Low-dose irradiation causes rapid alterations to the proteome of the human endothelial cell line EA.hy926
- First Online:
- Cite this article as:
- Pluder, F., Barjaktarovic, Z., Azimzadeh, O. et al. Radiat Environ Biophys (2011) 50: 155. doi:10.1007/s00411-010-0342-9
- 248 Views
High doses of ionising radiation damage the heart by an as yet unknown mechanism. A concern for radiological protection is the recent epidemiological data indicating that doses as low as 100–500 mGy may induce cardiac damage. The aim of this study was to identify potential molecular targets and/or mechanisms involved in the pathogenesis of low-dose radiation-induced cardiovascular disease. The vascular endothelium plays a pivotal role in the regulation of cardiac function and is therefore a potential target tissue. We report here that low-dose radiation induced rapid and time-dependent changes in the cytoplasmic proteome of the human endothelial cell line EA.hy926. The proteomes were investigated at 4 and 24 h after irradiation at two different dose rates (Co-60 gamma ray total dose 200 mGy; 20 mGy/min and 190 mGy/min) using 2D-DIGE technology. Differentially expressed proteins were identified, after in-gel trypsin digestion, by MALDI-TOF/TOF tandem mass spectrometry, and peptide mass fingerprint analyses. We identified 15 significantly differentially expressed proteins, of which 10 were up-regulated and 5 down-regulated, with more than ± 1.5-fold difference compared with unexposed cells. Pathways influenced by the low-dose exposures included the Ran and RhoA pathways, fatty acid metabolism and stress response.
Dulbecco’s Modified Eagle’s Medium
Hypoxanthine Aminopterin Thymidine
Single nucleotide polymorphism
High doses of ionising radiation, such as the tens of Grays used in radiotherapy, are known to increase the risk of cardiovascular disease (CVD). Observed effects of high-dose radiation on the heart include direct damage to the coronary arteries, diffuse fibrotic damage to the pericardium and myocardium, pericardial adhesions, stenosis of the valves and microvascular damage (Demirci et al. 2009; Adams et al. 2003). On the other hand, there are considerable mechanistic and epidemiological uncertainties surrounding the possible cardiovascular damaging effects of lower doses of ionising radiation (<0.5 Gy) (Little et al. 2008). It is necessary to explore potential biological and physiological effects of the low-dose radiation due to the increasing use of new radiation therapy protocols and imaging procedures applying such low radiation doses to the heart.
The data concerning the CVD risk from occupational low-dose exposures are controversial. Radiation workers in the Chernobyl liquidator cohort show an increased risk for ischaemic heart disease (Ivanov et al. 2006). Among employees at British Nuclear Fuels, as well as in Canadian nuclear workers and other occupationally radiation-exposed groups, there is also evidence for an increasing trend in circulatory disease mortality associated with the radiation dose (McGeoghegan et al. 2008; Ashmore et al. 1998). In contrast, no statistically significant increase in circulatory disease mortality could be attributed to inhaled radon and its progenies or external γ-irradiation among the German uranium miners (Kreuzer et al. 2006).
An excess radiation-associated risk for CVD has also been observed in the Life Span Study of the Japanese atomic bomb survivors. Importantly, even at doses as low as 0.5 Gy, the mortality and morbidity due to hypertension and myocardial infarction were increased (Preston et al. 2003; Yamada et al. 2004).
The vascular endothelium forms a continuous cellular monolayer lining the interior surface of blood vessels and serves as the barrier between circulating blood and the subendothelial matrix (Marsden et al. 1991). It plays an important role in the integration and modulation of many functions of the arterial wall (Luscher et al. 1990; Furchgott and Zawadzki 1980). Previous data show that increased production of reactive oxygen species (ROS) in the endothelium is associated with vascular endothelial dysfunction during the human ageing process (Herrera et al. 2009). Increased ROS production leads to oxidation of low density lipoproteins (LDL), accumulation of lipids into foam cells, growth of vascular wall intima layer and finally atherosclerotic plaque expansion and rupture (Ross 1999; Falk and Fernandez-Ortiz 1995).
Single-dose or fractionated irradiation of 14 Gy has been shown to accelerate the development of atherosclerosis and predispose to the formation of an inflammatory, thrombotic plaque phenotype in ApoE mutant mice, but not in age-matched controls (Hoving et al. 2008). Conversely, acute radiation doses in the range 0.1–1 Gy resulted in the down-regulation of the adhesion of leucocytes to the endothelium both in vitro and in vivo and thus had an anti-inflammatory effect (Little et al. 2008; Rodel et al. 2004). Furthermore, it may be that not only the total dose, but also the dose rate has implications for cardiovascular disease (Amundson et al. 2003), as it does both for acute deterministic damage (Mitchel et al. 2007; Okudaira et al. 2010) and for late stochastic effects such as cancer (Jacob et al. 2009). It is therefore possible that the biological responses of the endothelium to high and low doses of ionising radiation may be different and guided by different molecular mechanisms and that these responses are dose-rate-dependent.
The cardiovascular damage due to high-dose radiation can be visualised histopathologically (Stewart et al. 2006). The cardiac vasculature is a strong target candidate even for the low-dose initial damaging events, as endothelial cells exhibit one of the highest levels of cellular radiation sensitivity. In this study, we have used the human endothelial cell line EA.hy926, as it has been shown to be a good endothelial model for the proteomic analysis of the effects of low-dose non-ionising radiation (Nylund and Leszczynski 2004, 2006). EA.hy926 retains a primary endothelial cell-like transcriptome that also responds typically to statins that are widely used to reduce the risk of cardiovascular disease (Boerma et al. 2006). Moreover, as low radiation doses do not trigger cellular growth arrest or apoptosis in EA.hy926 as shown here, it is a valuable model for studying the immediate effects on the proteome.
To detect the small protein expression changes that are to be expected at low radiation doses we applied, we employed strategies to increase the sensitivity of the proteome analysis: (1) we focussed only on changes in the cytosolic proteome, (2) we used two overlapping pH ranges in a large-size gel format and (3) we applied 2D fluorescence difference gel electrophoresis (2D-DIGE) technology with an internal standard.
Materials and methods
The cell line EA.hy926 was originally derived from the PEG-mediated fusion of a thioguanine-resistant clone of human alveolar type II-like epithelial (A549) cells with primary human umbilical vein endothelial cells (HUVEC) and is maintained in HAT medium (Edgell et al. 1983). The cell line was tested for markers of origin (PCR of eight highly polymorphic microsatellite short tandem repeat loci) by the German Collection of Microorganisms and Cell Cultures (DSMZ) and for mycoplasma contamination with synthesised TaKaRa primers by nested PCR.
The cells were cultured in T75 culture flasks at 37°C with 5% CO2 in air. The culture medium (D-MEM supplemented with 10% foetal bovine serum and 1× HAT) was renewed every 3 days as described by Nylund et al. (Nylund and Leszczynski 2006).
Irradiation of cells
Cells were irradiated with a single 200 mGy dose using a Co-60 gamma-irradiator (37000 GBq Co-60 therapy source). Radiation was applied at two different dose rates, namely 20 and 190 mGy/min. Irradiations were performed at 37°C using a thermostatically controlled water bath in a 30 cm × 30 cm × 30 cm water-filled PMMA phantom. The dose applied to the cell samples was calculated as the absorbed dose to water (Bjerke et al. 1998), based on dose measurements using a cylindrical radiotherapy chamber. Sham-irradiated samples were treated exactly as irradiated samples, except that the Co-60 source was shielded.
Measurement of radiation-induced apoptosis
For the detection of apoptotic cells, exponentially growing cells were irradiated with 0, 0.2 and 5.0 Gy as described above. A total of 5.0-Gy irradiated cells served as a positive control for apoptosis detection.
To determine caspase-3 activity, cells were washed with PBS and collected by trypsinisation 24 h after gamma irradiation. The cells were incubated for 20 min at room temperature with a final concentration of 10 μM caspase inhibitor FITC-VAD-FMK (Promega, Madison, WI, USA) that binds selectively to active caspases. The cells were washed twice with PBS and analysed by FACS analysis (LSR II, Becton–Dickinson/FACS DIVA Software).
The subdiploid (subG1) DNA peak was analysed 48 h after irradiation. Cells containing less DNA than cells in the G1 phase of the cell cycle were considered to be apoptotic. Cells were harvested from plates by trypsinisation and washed twice with PBS. Pellet was gently resuspended in 300 μl of solution I containing 10 mM NaCl, 4 mM Na-citrate, 0.3% Nonidet P-40 (Roche, Germany), 10 μg/ml RNAase (Sigma, Germany) and 50 μg/ml propidiumiodide (PI; Sigma, Germay) and vortexed. The cell suspensions were incubated for 1 h at room temperature, followed by the addition of 300 μl of solution II containing 70 mM citric acid, 250 mM sucrose and 50 μg/ml PI. The cell suspensions were mixed and stored at 4°C before flow cytometric measurement. Cell cycle distributions were analysed on a FACScan LSR II (Becton–Dickinson) (excitation wavelength: 488 nm; emission wavelength: >610 nm, LSR II, Becton–Dickinson). The percentage of apoptotic nuclei (subdiploid DNA peak in the DNA fluorescence histogram) was calculated using WinMDI software v. 2.9 (Joe Trotter).
The chosen proteomic strategy was based on two-dimensional gel electrophoresis enhanced by the use of the DIGE technology (Unlu et al. 1997). Four experimental groups consisting of two time points after exposure (4 and 24 h) at each of the two dose rates (20 and 190 mGy/min) were analysed.
For each dose rate and time point, two independent biological samples with three technical replicates were used for statistical analysis.
Subfractionation, labelling and 2D-DIGE
Subfractionation of the cells was performed by differential detergent fractionation (Ramsby et al. 1994). All steps were carried out at 4°C. Briefly, the cell pellet was dissolved gently in lysis buffer [3 mM imidazol pH 6.8, 250 mM sucrose, 3 mM MgCl2, 5 mM EDTA, 1 mM tributylphosphine, with added protease (Roche Complete) and phosphatase inhibitors (Cocktail I and II, Sigma–Aldrich)], transferred to microcentrifuge tubes and resedimented for 5 min at 300×g. The plasma membranes were disrupted using 0.015% digitonin in lysis buffer, with gentle shaking for 10 min. The resultant lysate was centrifuged at 480×g for 10 min to remove organelle and membrane debris. The cytosolic supernatant was transferred to a new tube whilst the pellet was washed with 500 μl lysis buffer without digitonin, recentrifuged and both supernatants were pooled. The protein concentration was determined in triplicate by the Bradford assay using ovalbumin as standard (Sigma–Aldrich). The cytosolic proteins were precipitated with trichloracetic acid (TCA) in acetone for the subsequent CyDye-labelling. Two volumes of 20% TCA and 20 mM DTT in acetone chilled to −20°C were added. The proteins were precipitated at −20°C for 2 h and collected by centrifugation at 20,000×g, 4°C for 30 min. The pellet was washed twice with 20 mM DTT in acetone, and the supernatant was discarded.
Cytosolic fractions from sham-irradiated and irradiated samples were labelled with CyDye™ DIGE Fluor minimal dyes (200 pmol/50 μg). The irradiated samples were labelled with Cy3, the controls with Cy5 and the internal standard (a mixture of both) with Cy2. Samples were mixed and separated by 2-DE IEF was performed using an IPGphor III IEF system (GE Healthcare) with IPG strips pH 4–7 and 6–10 (24 cm, Serva). IPG strips were rehydrated o/n for the pH range 4–7 in a buffer containing 7 M urea, 2 M thiourea, 2% CHAPS, 10% glycerol, 0.2% DTT, 0.5% IPG buffer pH 4–7 and 0.5% IPG buffer pH 6–10. For the pH range 6–10, the rehydration of the strip was done in 7 M urea, 2 M thiourea, 2% CHAPS, 5% glycerol, 10% isopropanol, 0.5% DTT, 0.5% IPG buffer pH 4–7 and 0.5% IPG buffer pH 6–10. The amount of 150 μg (50 μg each) of combined samples were applied via cup loading on the anodic side. IEF was run at 20°C using voltages and running times as follows: 3 h at 500 V (step), 5 h to 1,000 V (gradient), 6 h to 10,000 V (gradient), to a total of 90 kVh at 10,000 V. During the IEF of pH 6–10 strips, the cathodic electrode wick was soaked in 0.4% DTT and exchanged after 8–10 h. For the 2nd dimension, the strips were equilibrated for 15 min in 1% DTT, 2% SDS, 50 mM Tris pH 8.8, 6 M urea and 30% glycerol, followed by incubation for 15 min in the same buffer with the 1% DTT replaced by 2% iodoacetamide. The strips were embedded on the top of the 12% gels (20 × 24 cm) using 0.5% agarose in Laemmli buffer. The second dimension was run using an Ettan DALTtwelvesystem (GE Healthcare) with Laemmli buffer (Laemmli 1970) using taurine as the trailing ion (Tastet et al. 2003). Electrophoresis was performed at 20°C 1 h at 5 mA/gel, 1 h at 8 mA/gel and then at 20 mA/gel until the bromophenol blue reached the end of the gel.
Gels were scanned at a resolution of 100 μm on a Typhoon 9400 (GE Healthcare). Cy2-, Cy3- and Cy5-Dye images of each gel were acquired at excitation/emission values of 488/520, 523/580 and 633/670 nm, respectively. After image acquisition, the gels were fixed overnight in 30% ethanol and 10% acetic acid and stored at 4°C. For mass spectrometry-based identification, the gels were post-stained with silver according to Heukeshoven and Dernick (1985).
Scanned images were cropped in ImageQuant software version 5.2 (GE Healthcare). The DeCyder version 6.5 software (GE Healthcare) was used for image analysis. The differential in-gel analysis (DIA) module was used for automatic spot detection. Abundance measurements for each individual gel were obtained by comparing the normalised volume ratio of each spot from a Cy3- or Cy5-labelled sample to the corresponding Cy2-signal from the pooled-sample internal standard. The DIA datasets from each individual gel were collectively analysed using the biological variation analysis (BVA) module, which allows inter-gel matching and the calculation of the average abundance for each protein spot among the six gels of each group. Statistical significance was assessed for each change in abundance using Student’s t-test and two-way-ANOVA analyses. We considered spots as differentially regulated if statistical significance at the 95% confidence level was achieved and if the standardised average spot volume ratio exceeded 1.50-fold. Calculation of experimental MW and pI for each differential protein spot was done with the help of the given pH range of the IPG-strips and with an externally applied molecular weight marker proteins. False discovery rate (FDR) correction was applied in the statistics.
Post-staining and MALDI identification
Proteins with statistically significant differential expression (Student’s t-test p < 0.05 and > ±1.50-fold) were manually picked from 2D-DIGE gels post-stained with silver. The gel spots were transferred to protein low-bind tubes and destained with 15 mM K3Fe(CN)6 and 50 mM Na2S2O3. The gel pieces were washed once with 500 μl water and afterwards twice with 200 mM NH4HCO3, each for 15 min. The liquid was removed and the gel pieces were shrunk three times with 25 μl acetonitrile (ACN) for 5 min. Subsequently, the gel pieces were dried and the samples were rehydrated with 10 μl of trypsin (Promega, 10 ng/μl in 50 mM NH4HCO3). After 10 min, the gel pieces were covered with 10–30 μl of 50 mM NH4HCO3 and digested o/n at 37°C. The resulting peptide mixture was extracted twice with 50 μl of 50% ACN, 2.5% TFA by sonication for 10 min. A third extraction was done by the addition of 10 μl of ACN and a 5-min sonication step. The supernatants were collected in a fresh protein low bind tube, frozen in liquid nitrogen and reduced to a volume of 10–20 μl in a speedvac. The peptides were bound to C18ZipTips (Millipore) according to the manufacturer’s instructions and eluted with 50% ACN, 0.1% TFA. Then, 0.5 μl of sample was spotted onto a stainless steel MALDI target plate by the dried droplet method. The matrix used was 3.75 mg/ml α-cyano-4-hydroxycinnamic acid in 60% ACN, 0.1% TFA.
Mass spectra were acquired using a 4700 Proteomics Analyser (MALDI-TOF-TOF) (Applied Biosystems). Measurements were performed with a 355-nm Nb:YAG laser in positive reflector mode with a 20-kV acceleration voltage. The mass range (m/z 900–4,000) was externally calibrated using the peptide calibration standard III (Applied Biosystems). For each MS and MS/MS spectrum, 3000 laser shots were accumulated. Tandem mass spectrometry was performed by CID with air as the collision gas. Precursor masses were selected in a data-dependent manner using the 8 most abundant ions (excluding masses of trypsin autolytic products and common keratin peptides). Spectra acquisition and processing was done in automatic mode with 4000 Series Explorer software (version 3.6, Applied Biosystems).
Criteria for protein identification
The GPS Explorer™ Software (version 3.6, Applied Biosystems) was used for spectra analyses. The database search was performed with MASCOT (Version: 2.2.06) using the human/mammalia Uniref 100 [UniRef100 version from 20090718 (8663575 sequences; 3067231997 residues)] and SwissProt databases [SwissProt version from 20090708 (495880 sequences; 174780353 residues)] with human/mammalia as taxonomy. One missed trypsin cleavage was selected. Carbamidomethylation was set as the fixed modification and oxidised methionine as the variable modification. Precursor tolerance was set to 75 ppm and MS/MS fragment tolerance to 0.3 Dalton. The shown MOWSE protein scores are a summary of scores for each MS/MS spectra and an additional score for the peptide mass fingerprint. The significance level (p value < 0.05) for a protein score is usually higher than a MOWSE score of 50–60 (for an analysis of this dataset against Swiss-Prot, this corresponds to a MOWSE score >56). MS/MS spectra of protein identifications with scores nearby the significance level have always been manually checked.
Western blot analysis
For the validation of protein expression changes by Western blotting (Burnette 1981), 20 μg of proteins from the cytosolic fractions from novel biological samples treated similarly to those used for the proteome analysis were separated on 8 and 12% SDS polyacrylamide gels according to Laemmli (1970). Proteins were transferred to nitrocellulose membranes (GE Healthcare, Germany) using a semidry blotting system at 100 mA for 90 min. Membranes were saturated for 1 h with 5% advance blocking reagent (GE Healthcare) in Tris buffered saline (50 mM Tris–HCl, pH 7.6 and 150 mM NaCl) containing 0.1% Tween 20 (TBS/T). Blots were then incubated o/n at 4°C with antibodies against either cofilin, serine/threonine-protein phosphatase 2A, transitional endoplasmic reticulum ATPase (all from New England Biolabs), Ran-specific GTPase-activating protein, GTP-binding protein SAR1a, flavin reductase (all from Abcam UK), Rho GDP-dissociation inhibitor 1 (Invitrogen, Germany) or α-tubulin as loading control (Sigma–Aldrich). After washing three times in TBS/T, blots were incubated for 1 h at room temperature with horseradish peroxidase-conjugated anti-mouse, anti-rabbit or anti-goat secondary antibody (Santa Cruz Biotechnology) in blocking buffer (TBS/T with 5% w/v advance blocking reagent). Immunodetection was performed with ECL advance Western blotting detection kit (GE Healthcare). The protein bands were quantified using ImageQuant 5.2 software (GE Healthcare) by integration of all the pixel values in the band area after background correction, normalised to the α-tubulin expression.
Analysis of protein–protein interaction and signalling network
Analysis of protein–protein interaction and signalling networks was performed by the search tool Ingenuity Pathway Analysis (Ingenuity System, http://www.ingenuity.com). After inserting the accession numbers of the differentially up- and down-regulated proteins, Ingenuity displayed all predicted associations in a summary view. The nodes represent proteins that are connected with one or several arrows; the identified proteins identified in grey. The solid arrows represent direct interactions and the dotted arrows indirect interactions. We searched for direct and predicted protein interactions.
The aim of the study was to analyse time- and dose-rate-dependent variations in the proteome of the human endothelial cell line EA.hy926 after exposure to low-dose irradiation. There are only a few studies describing the effect of dose rate on cellular gene expression (Sokolov et al. 2006; Sugihara et al. 2008; Taki et al. 2009) and even fewer investigations at the proteome level (Nakajima et al. 2008).
Effect on apoptosis and cellular growth
General proteomic alterations
Proteomic analysis was performed by comparing the expression patterns of the cytosolic fractions of sham-irradiated and 200 mGy low-dose gamma-irradiated EA.hy926 human endothelial cells at two different time points after irradiation (4 h and 24 h). The number of protein spots detected with both pH ranges using DIGE labelling averaged 1,300. Compared to the sham-irradiated EA.hy926 cells, Decyder image analysis showed a total of 33 (pH 4–7) and 25 (pH 6–10) protein spots that were significantly altered (change >1.50-fold, p < 0.05). Most of the differentially expressed protein spots showed low intensities, indicating their non-abundant character. It was not possible to identify the lowest abundance spots using MALDI-TOF/TOF mass spectrometry, even after pooling material from several preparative gels.
Proteins identified in this study
Gene name ExPASy
MW kDa (obs./calc.)
No. peptides matched
Sequence coverage (%)
Rho GDP-dissociation inhibitor 1a,b
GTP-binding nuclear protein Ran
Ran-specific GTPase-activating proteina
Eukaryotic translation initiation factor 5A-1
Acetyl-CoA acetyltransferase, mitochondrial
Transitional endoplasmic reticulum ATPasea
GTP-binding protein SAR1aa
Heat shock protein HSP 90-alpha
Peptidyl-prolyl cis–trans isomerase A
Serine/threonine-protein phosphatase 2A 65 kDa regulatory subunit A alpha isoforma
SH3 domain-binding protein 2
Protein mago nashi homolog
The biological processes and average ratios of differentially expressed proteins with the two-way-ANOVA
Average ratio 4 h 20 mGy/min
Average ratio 4 h 190 mGy/min
Average ratio 24 h 20 mGy/min
Average ratio 24 h 190 mGy/min
Rho GDP-dissociation inhibitor 1
Cellular adhesion and migration
Cellular adhesion and migration
GTP-binding nuclear protein Ran
Ran-specific GTPase-activating protein
Eukaryotic translation initiation factor 5A-1
Acetyl-CoA acetyltransferase, mitochondrial
Fatty acid metabolism
Transitional endoplasmic reticulum ATPase
GTP-binding protein SAR1a
Heat shock protein HSP 90-alpha
Protein folding, stress response
Peptidyl-prolyl cis–trans isomerase A
Serine/threonine-protein phosphatase 2A 65-kDa regulatory subunit A alpha isoform
Cell growth and division
Methionine salvage pathway
SH3 domain-binding protein 2
Protein mago nashi homolog
Five out of fifteen differentially expressed proteins (CLF-1, RANBP1, EIF5A, SH3BP2, MAGOH) were significantly up- or down-regulated at more than one time and dose-rate point (Table 2). From the proteins found significantly up- or down-regulated at only one time or dose-rate point (10 out of 15), five were confirmed with Western blot analysis (see “Proteomic analysis”). The two-way-ANOVA showed that four proteins were regulated in a time- or interaction-dependent manner (Table 2).
Proteomic analysis data of differentially regulated proteins identified in this study (modifications, sequence coverage, matched peptides) are shown in Supplementary Table 1.
Time-dependent proteomic alterations
Independently of the dose rate, more alteration in the protein expression was seen after 4 h compared with 24 h suggesting that most of the proteome alterations are rapid and transient. The largest number of differentially expressed proteins (10) was found 4 h after the irradiation at the higher dose rate (190 mGy/min). Similarly, the lower dose rate (20 mGy/min) induced more alterations (5) in the protein expression pattern after 4 h than after 24 h (three proteins). Proteins showing differential expression only after 4 h include ARHGDIA, RAN, VCP, HS90AA1, PPP2R1A, SAR1A and ADI1. Protein expression levels of ACAT1, BLVRB and PPIA were significantly changed only after 24 h. It seems that many biological pathways show immediate response (after 4 h) to low-dose radiation, but this response is no longer seen after 24 h. The processes involved include ER transport, heat shock response, cell growth and division and methionine salvage pathways (Table 2).
Dose-rate dependent proteomic alterations
Although the difference in the dose rates used was only about tenfold, it could be clearly seen that the higher dose rate (190 mGy/min) affected the proteome alteration pattern much more effectively than the lower dose rate (20 mGy/min). Eight proteins (ARHGDIA, RAN, ACAT1, BLVRB, VCP, HS90AA1, PPIA, PP2R1A and MAGOH) vs. two proteins (SAR1A and ADI1) were significantly up- or down-regulated with only high or only low dose rate, respectively (Table 2).
Signalling network by bioinformatics analysis
To our knowledge, this is the first study investigating the rapid response of the cytoplasmic proteome of endothelial cells to a challenge with two different dose rates of low-dose ionising radiation. We show that an exposure as low as 200 mGy has an impact on the human endothelial cell proteome although it does not cause any effect on cellular growth or apoptosis. Not surprisingly, the protein expression changes were subtle, only one protein expressing higher than a 2.5-fold increase. The threshold of 1.5 for statistical significance typically used in 2D-DIGE experiments showed that five of the fifteen differentially up- or down-regulated proteins were consistently altered in expression over different time and dose-rate conditions. Five proteins, the expression of which was significantly changed only if exposed to one experimental condition, were analysed with Western blotting that confirmed the 2D-DIGE results. Therefore, in spite of the modest degree of expression alterations between the sham- and low-dose irradiated proteomes, the different methods used to validate the results give confidence that the study has identified proteins and corresponding pathways biologically relevant to the low-dose response of the endothelium.
Most of the protein amount alterations occurred 4 h after irradiation when the higher dose rate (190 mGy/min) was used. The classical stress response proteins, such as HSP 90-alpha and endoplasmic reticulum stress indicators, showed rapid and transient responses to the exposure detectable at 4 h, but were resolved after 24 h. In contrast, the altered expression seen for a marker of oxidative stress, flavin reductase, persisted for at least 24 h after exposure, suggesting a longer-lasting influence.
The role of the time dependence on the protein expression is more dominant than that of the dose rate (Table 2). This may be due to the fact that the difference in the dose rates used in this study is only about 10-fold. Nevertheless, although the proteome patterns arising after high- and low-dose rate irradiation are overlapping, they are not identical. The high dose rate induces differential expression of nine unique proteins that are not found responding to the low dose rate, whereas the low dose rate affects the expression of only two unique proteins.
Further analysis of the 15 identified deregulated proteins in our study showed that the proteins can be divided into a number of functionally related groups (Table 2): (1) proteins involved in the RhoA pathway that regulates cellular adhesion and migration; (2) proteins of the Ran signalling pathway involved in RNA transport between the nucleus and cytoplasm; (3) proteins involved in fatty acid metabolism and (4) endoplasmic reticulum and oxidative stress response proteins.
We found biological pathways distributed across the functional groupings, involving the action of small Ras-like GTPases, namely Ran, RhoA and Sar1a. The proteins share a significant sequence homology with the GDP/GTP-binding pocket being especially conserved (Neuwald et al. 2003). Many Ras-superfamily small GTPases are components of signalling pathways that link extracellular signals via transmembrane G-protein-coupled receptors to cytoplasmic or nuclear responses (Bhattacharya et al. 2004). It is known that not only RhoA and Ran pathways, but also the expression of serine/threonine-protein phosphatase 2A is regulated by the level of oxidative stress (Heo 2008; Heo and Campbell 2005; Foley et al. 2007) providing a possible link between these pathways and the oxidative stress response that we find affected.
The RhoA pathway proteins identified here are the Rho GDP-dissociation inhibitor 1 and cofilin. Rho GDP-dissociation inhibitor 1 is down-regulated, what might lead to an alteration in levels of the RhoA-GTP complex (active form). In addition, cofilin, a downstream target of RhoA, is up-regulated, suggesting that the RhoA/ROCK/Lim/cofilin pathway that regulates the actin cytoskeleton assembly and thereby cellular adhesion and migration (Kuzelova and Hrkal 2008) is affected after low-dose radiation. Previous studies suggest that activation of this pathway by inflammatory cytokines and agonists of G-protein-coupled receptors effectively triggers vascular disease (Seasholtz and Brown 2004). Indeed, ROCK (Rho kinase) inhibitors have been proven to have therapeutic potential for preventing and treating a wide range of cardiovascular diseases (Shimokawa and Takeshita 2005; Budzyn and Sobey 2007).
GTP-binding protein Ran is essential for the translocation of RNA and proteins through the nuclear pore complex. Among other functions, Ran mediates the transport of microRNAs that form an endogenous system for the regulate and coordinated expression of genes on a post-transcriptional level (Wiemer 2007; Bohnsack et al. 2004). Ran is also involved in chromatin condensation, control of the cell cycle and DNA synthesis. Ran-specific GTPase-activating protein (RANBP1) inhibits GTP exchange on Ran; its downregulation may thus increase the Ran turnover. The precise role of the eukaryotic translation initiation factor 5A-1 that is up-regulated in our study is not known, but it is known to form a complex with Ran.
Our study indicates furthermore alterations in the expression of proteins involved in the cellular energy metabolism. Acetyl-CoA acetyltransferase, the level of which is up-regulated is able to produce CoA using fatty acids as a source.
The endoplasmic reticulum (ER) is a finely tuned organelle, the main task of which is to guarantee the proper folding of proteins. It is composed of molecular chaperones catalyzing the folding, and sensors detecting misfolded or unfolded proteins (Malhotra and Kaufman 2007; Malhotra et al. 2008). Recent data suggest that oxidative stress is associated with the development of endoplasmic reticulum stress and a subsequent degradation of misfolded proteins (Guo et al. 2009). In agreement with this, our data indicate that the levels of proteins involved in either oxidative stress (flavin reductase), ER stress and transport (transitional endoplasmic reticulum ATPase, GTP-binding protein Sar1a) or protein folding and degradation (heat shock protein HSP 90-alpha, peptidyl-prolyl cis–trans isomerase A, proteasome subunit beta type-3) are all affected by low-dose ionising radiation.
Our study indicates that low-dose ionising radiation affects the cytoplasmic proteome of a human endothelial cell line in a subtle but significant manner, the biggest impact being observed immediately (4 h) after irradiation. The higher dose rate used here (190 mGy/min) induced far more proteome alterations than the low dose rate (20 mGy/min).
A common feature of many of the cytoplasmic proteins identified as low-dose radiation-responsive in our study is that the signalling pathways they represent are directly dependent on the redox potential and the level of reactive oxygen species. A large number of recent data suggest that altered levels of oxidative stress are essential in the development of cardiovascular disease and that endothelial mitochondria may play an important role both as a target and source of reactive oxygen species (Landar et al. 2006; Zhang and Gutterman 2007; Ballinger 2005; Zmijewski et al. 2005; Davidson and Duchen 2007). Therefore, it is our future aim to further investigate the role of mitochondria in the low-dose radiation response.
Although the cell line EA.hy926 displays a suitable model for endothelial functions, it originates from human umbilical vein endothelial cells that may differ from cardiac endothelial cells. The studies with primary human coronary artery cells exposed to low-dose ionising radiation in a similar manner as EA.hy926 in this study are ongoing. In addition, using EA.hy926, we will investigate post-translational modifications and include radiation-induced proteome alterations of nuclear proteins in the future studies.
Taken together, using 2D-DIGE proteomic approach and systematic bioinformatics analysis, this study has been able to identify biological pathways that enable a better understanding of a low-dose radiation response in a target tissue.
The research leading to these results is supported by a grant from the European Community’s Seventh Framework Programme (EURATOM) contract no. 211403 (CARDIORISK). We thank Dr. Ludwig Hieber and Dr. Herbert Braselmann for giving valuable advice in statistical questions.