+ free. Replacement of capillary blood with these ligand-free solutions for up to 30 minutes failed to inhibit the [Ca2+]i oscillations. Therefore, the oscillations were attributable neither to ligation of endothelial receptors by soluble ligands in blood nor to entry of extracellular Ca2+ into the endothelial cell. However, thapsigargin, which blocks endosomal uptake of Ca2+ by inhibiting the endosomal Ca2+-ATPase pump,27 completely inhibited the oscillations when given with Ca2+-free dextran. This thapsigargin-induced inhibition indicates that the oscillations were entirely attributable to intracellular Ca2+ release from a thapsigargin-sensitive endosomal pool.
Although intercellular and intracellular Ca2+ waves have been reported under stimulated conditions in several cell types in vitro,2 33 34 the unique feature of the capillary Ca2+ waves reported in the present study is that they were generated under quiescent unstimulated conditions. Our findings in cultured endothelial cells confirm several reports that [Ca2+]i oscillations are absent in resting endothelial cells in vitro,25 35 although some have reported the presence of sporadic oscillations.1 36 Spontaneous Ca2+ waves and oscillations have been described thus far only in excitable cells,2 34 such as skeletal, cardiac, and smooth muscle.34 For example, in vascular smooth muscle cells, intracellular Ca2+ waves (Ca2+ sparks), which may subserve pacemaker function, are endosomally generated by IP3R-gated channels.34 The present capillary Ca2+ waves are similar to smooth muscle Ca2+ sparks in amplitude but have lower frequency. The propagation of intercellular Ca2+ waves is attributable to a Ca2+-release cascade induced by gap junctional diffusion of IP3 between cells.3 IP3R ligation releases Ca2+, which in turn may stimulate further IP3 production by inducing, for example, activation of the Ca2+-sensitive phospholipase A237 pathway and, consequently, arachidonate-induced [Ca2+]i increase.38 The extent to which these mechanisms are applicable here remains unclear.
FFT analyses of the endothelial [Ca2+]i oscillations revealed several wave components, of which the dominant had an amplitude of 37 nmol/L and frequency of 0.4 min−1. We restricted further analyses to the dominant wave, which, at a wave velocity of 5 μm/s, is among the slowest reported for mammalian Ca2+ waves.33 FFT analyses also confirmed that pacemakers were phase-advanced over adjoining cells, further supporting their role as possible origins of the Ca2+ wave. Although the oscillations were well phase-locked within a nonbranched segment of the capillary, phase locking was considerably worse between cells taken from capillary segments separated by a branch. We interpret from these findings that each nonbranched segment of the lung capillary generates its own Ca2+ wave, which causes phase-locked [Ca2+]i oscillations among cells within the segment. The wave appears not to propagate significantly beyond the segment; hence, under resting conditions, [Ca2+]i regulation is probably better coordinated within the segment than at the intersegmental level.
The resolution of oscillatory data by FFT analyses, as used in the present study, may incur underestimates of wave resolution attributable to sampling period (ie, time interval between successive data points) and sample size (ie, the total number of points included in the analysis). Sampling period determines the extent of the frequency range returned by the analysis; sample number correlates directly with the frequency resolution. To avoid possible fluorescence quenching attributable to fluorescent excitation of capillaries for prolonged periods, we restricted our imaging runs to 5 minutes and obtained 32 images at 10-second intervals. Although our analyses returned dominant amplitudes in the frequency range of 0.2 to 1 min−1, it is possible that this wave is itself a composite of other waves that may be further resolved with larger sample size.
The [Ca2+]i Response to Histamine
The principal aim of the histamine experiments was to determine the effects of the pacemaker wave on capillary [Ca2+]i regulation under agonist-stimulated conditions. We modeled histamine-induced endothelial [Ca2+]i increases on the following mechanisms: (1) external Ca2+ entry,29 39 (2) endosomal Ca2+ release,29 39 and (3) transmission of the pacemaker wave. Control responses represented the combined effects of all mechanisms. The external Ca2+-free condition ruled out mechanism 1. A combination of the external Ca2+-free condition with heptanol ruled out mechanisms 1 and 3.
In pacemakers, ≈46% of the total mean [Ca2+]i increase was attributable to Ca2+ entry, as judged from the abrogated response in external Ca2+-free conditions (Table 1⇑, column 2). The remaining component of the response was evidently Ca2+-entry independent and may be attributed to endosomal Ca2+ release, because the lack of a heptanol effect on pacemaker [Ca2+]i leads us to consider intracellular rather than intercellular factors in pacemaker [Ca2+]i regulation. The importance of pacemaker endosomal mechanisms was directly indicated in the complete inhibition of the histamine-induced [Ca2+]i increase following depletion of endosomal stores by thapsigargin. These findings reaffirm our view that thapsigargin-sensitive endosomal mechanisms40 played a major role in pacemaker [Ca2+]i regulation.
These considerations may also be relevant to the impressive histamine-induced increase of oscillation amplitude, which may be attributable to endosomal mechanisms analogous to those of Ca2+-induced Ca2+ release. In Ca2+-induced Ca2+ release, Ca2+ entry primes Ca2+-dependent endosomal receptors, such as IP3R,41 and augments receptor-mediated endosomal Ca2+ release.2 The lack of such priming may explain the reduction of oscillation amplitude in Ca2+-free conditions (Table 2⇑).
In nonpacemakers, removal of external Ca2+ inhibited the increase of mean [Ca2+]i to 59% of control (Table 1⇑, column 1). Hence, similar to pacemakers, the nonpacemaker mean [Ca2+]i increase to histamine was dominated by Ca2+ entry, and 41% of the response was attributable to Ca2+ entry–independent mechanisms. But by contrast with pacemakers, the Ca2+ entry–independent component was completely blocked by heptanol (Fig 7C⇑). These findings lead to the important conclusion that a significant part of the nonpacemaker [Ca2+]i response to histamine was attributable to the heptanol-inhibitable pacemaker wave. Importantly, we note that under heptanol inhibition, endosomal mechanisms in the nonpacemaker played no detectable role in rescuing the [Ca2+]i response to histamine. Hence, we conclude that endosomal Ca2+-release mechanisms respond differently in nonpacemaker and pacemaker cells, although the basis for this difference remains unknown.
In conclusion, the present results indicate that spontaneous intercellular spatiotemporal Ca2+ signaling, attributable to pacemaker-generated Ca2+ waves, is a constitutive property of the in situ lung microvascular endothelium. Endothelial [Ca2+]i regulation is complex in the lung capillary and differs in pacemaker and nonpacemaker cells. Under stimulated conditions in the presence of histamine, [Ca2+]i increases in pacemaker cells are regulated by endosomal Ca2+ release and external Ca2+ entry. By contrast, endosomal mechanisms are less important in the nonpacemaker cell, in which external Ca2+ entry and the pacemaker-generated Ca2+ wave regulate the bulk of the [Ca2+]i increase. The physiological advantage gained from endothelial Ca2+ intercommunication appears to be that [Ca2+]i increases are established in all cells of the capillary segment. Such intercommunication may be important in the organization of multicellular endothelial responses in the lung microvessel. In hepatocytes in situ, vasopressin induces pacemaker-generated intercellular Ca2+ waves that propagate codirectionally with pericentral-periportal bile flow.5 In lung microvessels, periodic [Ca2+]i increases may optimize resting endothelial functions, such as NO production28 42 and maintenance of the mitochondrial redox potential.43 By generating intercellular Ca2+ waves, pacemakers may coordinate [Ca2+]i in capillary endothelial cells and may thereby achieve functional homogeneity in the capillary wall. The extent to which these intercellular Ca2+ signals modulate global capillary responses under inflammatory conditions requires further consideration.
Selected Abbreviations and Acronyms
|diI-AcLDL||=||diI-acetyl low-density lipoprotein|
|FFT||=||fast Fourier transform|
This study was supported by National Institutes of Health grants HL-36024 and HL-53625. R. Kashyap assisted in analyses of imaging data.
- Received April 1, 1996.
- Accepted August 8, 1996.
Demer LL, Wortham CM, Dirksen ER, Sanderson MJ. Mechanical stimulation induces intercellular calcium signaling in bovine aortic endothelial cells. Am J Physiol. 1993;264:H2094-H2102.
Boitano S, Dirksen ER, Sanderson MJ. Intercellular propagation of calcium waves mediated by inositol trisphosphate. Science. 1992;258:292-295.
Robb-Gaspers LD, Thomas AD. Coordination of Ca2+ signaling by intercellular propagation of Ca2+ waves in the intact liver. J Biol Chem. 1995;270:8102-8107.
Nathanson MH, Burgstahler AD, Mennone A, Fallon MB, Gonzalez CB, Saez JC. Ca2+ waves are organized among hepatocytes in the intact organ. Am J Physiol. 1995;269:G167-G171.
Stauffer PL, Zhao H, Luby-Phelps K, Moss RL, Star RA, Muallem S. Gap junction communication modulates [Ca2+]i oscillations and enzyme secretion in pancreatic acini. J Biol Chem. 1993;268:19769-19775.
Nedergaard M. Direct signaling from astrocytes to neurons in cultures of mammalian brain cells. Science. 1994;263:1768-1771.
McDonald DM. Endothelial gaps and permeability of venules in rat tracheas exposed to inflammatory stimuli. Am J Physiol. 1994;266:L61-L83.
Little TL, Xia J, Duling BR. Dye tracers define differential endothelial and smooth muscle coupling patterns within the arteriolar wall. Circ Res. 1995;76:498-504.
Grynkiewicz G, Poenie M, Tsien RY. A new generation of Ca2+ indicators with greatly improved fluorescence properties. J Biol Chem. 1985;260:3440-3450.
Tsukada H, Ying X, Fu C, Ishikawa S, McKeown-Longo P, Albelda S, Bhattacharya S, Bray BA, Bhattacharya J. Ligation of endothelial αvβ3 integrin increases capillary hydraulic conductivity of rat lung. Circ Res. 1995;77:651-659.
Ying X, Qiao R, Ishikawa S, Bhattacharya J. Removal of albumin microinjected in rat lung perimicrovascular space. J Appl Physiol. 1994;77:1294-1302.
Sadurski R, Tsukada H, Ying X, Bhattacharya S, Bhattacharya J. Diameters of juxtacapillary venules determined by oil-drop method in rat lung. J Appl Physiol. 1994;77:718-725.
Bhattacharya J, Gropper MA, Shepard JM. Lung expansion and the perialveolar interstitial pressure gradient. J Appl Physiol. 1989;66:2600-2605.
Ishikawa S, Tsukada H, Bhattacharya J. Soluble complex of complement increases hydraulic conductivity in single microvessels of rat lung. J Clin Invest. 1993;91:103-109.
Bhattacharya S, Bhattacharya J. Segmental vascular responses to voltage gated calcium channel potentiation in rat lung. J Appl Physiol. 1992;73:657-663.
Bolsover SR, Silver RA, Whitaker M. Ratio imaging measurement of intracellular calcium and pH. In: Shotton DM, ed. Electronic Light Microscopy. New York, NY: Wiley-Liss; 1993:181-210.
Brigham EO. The Fast Fourier Transform. Englewood Cliffs, NJ: Prentice-Hall; 1974.
Shepro D, Morel NML. Pericyte physiology. FASEB J. 1993;7:1031-1038.
Virgilio FD, Steinberg TH, Swanson JA, Silverstein SC. Fura-2 secretion and sequestration in macrophages: a blocker of organic anion transport reveals that these processes occur via a membrane transport system for organic anions. J Immunol. 1988;140:915-920.
Oike M, Droogmans G, Nilius B. Mechanosensitive Ca2+ transients in endothelial cells from human umbilical vein. Proc Natl Acad Sci U S A. 1994;91:2940-2944.
Buckley BJ, Mirza Z, Whorton AR. Regulation of Ca2+-dependent nitric oxide synthase in bovine aortic endothelial cells. Am J Physiol. 1995;269:C757-C765.
Hamilton KK, Sims PJ. Changes in cytosolic Ca2+ associated with von Willebrand factor release in human endothelial cells exposed to histamine. J Clin Invest.. 1987;79:600-608.
Mello DD, Reid L. Arteries and veins. In: Crystal RG, West JB, Barnes PJ, Chemlock NS, Weibel ER, eds. The Lung: Scientific Foundations. New York, NY: Raven Press Publishers; 1991:767-778.
Jaffe LF. The path of calcium in cytosolic calcium oscillations: a unifying hypothesis. Proc Natl Acad Sci U S A. 1991;88:9983-9887.
Nelson T, Cheng H, Rubart M, Santana LF, Bonev AD, Knot HJ, Lederer WJ. Relaxation of arterial smooth muscle by calcium sparks. Science. 1995;270:633-636.
Roudbaraki MM, Vacher P, Drouhault R. Arachidonic acid increases cytosolic calcium and stimulates hormone release in rat lactotrophs. Am J Physiol. 1995;268:E1215-E1223.
Papp B, Paszty K, Kovacs T, Sarkadi B, Gardos G, Enouf J, Endyedi A. Characterization of the inositol trisphosphate-sensitive and insensitive calcium stores by selective inhibition of the endoplasmic reticulum-type calcium pump isoforms in isolated platelet membrane vesicles. Cell Calcium. 1993;14:531-538.