The relative radiation sensitivities of the various compartments of the heart are poorly characterized. Cardiac fibrosis is a common side effect of radiotherapy, suggesting that endothelial barrier function is an important factor in radiation-induced pathology. We employed Electric Cell Substrate Impedance Sensing (ECIS) to assess cytoskeletal rearrangement, permeability changes and endothelial barrier function changes in response to radiation in studies of human coronary arterial endothelial cells (HCAECs). A 5-Gy dose of γ radiation resulted in a significant sixfold transient decrease in transmonolayer resistance 3 h postirradiation (P = 0.001). This decrease in resistance coincided with changes in fluorescent tracer flux (P = 0.05) and display of an actin bundling phenotype. After irradiation, decreases in wound healing (P = 0.03) and micromotion within the monolayer (P = 0.02) were also observed. Time-lapse video studies confirmed that the monolayer is dynamic and showed that cells are extruded from the monolayer at a higher frequency after irradiation. These findings suggest that perturbed endothelial barrier function in the heart can occur at lower doses of γ radiation than previously reported.
Radiation-induced heart disease is of concern given the exposure of the heart not only in therapeutic management of breast cancer and Hodgkin's disease but also prophylaxis in prostate cancer patients to counter gynecomastia (1, 2). More recently, interest in the effects of occupational, exposure to low-dose radiation on the heart has increased (3, 4). At doses greater than 50 Gy, overt damage to the relatively radioresistant myocardium is seen (5, 6). However, at doses of approximately 2 Gy encountered in clinical fractionated radiotherapy, radiation-induced heart disease is believed to be a disorder of the endothelial cells in the heart.
Several aspects of normal vascular function are ablated by ionizing radiation. Normal function entails the ability of the vasculature to contract or relax in response to vasoactive hormones and cytokines as well as the ability to form a competent barrier that will resist thrombosis and respond to signals from the immune system. In response to injury or infection, normal endothelial immune response involves altered surface expression of selectins and several other adhesion molecules to facilitate rolling of leukocytes, which, after activation, extravasate where they then function (7, 8). Radiation can inappropriately activate this sequence of events, presumably in response to oxidative damage induced at the site (9). Contraction and relaxation kinetics are altered in ex vivo experiments where sections of great vessels are irradiated (10–12). These mechanisms, involving nitric oxide, are important in maintenance of vascular tone (13). The contribution of prostanoid biology in the vascular radiation response is exemplified by the reduction in radiation-induced atherosclerotic plaque formation in the carotid arteries by COX2 inhibition (14). Thus prostanoid-dependent vascular contractility, prostanoid-independent vascular contractility, and immune response are all ablated by radiation.
In addition to effects on contractility and paracrine signaling, radiation also causes overt changes in barrier function to the extent that dyes and tracers can flow out of the vasculature after irradiation (15). The ramifications of radiation-induced barrier failure in smaller vessels has been studied, and the result appears to be a modified immune response resulting in the deposition and remodeling of the extracellular matrix, leading to fibrosis (16, 17). In the heart, ectopic remodeling of the matrix will have deleterious effects on cardiac function (18, 19). Failure of this barrier can inappropriately admit soluble factors from blood that can then serve as stimulants for leukocyte homing and smooth muscle cell proliferation, resulting in an atherosclerotic lesion (14, 20, 21). Such lesions are also induced in young radiotherapy patients with no pre-existing atherosclerosis (22). The mechanism underlying the iatrogenic generation of both these lesions and fibrosis is of great interest.
At the molecular level, the initiation of endothelial cell death in response to radiation is concomitant with the activation of stress-activated kinase cascades, including the transduction of a radiation damage signal via p38 and JNK (23). In primary cells from the aorta, radiation-induced apoptosis is mediated by NF-κB signaling (24). In spite of this assembled body of work, the details of endothelial dysfunction in radiation-induced heart disease remain unclear. Further, the relative sensitivities of cardiac substructures to radiation damage have not been determined. In addressing these questions, we considered that endothelium from bovine adrenals exhibits a compromised ability to organize the cytoskeleton into a migration-competent apparatus when assessed 24 h after irradiation, producing a delay in wound closure that is also seen in dermal microvascular endothelial cells (DMVECs) (25, 26). Furthermore, at least one type of endothelial cells are known to respond to radiation with a robust Rho-mediated cytoskeletal rearrangement at doses of 20 Gy with relatively rapid kinetics of 10–30 min (27). A key point of study, then, is the examination of very subtle changes in barrier function and cell dynamics of coronary endothelium in response to lower doses of radiation between 0 and 5 Gy.
Measurement of electrical impedance across layers of cells can provide information regarding cellular morphology and barrier function (28, 29). For this reason and several others, we selected Electric Cell Substrate Impedance Sensing (ECIS) as an investigative tool. First, the technique is non-invasive owing to the very small amplitude of interrogating currents employed (<40 μA/cm2). Second, the method is exquisitely sensitive, successfully reporting not only micromovement in confluent monolayers but also cellular changes in response to the fluctuations in ambient CO2 of ±0.5% in tissue culture incubators (30, 31). When cell monolayers are seeded on gold planar arrays, a variety of biological processes can be assessed, including spreading, adhesion, migration and transgression of the monolayer by invasive, heterotypic cells (32, 33). The transmonolayer impedance of endothelial cells in particular responds robustly when treated with compounds known to perturb endothelial barrier function, including thrombin, histamine and calcium chelating agents, and a model has been tested allowing for electronic assessment of permeability changes in response to these stimuli (34–36). Moreover, this technology allows for continuous cell monitoring over tens of hours, which is a unique and powerful feature of this experimental approach. The goal of this work was to assess radiation-induced barrier changes in human coronary endothelial monolayers using an impedance-based micromorphometry technique. Particular attention was paid to rearrangements of the cytoskeleton that could alter the barrier function of the tissue.
MATERIALS AND METHODS
Human coronary arterial endothelial cells (HCAECs) were obtained from Lonza (Rockville, ME) as either actively growing cultures or frozen ampoules. Cells were subcultured in a modified form of commercially available medium in which all additives were used with the exception of hydrocortisone (EGM-2-MV Bullet Kit, Lonza, Rockville, ME).
Conventional cultureware was coated with collagen by preparing a 30-μg/ml solution of Rat Tail Col I (BD) in water and overlaying the cultureware with the coating solution for 1 h at 37°C prior to seeding cultures on the growth surface. Cells were maintained in humidified incubator at 37°C and 95% air/5% CO2 and were fed three times per week.
Cells were transported to a nearby Gammacell 40 apparatus (Atomic Energy, Ontario, Canada) that delivered 137Cs γ rays at a dose rate of 0.79 Gy/min. Cultureware was insulated during transport and irradiation to limit the effects of temperature changes.
Electric Cell Substrate Impedance Sensing
All measurements were performed using an Applied Biophysics ECIS-Zθ instrument that allows measurement of complex impedance of cell monolayers over a range of frequencies. For the impedance-related experiments, “8W10E+” type ECIS arrays (Applied Biophysics, Troy, NY) were coated with extracellular matrix substrate by addition of 400 μl of a Rat Tail collagen 1 solution to each well. This array type (see Fig. 1A) allows for measurements at 40 regularly spaced working electrodes over the 0.8-cm2 surface of the culture with an interdigitated array of counterelectrodes forming the other terminus of the circuit. Collagen densities were 50 μg/cm2 culture area, and arrays were incubated at 37°C for 1 h. After coating, the wells were aspirated dry and blocked with a 400-μl volume of a 1% w/v solution of bovine serum albumin (BSA) in PBS. Blocking was also carried out for 1 h at 37°C. After blocking, wells were aspirated dry and HCAECs were seeded on gold planar electrodes at a density of 75 × 103 cells/cm2 in medium permitting an almost confluent density at the outset of the experiment. Impedance measurements made at 11 frequencies with an ECIS-Zθ instrument at 5-min intervals and logged on a command laptop computer running proprietary ECIS software.
For wound recovery experiments, “8W1E” type ECIS arrays were used. This second type of array allows for measurements at a single, central point (Fig. 1A, lower portion) with a working electrode 250 μm in diameter with a larger surrounding counterelectrode completing the circuit. Cells were allowed to mature to competent monolayers and were then subjected to an alternating current with an amplitude of 1400 μA for 20 s at a frequency of 60 kHz, effectively denuding the electrode surface (∼0.05 mm2) of cells. Impedance measurements then resumed, permitting surveillance of the rate at which cells migrated over the freshly denuded surface, allowing the monitoring of wound closure.
A survey of micromotion was conducted by investigating cells on “8W1E” electrodes for 5 h after wound recovery 5 h after the initial irradiation. Ohmic resistance values recorded at 4000 Hz on this single point array reflect the coverage of a 250-μm-diameter region by the cells residing in the monolayer. These acquired values were compiled, and a micromotion index (mmi) was calculated using the sum of the absolute values of differences between the data points. This index reflects the total swept distance of the recording trace and is described by Eq. (1) below, where R represents the resistance value in ohms/cm2 at 5-min intervals.
Cells were cultured on collagen-coated glass cover slips and blocked as above at area-corrected seeding densities relative to culture on ECIS arrays. Wells were aspirated dry, washed with PBS, fixed by addition of 4% paraformaldehyde in PBS for 10 min, and then stored overnight in PBS at 4°C. The next day, cells were permeabilized by addition of 0.5% Triton X-100 in PBS for 5 min and then left in PBS for 10 min prior to staining. Cover slips were placed face down on a 97-μl volume of staining solution and incubated for 1 h at 37°C. Staining solution was comprised of 10% normal goat serum in PBS supplemented with a 1:40 dilution of Alexa488-conjugated phalloidin (Molecular Probes, Eugene, OR). After staining, cover slips were immersed in PBS three times before being allowed to destain in a 5-ml volume of PBS for a subsequent 10 min. Cover slips were then mounted on 50-μl volumes of Lerner Aquamount aqueous mounting medium (Thermo Scientific, Kalamazoo, MI) prior to imaging. Confocal images were obtained using a Nikon C1 confocal scanning head coupled to a TE-2000-U microscope. An infinity-corrected Plan Fluor 40X extra long working distance objective with a numerical aperture of 0.6 was used. Data from the confocal sensor were captured with EZ C1 software generating 512 × 512-pixel images that were exported in TIFF format.
Lucifer Yellow Dye Flux
Transwell inserts with diameter of 12 mm and pore size of 0.4 μm (Costar, Corning, NY) were coated with collagen and blocked with BSA at area-corrected volumes and densities to permit comparable coating to those conditions used on ECIS arrays as described above. Cells were seeded at area-corrected densities and monolayers were allowed to mature on the porous substrates for 20 h, at which point cells were irradiated. Immediately after irradiation (or sham control treatment), 50 μl of a concentrated Lucifer Yellow solution was added to the lower compartment of the culture at a final concentration of 0.5 mg/ml. Samples of 10 μl were removed from the upper part of the culture at subsequent intervals and stored at 4°C until analysis. For fluorimetric analysis, a 1:20 dilution was made of a 5-μl portion of each sample in water, and this volume was subject to fluorimetry using a Synergy2 96-well plate reader under the command of Gen5 software (Biotek, Highland Park, VT). The reader was configured for excitation and emission wavelengths appropriate to Lucifer Yellow.
Radiation-Induced Changes in Electrical Properties of Coronary Endothelial Monolayers
HCAECs were seeded in all experiments at a uniform density 75 × 103 cells/cm2 on substrates that were coated with rat tail collagen I at a density of 50 μg/cm2. To assess the effects of radiation on this coronary tissue, we employed ECIS technology and seeded cells on collagen-coated gold planar arrays optimized for surveillance of transmonolayer resistance and associated sealing and permeability. We observed an immediate and early development of resistance attributed to contact and spreading of the cells on the array surface (Fig. 1B). This process appeared to resolve by 4 h after seeding. The next phase, marked by a further increase of resistance, appeared to be sealing and maturation of lateral contacts. This maturation appeared to peak in a range of 18 to 22 h and was characteristic of several experiments using cells from three separate human donors. Finally, the electrical resistance of monolayers begins to decline over the remainder of the time course to values near the initial immature monolayer. We irradiated HCAEC monolayers with a dose of 5 Gy because this dose constituted the threshold of detection for cytoskeletal rearrangement in endothelial cells in a survey of irradiated endothelia in the literature (27). Higher doses of 10 and 20 Gy generated more robust rearrangements. When competent, mature monolayers are irradiated with 5 Gy, a small initial rise in the resistance values precedes a statistically significant (P = 0.001) drop in irradiated cells that appears at 3 h postirradiation (Fig. 1, inset).
Impedance-Based Structural Modeling of Coronary Endothelial Monolayers
A notable advantage of ECIS is the ability to electronically dissect the contribution of cellular shape to the resultant impedance. Multifrequency transmonolayer impedance measurements can be input into a model of monolayer structure that makes use of 0 and first-order Bessel transforms to dissect paracellular and cell-substrate interactions (30). Transmonolayer permeability in endothelial cells is known to be associated with changes in the Rb parameter of this model, giving a readout of paracellular conductance changes (35). We found that the Rb parameter was significantly altered after irradiation with 5 Gy and, more importantly, that the α parameter, associated with cell-substrate interaction, was not altered as a result of this treatment (Fig. 1C).
Radiation-Induced Cytoskeletal Changes
Cytoskeletal rearrangement is known to occur in endothelial cells exposed to 15–20 Gy (27). We conducted a survey of the cytoskeleton in HCAECs at various times after irradiation, including 3 h, at which the impedance measurements suggested that such a rearrangement might take place. In comparison with sham-irradiated controls, phaolloidin-stained preparations of irradiated HCAECs showed cells with a slightly more polygonal shape with bundling of actin fibers at the periphery of the cells and a notable retraction from neighboring cells (Fig. 2A). In a separate experiment, HCAEC monolayers were followed over 6 h; the prominence of this phenotype was greatest in 3-h samples before resolving thereafter (Fig. 2B). Taken together, these findings reveal a transient rearrangement of the cytoskeleton in response to radiation that coincides with a loss in barrier function as assessed by ECIS.
Radiation-Induced Changes in Monolayer Permeability
To further confirm that cytoskeletal rearrangements were coincident with changes in monolayer permeability as predicted by the impedance measurements, the flux of a neutral, fluorescent tracer dye across the monolayer was measured. Cells were seeded on collagen-coated, porous substrates and allowed to mature into monolayers that were irradiated and monitored over 4 h. Fluorescent dye concentrations increased in the upper compartment over the time course, with vacant inserts always permitting greater amounts of dye across the substrate. While little to no difference in dye concentration in the upper compartment of the culture system was seen at 2 h, a statistically significant increase in permeability to the dye was seen at 3 h postirradiation (Fig. 3A, n = 3; P = 0.05).
Time-Lapse Live Imaging Experiments
In all studies on collagen-coated glass substrates, we observed intermittent cell loss from the monolayer. This was initially attributed to disturbance during washing, fixation and staining. Based on results from time-lapse live imaging experiments, we believe these cells to be representative of a minority population that intermittently detaches in the course of culture (Supplementary Information; http://dx.doi.org/10.1667/ RR2665.1.S1 (10.1667_RR2665.1.S1.pdf),, , ). Also evident in videos was the extrusion of dead cells from the monolayer with concomitant closure of the monolayer by surrounding cells. This was a satisfying result that corroborated the finding of intermittent extrusion of cells seen in immunofluorescence microscopy studies. This intermittent extrusion of endothelial cells comprising turnover in the monolayer has been observed previously in primary culture (37). The accrual of these dead cells was seen to increase gradually over time in both irradiated monolayers and sham controls. In some instances, extruded cells were able to spread and reintegrate into the underlying monolayer again. The increase in accrued, extruded cells was greater in irradiated cells, with the onset of accrual beginning at 1.5 h and stabilizing at 3.5 h after irradiation (Fig. 3B). Other studies of irradiated endothelial cells from the lung showed a marked depolymerization of the actin cytoskeleton with a resultant pulmonary edema (38–40). In contrast, we found that coronary endothelial monolayers remained ostensibly intact when assessed with time-lapse light microscopy with no obvious retraction of cells from one another or visible gaps appearing between cells. Together, the impedance findings, microscopy and videos suggest that the transient loss in transmonolayer resistance is an amalgam of cytoskeletal rearrangement and increased cell extrusion from the monolayer.
Radiation-Induced Changes in Cell Migration
We sought to further characterize cytoskeletal dynamics in HCAEC monolayers by conducting a wounding experiment on single-point ECIS arrays with a single 250-μm-diameter working electrode from which the interrogating electrical current originates, then passes through the monolayer and finally reaches the surrounding counterelectrode (Fig. 1A, lower portion). Again, monolayers were seen to mature when measured electrically. Given that the measurements were made at a single 0.05-mm2 point instead of at 40 regularly spaced points over a 0.8-cm2 area, the maturation was reported by a more irregular trace (Fig. 4A) compared with the resistance data presented in Fig. 1. Mature monolayers were irradiated with 5 Gy of γ rays as before, but after 30 min, monolayers were electrically removed from the central working electrode by driving 1400 μA at 60 kHz through the circuit for 20 s. This bolus of energy killed and removed cells from the electrode. The kinetics of wound closure over the electrode was reported electrically. Irradiated cells were seen to lag behind sham-irradiated controls in wound closure, and, when assessed 3 h after irradiation, the difference between the two conditions was statistically significant (Fig. 4, inset, n = 3; P = 0.03).
After completion of the wound closure process, monolayer resistance was seen to return to values similar to preirradiation levels, and the recorded resistance values diminished slowly over the remainder of the time course as seen before. In this period, micromotion of cells transiting the 250-μm-diameter working electrode was recorded by examining fluctuation in the resistance of the monolayers. Ohmic resistance values acquired at 5-min intervals over a 5-h period spanning 10–15 h after irradiation were used to create a micromotion index (mmi) as described in the Materials and Methods. Given our suspicion that cytoskeletal dynamics were somehow compromised by radiation, we assessed the significance of the difference in calculated mmi values and found them to be statistically significant in a single-tailed Student's t test (Fig. 4B; n = 3; P = 0.02). Thus cytoskeletal dynamics assessed by dynamic motion in the monolayer is ablated by radiation, and wound closure is delayed as well.
Several studies have shown that endothelial function is compromised after doses of radiation over 15 Gy. These studies include demonstrations of perturbed barrier function, compromised transduction of relaxation signals, and altered prostanoid biology (9–14). Since the relative sensitivities of cardiac substructures to radiation damage have not been assessed the details of endothelial dysfunction in radiation-induced heart disease remain unclear, we studied the coronary endothelium. We investigated the changes in coronary endothelial function using ECIS. This is a unique, highly sensitive technology that allows us to monitor changes in live cell monolayers for days after irradiation.
The transmonolayer resistance measurements in our studies are made at 4 kHz and represent only the ohmic component of impedance for simplicity. We chose 4 kHz because this frequency exhibits the greatest response seen in a multifrequency spectrum for our monolayers in comparisons of early and mature monolayers. This frequency also permits comparison of our findings to a considerable number of ECIS-related publications dealing with endothelial biology where the same frequency is employed (41, 42). After irradiation, we observed changes in the transmonolayer impedance of HCAEC monolayers. These changes represent a compound response comprised of cytoskeletal rearrangements as seen in immunofluorescence surveys, increased monolayer permeability reported both electrically and in Lucifer Yellow flux studies, and increased extrusion of cells from the monolayer as seen in time-lapse imaging studies.
The significant and unique aspects of this work are the kinetics and morphometry of radiation-induced cytoskeletal rearrangement in an intact monolayer of HCAECs in response to a dose of γ radiation. Given our sensitive methodology, this rearrangement becomes discernible at much lower doses than those already used to investigate this response in other endothelia. We found that, 3 h after a 5-Gy dose, a transient loss of transmonolayer resistance occurred in irradiated samples and that this resolved over a subsequent 1–2-h period. This drop in resistance was coincident with the rearrangement of the actin cytoskeleton in similar preparations when the status of the cytoskeleton was analyzed by confocal microscopy. We also found that HCAECs exhibited delayed wound closure when wounded 30 min after irradiation and that the micromotion of these cells over a 250-μm-diameter electrode was reduced when a 5-h segment of the recording was assessed for this parameter. Together, these findings indicate that HCAECs have an altered cytoskeletal biology after exposure to ionizing radiation. Since reduction in transendothelial resistance is generally associated with increased permeability (34, 35), we believe that the more immediate and transient loss in resistance we observed is conducive to fibrosis and atherogenesis in the coronoary circulation.
The limitations of the study include its restriction to a two-dimensional analysis format and the induction of stress in the primary cells we used as a result of subculture outside their normal environment. The commercially available HCAECs halted division and senesced 10 to 14 days after procurement and culture. These stresses could have an impact the nature of the response we observed.
The strengths of the study include the use of ECIS technology, which made lengthy recordings of the monolayers possible. The modeling component of the instrumentation allows for angstrom-scale measurements of the compartment between the cells and the collagen-coated substrate (30). The ECIS method also permits a unique measurement of the micromotion parameter, providing information on the motility of resident cells in the endothelial monolayers and, indirectly, the ability of the cells to manipulate their cytoskeleton in the process of locomotion. Our use of primary HCAECs allowed a particularly useful insight into this important cardiac substructure.
Our findings in coronary endothelium differ from other studies in which cells were obtained from other tissues. Pulmonary endothelium is known to undergo a depolymerization of the actin cytoskeleton, and this results in a physiologically significant increase in the wet weight of the lungs as edema accrues (39, 40). Gabrys et al. (27) demonstrated that a Rho-mediated, radiation-induced cytoskeletal rearrangement occurred in certain endothelial cells and that only in DMVECs was this process independent of p38. Others found that both p38 and JNK were activated in irradiated human umbilical vein endothelial cells (HUVECs) (23). Endothelial cells will undoubtedly have different characteristics based on the tissue from which they are obtained. DMVECs will have a role in thermoregulation, for example, which is not relevant for HUVECs. Endothelial cells from bovine adrenal glands (BAECs) are exposed to a very different endocrine milieu compared to the endothelium in the blood-brain barrier. Similarly, coronary endothelium resides in a physiologically unique location that has one of the highest oxygen consumption sites in the body. In a comparison between various endothelial subtypes plated on laminin, only HCAECs did not exhibit a statistically significant increase in proliferation (43). It is not unreasonable, therefore, to expect a different radiation response from this endothelium given the increased turbulence, shear and incessant flexion of the adjacent epicardium.
The kinetics of our observed radiation-induced cytoskeletal rearrangements differs from the effects observed in other studies of endothelial tissue ex vivo. Immediate and early responses to radiation have been observed with kinetics of less than 30 min. Such rapid responses include conversion of sphingomyelin to ceramide and the induction of endothelin transcripts (44, 45). The differences in kinetics could arise from the endothelial cell type as discussed above or from a difference in the extracellular matrix (ECM).
Several different radiation biology studies indicate that the ECM is an important factor in the biology of endothelial cells. A number of studies of endothelia indicate that there is an interplay between radiation-induced damage and survival signaling from matrix through ECM receptors (integrins) along the PI3K-Akt axis (43, 46–48). All of the work conducted in this study involved HCAECs plated on substrates coated with collagen I, so the migration and signaling we observed was mediated primarily by α1β1 and α2β1 integrins. Adhesion to and migration on denatured collagen (gelatin) or vitronectin is mediated by αvβ3, which is upregulated in endothelium that is localized near tumors or wound tissue (49, 50). The surface expression of integrins in HCAECs was not measured here, but given that HCAECs are dissociated from their resident tissue and are in culture ex vivo, it is likely that αvβ3 is expressed on the surface of these cells. The magnitude or kinetics of our observed response may vary with the composition of the ECM on which the monolayer is seeded.
Endothelial barrier function is compromised after irradiation, and this can be attributed to paracellular or transcellular communication between the blood and tissue. More definitive results have been seen in the endothelial barrier in the brain, where the paracellular pathway has been found to be modulated after a single dose of 20 Gy (51). In the heart and in irradiated tumor vasculature, findings support transcellular communication between tissue and blood after fractionated radiotherapy or a single dose of 10–13 Gy (52, 53). Paracellular communication would seem to be more relevant at doses where overt killing of the endothelium is observed (54–56), and the resultant access to ECM in the context of dysfunctional endothelium could account for the accrual of von Willebrand factor by means of collagen binding domains of this platelet aggregating factor (57). Our findings suggest that lower doses where endothelial cells are not killed could still induce transient communication between the blood and myocardium.
Finally, it is important to consider that the coronary endothelium is a component of a complete cardiopulmonary system. The ultimate detrimental effect of radiation-induced heart disease is reduction of cardiac function. Irradiation of the thorax for diagnostic or therapeutic reasons includes exposure of the lungs, myocardium, pericardium, great vessels and coronary circulation. Radiation-induced effects on any of these components can cause feedback into others in a decompensating feedback loop, resulting in failure (58). We anticipate that future studies will further detail the role of endothelial dysfunction in radiation-induced heart disease and that additional sub-compartments of the heart will be assessed for their respective radiation sensitivities.
The authors are deeply indebted to I. Giaver, C. Keese and Applied Biophysics for access to an ECIS instrument, without which this work would not have been possible. The authors are also thankful to S. Amundsen and S. Ghandhi for helpful discussions and to T. Templin for assistance with the assembly of the manuscript. This work was supported with funding from the Department of Energy (DEPS0208ER0820) and NASA (NNX07AT41G).
 Note. The online version of this article (DOI: 10.1667/RR2665.1) contains supplementary information that is available to all authorized users.