Final Report Summary - CARDIORISK (The mechanisms of cardiovascular risks after low radiation doses)
Project context and objectives:
The aim of the CARDIORISK collaborative research project is to elucidate the pathogenesis of early and late alterations in the microcirculation of the heart and of atherosclerotic lesions in arteries after exposure to low radiation doses in comparison to high radiation doses. A major goal was investigation of early molecular, proinflammatory and prothrombotic changes as well as alterations of myocardial perfusion, cardiac cell integrity and immunological processes.
CARDIORISK aimed to achieve the following objectives:
1. elucidation of the pathogenic mechanisms of radiation-induced heart disease and of radiation-induced vascular damage after low and moderate radiation doses at the tissue, cell and molecular level (the primary objective of this project). In the context of this experimental study, low radiation doses are defined as average heart doses between 0.01 and 0.5 Gy, while moderate doses as average heart doses between 0.5 and 5 Gy
2. determination of the radiation dose dependence of the severity, latency, and rate of progression of cardiovascular radiation damage
3. elucidation of the interaction between radiation-induced cardiovascular damage with other risk factors e.g. elevated cholesterol levels for atherosclerotic vascular damage and for microvascular damage
4. clarification of the histopathological and biochemical (proteomic) development of cardiovascular radiation damage, in particular elucidation of differences in responses at high and low radiation doses
5. development of a dose specification system for inhomogeneous dose distribution in the heart and the cardiovascular system, which will be based on the identification of critical subvolumes in the heart and their anatomical distribution in the organ.
Project results:
The CARDIORISK project addressed both macrovascular and microvascular radiation damage after local irradiation. Different groups were developing and implementing various experimental methods and models making use of the same cells and tissues most of which were centrally prepared and provided to all members of the research consortium. This way, results obtained in this research programme apply to identical biological material, treated identically with radiation. In vitro assays and in vivo experiments were all based on cells, tissue or animals from two genetically different mouse strains: Apoliprotein E (ApoE)-/- mice deficient in ApoE and prone to the development of atherosclerosis and C57Bl6 mice as normal counterpart. Molecular and cellular responses at different times can thus be related directly to histopathological and functional changes of the irradiated cardiac microvasculature or the irradiated arteries.
Development of experimental methods
The major tasks of the CARDIORISK consortium within the first two years were the development and establishment of experimental systems to study the radiation biology of cardiovascular effects. Both in vivo and in vitro model systems were deployed to describe and quantify the effects of ionising radiation occurring at low doses.
Local irradiation of the heart, the a. carotis and the A. saphena, partners TUD, NKI
Small animal irradiation setups were developed to reproducibly irradiate either the heart, the a. carotis with the aortic arch or the peripheral A. saphena. Irradiation consisted of doses ranging from 0.2 to 16 Gy. Animals or heart tissue were then distributed to all consortium members for further analysis on comparable biological material. As radiation-induced cardiovascular disease represents a late occurring event, animals were followed up to 60 weeks and investigated at different time points after irradiation (20, 40 and 60 weeks).
In Dresden, a total of 1648 animals were irradiated locally to the heart at doses between 0 (sham irradiation) and 16 Gy. Before irradiation, the correct position of the hearts was verified by digital radiographs, resulting in a total body dose of 4 mGy. Subsequently, single local heart doses of 0.2 2, 8 or 16 Gy were applied. A control group received sham irradiation with 0 Gy. Live animals or heart preparations were then distributed between the partners for the respective experiments. An additional 291 animals were irradiated locally to the heart at the NKI (doses of 0, 2, 8 and 16 Gy). These animals were used for assessment of cardiac function, prior to evaluation of morphological damage and changes in gene and protein expression, by NKI and CARIM.
Preparation of cardiovascular endothelial cells, cardiomyocytes and tissue for functional assays, partners USFD, MSCCI, TUM
As endothelial cells (EC) are among the putative target cells, methods were developed to extract primary EC from non-irradiated and irradiated hearts after different radiation doses (0.2-16 Gy) and different follow-up times.
Cardiac endothelial cell extraction methods were developed and optimised by three partners. They consistently yielded about 100 000 cells / heart after different radiation doses and at different time points after irradiation up to 68 weeks.
Development of ex vivo angiogenesis assays, partner USFD
Cardiac injury triggers repair processes in the myocardium that are dependent on neo-vascularisation. Angiogenesis is therefore central to cardiac repair. To determine whether radiation damage modulates the capacity of cardiac tissues to repair and re-vascularise, novel assays/approaches to determine 'angiogenic' activity of cells derived from the mouse heart at predetermined time points after in vivo irradiation were developed by USFD:
1. In the first approach, explants of both atrial tissue and ventricular myocardium were established in three-dimensional fibrin gels and outgrowth of endothelial sprouts from these explants were used to assess endothelial sprouting after irradiation Each heart section was dissected and around 50 tissue explants about 1 mm3 in size were embedded in a fibrin-gel and overlaid with growth medium, incubated for 10 days and then assessed for extent of sprouting using a semi-quantitative approach. Explants with no sprouts were scored as zero. Explants with a few sprouts were scored as one or two, while explants with extensive sprouting were assigned scores of three or four.
In a second approach, a 'novel' fibroblast-endothelial self-assembling angiogenesis assay was developed and used to monitor cardiac endothelial cell migration, and remodelling into capillary-like structures among fibroblasts/ myofibroblasts/ pericytes. Hearts of irradiated animals were enzymatically digested into a single cell suspension, counted and equal numbers of viable cells plated in 24-well cluster plates. After 7 to 10 days, cultures were fixed and stained with lectin to visualise endothelial cells forming capillary like structures surrounded by fibroblasts/pericytes.
In a third approach, a co-culture model for cardiac EC (CEC) and cardiac fibroblast extracted from (sham-)irradiated mouse hearts was developed to investigate the influence of fibroblast on the angiogenic properties of CEC.
Fibroblasts were plated, allowed to grow for 7 - 10 days to establish post-confluent conditions. Fibroblasts were then irradiated. Endothelial cells were plated on these irradiated fibroblast beds and allowed to grow for 10 days in cultures after which cultures were fixed and stained with lectin to visualise endothelial cells forming capillary like structures.
In vivo imaging of cardiovascular function, partners TUD, NKI, CARIM
To study cardiovascular changes in vivo over the whole follow-up period of 60 weeks within the living animals, non-invasive imaging methods were implemented.
For imaging of the peripheral A. saphena, a new imaging technology was validated by TUD called optical coherence tomography (OCT) which allows structural as well as functional vessel imaging in vivo at specific time points. With OCT imaging, arterial diameter of A. saphena under normal conditions, after vasoconstriction and vasodilatation at day one and three months after irradiation with 0-16 Gy were investigated. The left A. saphena was locally irradiated with 2 to 10 Gy and investigated 3, 6, 9, and 12 months after irradiation to validate the suitability of the method for investigations of late functional changes in arteries exposed to different radiation doses. The arteria of the right leg served as an individual control. OCT after local application of saline served as a control as well. Vasodilation and vasoconstriction was pharmacologically induced. The resulting two dimensional (2D) OCT cross sections of the A. saphena are used to define the respective inner arterial diameters under the different conditions.
For imaging of the heart, single photon emission computed tomography (SPECT-CT) was validated and used as well as ultrasound imaging. With these imaging modalities cardiac blood volume, end diastolic and systolic volumes, ejection fraction and fractional shortening were measured, as relevant cardiac function parameters.
Development of dynamic adhesion assay, partner QUB
The dynamic adhesion assay was further developed by creating a sealed system that allows to carry out the whole experiment under sterile conditions. A preconditioning approach is now used under flow conditions, which leads to re-organisation of the cellular cytoskeleton and pre-orientation of the cells in the direction of the flow. Significant challenges still remain with non-uniform adherence of monocytes and currently work is continuing to improve this.
Investigation of macrovascular effects
Structural changes, partners TUD, NKI
A. carotis: Atherosclerotic changes are observed at high doses only and are mouse strain dependent. Wild type C57Bl6 mice did not develop atherosclerotic lesions within a 30-week follow-up period after carotid artery irradiation (below, at and above the bifurcation) irrespective of dose. Around 20 % of mice developed early fatty streaks at 30 weeks after 14 Gy, but no mature atherosclerosis was observed.
Local high dose (8 and 14 Gy) irradiation to the carotid region of ApoE-/- mice resulted in more atherosclerosis and an inflammatory lesion phenotype in the long term ( around 30 wks follow-up). At 22-30 weeks after carotid artery irradiation of ApoE-/- mice there was a two-fold increase in the number of atherosclerotic lesions and total plaque burden after 8-14 Gy, compared with age and sex-matched controls. The majority of lesions in the irradiated carotid arteries of ApoE-/- mice irradiated with 14 Gy were granulocyte rich, with thrombotic features (iron containing macrophages, thrombin deposits, endothelial cell bands indicative of stimulated angiogenesis), whereas these features were much less common in age-matched controls or after irradiation with 8 Gy. In the short term (1-4 wks), 14 Gy also resulted in a higher incidence of fatty streaks. On the other hand, a lower dose of 2 Gy did not affect lesion size or phenotype.
Heart
There were no obvious atherosclerotic changes in after irradiation of hearts of C57Bl6 mice. However, at 20 weeks after 16 Gy to hearts of ApoE-/- mice there was foam cell accumulation in the endocardium of 17/20 hearts examined, versus 1/14 in age matched controls. There were also atherosclerotic lesions in the coronary arteries of 3/15 of the hearts irradiated with 16 Gy, compared with 1/14 in controls. Atherosclerotic changes at longer follow-up times are currently being evaluated.
A. saphena: No atherosclerotic lesions regardless of radiation dose and time of follow-up. In contrast, no atherosclerotic lesions in A. saphena of C57Bl6 or ApoE-/- mice were observed irrespective of dose and interval after irradiation. For detection of fatty streaks, fat uptake and foam cells oil red and CD45 stainings were conducted in all 10 weeks animals. Some fat vacuoles were detected in Tunica intima in four animals (two of the 2 Gy and the 16 Gy group). CD45 staining showed no monocytes in any layer of the A. saphena. Hematoxylin - Eosin staining showed no pathological changes after irradiation (e.g. edema, plaques). Circumference of A. saphena was measured in 18 animals and varied by 10 % without effect of irradiation. CD45 stainings and circumference measurements will therefore not be followed up in the main cohort.
In vivo optical imaging of vascular function, partner TUD
For imaging of the peripheral A. saphena , the OCT imaging technology was validated which allows structural as well as functional vessel imaging in vivo at specific time points.
Radiation does not alter the functional compliance of the A. saphena
Independent of post-irradiation interval and dose, the diameter at maximum vasoconstriciton was about 50 % of the initial control diameter. After vasodilation, the diameter increased to 1.5x the control diameter, again independent of time-interval and dose. Therefore, no radiation induced changes in functional compliance of the A. saphena were observed. Clearly higher vasodilatatory values are found for ApoE-/--mice at 3 months, compared to C57BL/6-mice. No leg contracture was observed.
The role of inflammatory, adhesive and thrombogenic responses, partners TUD, NKI, UL, UROS
Investigation of A. carotis, NKI:
To investigate whether the increased incidence of fatty streaks (short term) and inflammatory lesion phenotype (long term) is caused by upregulation of inflammatory molecules directly after irradiation, expression of several inflammatory and thrombotic molecules were analysed up to 60 weeks after irradiation using immunohistochemistry and micro-array analysis.
Inflammatory markers were examined, namely P-selectin; intercellular adhesion molecule one (ICAM1); vascular cell adhesion molecule one (VCAM1); thrombomodulin (TM); endothelial oxide synthase (eNOS); monocyte chemoattractant protein one (MCP1); tissue factor (TF); endoglin; von Willebrand factor (vWF) and plasminogen activator inhibitor-1 (PAI-1). The amount (percentage area positive for signal) and intensity of the inflammatory proteins in the endothelium of the carotid arteries were quantitatively analysed at three positions: above, around and below the bifurcation on the carotid arteries.
Expression pattern of inflammatory and prothrombotic factors is independent of structural changes
There was significantly less VCAM1 expression in unirradiated carotid arteries of C57Bl6 mice than in ApoE-/- mice. For the other markers, there were no base line differences between the strains.
One day after 14 Gy irradiation CD31 expression was significantly decreased in the ApoE-/- mice. There were no significant differences in the percentage of EC expressing ICAM1, VCAM1 or TM, and no differences in TF expression in the arterial wall. 1 week after irradiation with 14 Gy, but not lower doses, the percentage of EC expressing ICAM1 was reduced in both strains of mice and VCAM1 was reduced in ApoE-/- mice only. Four weeks after irradiation with 14 Gy the expression of TF and TM was increased in ApoE-/- mice but was reduced in C57Bl6 mice. Irradiation did not cause significant changes in expression of CD31, eNOS, MCP1 or endoglin in the carotid artery at one or four weeks. Eight Gy irradiation did not result in changes in expression of ICAM1, VCAM1, TM, TF, one and four weeks after irradiation in the C57BL6 mice.
Messenger ribonucleic acid (mRNA) analyses:
The first analyses have been performed on the ApoE-/- carotid arteries at one and four weeks after irradiation with 2, 8 and 14 Gy. TM and MCP1 were the only genes that were significantly upregulated at four weeks after 8 Gy.
Investigations of A. saphena, partners TUD, UL, UROS:
Sections of irradiated A. saphena were prepared for immunohistochemical analysis of thrombotic and inflammatory changes in C57/Bl6 and ApoE-/- mice 3, 6, 9, 12, 18 months after irradiation. Twelve different markers were tested for suitability and six (VCAM, E-Selectin, MCP1, TM, CD31 and ICAM) were finally chosen for evaluation. Six markers were not included in the final studies (CD45, Thy-1, Alkaline phospahtase), as the tunica intima is negative for these markers, vWF, KC and TF stained too weak for quantitative evaluation despite of testing different antibodies and protocols.
Strain specific expression of inflammatory and prothrombotic factors without structural atherosclerotic changes
Local increases of CD31 mean fluorescence in C57/Bl6 mice six months after irradiation were observed. CD31 showed only local significances and a tendency to decrease three months after irradiation in Bl6.
ICAM1 mean fluorescence intensity showed increases at early time points (three and six months Bl6; three months ApoE) already at intermediate doses higher than 2 Gy, whereas after nine months no changes were found.
In ApoE-/- mice VCAM showed a partial increase at three months post radiation mainly at 16 Gy. C57/Bl6 mice showed a higher and dose- dependent increase also at three months and additionally at six months. The total area of VCAM-expression showed a tendency to expand at later time points and higher doses.
In ApoE deficient mice local increases in total area of TM 6 and 12 months after radiation exposure were found. In C57/Bl6 fluorescence intensity partially increased 6 months post irradiation. In ApoE-/- a dose-dependent three months post radiation could be revealed. Irradiation with 2 Gy increased the fluorescence intensity whereas 8 and 10 Gy resulted in decreased fluorescence intensities.
MCP-1 positive area showed high variation but no significant and consistent changes. No significant effects in E-Selectin expression in both ApoE and Bl6 were detected.
Investigation of microvascular effects in irradiated heart
In vivo functional imaging of microvascular perfusion, partners NKI, CARIM
Novel in vivo imaging methods (micro SPECT-CT, ultrasound) were used for evaluation of cardiac blood volume and function. One of the tracers used (myoview) can be used to measure changes in cardiac microvascular perfusion in humans undergoing SPECT-CT. We had hoped to use this tracer quantify focal and/or overall reduction of blood perfusion in the irradiated mouse heart. However, this turned out not to be possible due to the extremely high heart beat rate in the mouse. We therefore had to rely on invasive methods, i.e. immunohiostochemistry after injection of fluoresceine-isothyocynate (FITC)/lectin, for evaluation of perfusion in the mouse heart.
SPECT-CT using the blood pool tracer Tc99m-HSA, performed at NKI, showed that cardiac blood volume of C57Bl6 mice decreased significantly at 20 weeks after 2 to 16 Gy (18 to 22 %), with a further decrease (20 to 36 %) at 60 weeks after two to eight Gy.
Gated SPECT-CT using the Tc99m-myoview (at NKI) demonstrated a significant reduction in end diastolic volume (EDV) and end systolic volume (ESV) and increased ejection fraction (EF) at 20 and 40 weeks after irradiation.
Ultrasound measurements (at CARIM) on C57Bl6 mice also showed significantly decreased EDV and ESV volumes and increased EF at 20 weeks after 16 Gy, but these parameters, as measured by US, had largely normalised by 40 weeks. Cardiac function of ApoE-/- mice was measured at 20 weeks after 0 and 16 Gy, by both ultrasound and gated SPECT (myoview). Both analyses indicated very similar reductions in EDV and ESV and increased EF in ApoE-/- mice as in C57Bl6 mice at 20 weeks after irradiation.
Similar reductions in EDV and ESV were seen in mice irradiated in Dresden and assessed by ultrasound investigations (at TUD). In these studies there were no consistent changes in EF in irradiated mice, however, there was a trend for increased EF at 20 weeks after irradiation in all dose groups.
Morphometry of microvascular density, partners TUD, NKI, UL:
Three groups investigated capillary density of irradiated hearts using different methods and criteria of quantification looking at different parts of the left ventricle.
1. NKI: Microvascular density (MVD) was determined in transverse frozen sections of mouse hearts (male C57Bl6 and ApoE-/-) at 20, 40, and 60 weeks after 0, 2, 8 and 16 Gy (uniform, whole heart irradiation). An anti-CD31 antibody with diaminobenzidine (DAB) staining was used to visualize cardiac vasculature of the central or apical part of the heart. To quantify the percentage of perfused microvessels, FITC-lectin was injected intravenously (i.v.) five minutes before tissue harvesting. For quantification of microvessels, five random fields (40x objective) were photographed and a computerised morphometry system was used to quantify MVD. Vessels beneath a size of 1.5 or above 200 µm2 were automatically excluded from the measurements.
2. UL: Apical parts of transverse frozen sections from male ApoE-/- and C57/Bl6 mice were evaluated at 20, 40 and 60 weeks after 0, 0.2 2, 8, and 16 Gy. A fluorescent CD31 antibody was used to detect the vasculature and results are expressed as CD31+ events normalised per 4'-6 diamidino-2-phenolindole (DAPI) positive nuclei.
3. TUD: Formalin fixed, paraffin-embedded sections were used to evaluate the number of microvessels per cardiomyocyte in the papillary muscle. This was based on a staining of adjacent sections with CD31/DAB, to detect endothelial cells, and laminin, to distinguish the cell boundaries of the cardiomyocytes. The number of microvessels per cardiomyocyte was counted manually, at 20 and 40 weeks after 0, 8 and 16 Gy (male ApoE-/- and C57/Bl6 mice).
Dose and time dependent decrease in MVD is accompanied by decrease in perfusion, but not hypoxia. In C57Bl6 hearts of mice irradiated at the NKI, there was a transient increase in microvascular density in the central part of the left ventricle (16-20%) at 20 weeks after two to eight Gy, and a trend for a decrease after 16 Gy (no significant changes in apical regions). A significant decrease in microvascular density was observed at 40 weeks after 16 Gy and 60 weeks after 8 Gy, indicative of progressive loss of microvessels after mid to high doses. In ApoE-/- mice, capillary loss occurred earlier and at lower doses (from 20 weeks after 8 Gy).
Analysis of CD31 and FITC-lectin stained sections showed that 92 % of microvessels in hearts of control C57Bl6 mice were perfused at 20 weeks after treatment. There was no significant change in the number of perfused vessels after irradiation with 2 to 16 Gy. At 40 weeks after 16 Gy, the reduction in MVD was accompanied by a very small, but significant, decrease in perfusion of those vessels (84 % of the remaining microvessels were perfused after 16 Gy, versus 87 % in controls). In ApoE-/- mice, the reduction in MVD at 40 weeks after 8-16 Gy was also accompanied by a small reduction in the perfusion of remaining vessels (86% versus 91 % in age matched controls). However, these small reductions in perfusion did not lead to any detectable hypoxia (EF5 staining).
In C57/Bl6 mice irradiated at TUD and analysed at UL, there were no changes in CD31+ events, normalised per DAPI nuclei, in the apical part of hearts at 20 weeks after irradiation, but at 40 to 60 weeks after 8 to 16 Gy there were decreases in MVD (only significant at 40 weeks after 16 Gy). In ApoE-/- mice, there were no consistent changes in MVD.
The morphometric analysis of the microvascular density in the papillary muscle by TUD, however did not reveal any consistent dose or time dependent changes.
Loss of alkaline phosphatase already occurs at low radiation doses and precedes reduction in micro vessel density, partner NKI
To investigate whether structural changes in MVD were associated with functional changes, changes in expression of alkaline phosphatase (ALP) vWF in microvessels, evidence of vascular leakage (staining for albumin) and collagen deposition in interstitial areas was investigated by NKI.
At 20 to 60 weeks after irradiation with 2 to 16 Gy, the amount of capillary staining for ALP was reduced by up to 50 % in both strains of mice, with more pronounced response in ApoE-/- mice.
In C57Bl6 mice, there was a significant decrease (30 to 44 %) in percentage tissue stained for ALP at 20 weeks after irradiation with 8 to 16 Gy. By 40 weeks, the 2 Gy dose group also had significantly less ALP expression, indicative of further progression of endothelial damage in small blood vessel. At 60 weeks the ALP expression in irradiated groups (2 and 8 Gy) was 50 % of the mean control value, but these differences were borderline significant (p=0.05).
In ApoE-/- mice, there was an earlier reduction in microvascular density than in C57Bl6 mice. At 20 weeks after 8 to 16 Gy MVD was already significantly less than in the controls and this persisted until at least 40 weeks after 8 to 16 Gy (longer times were not evaluated). There was also an earlier, and more pronounced, decrease in ALP expression in microvessels of irradiated ApoE-/- hearts than in C57Bl6 mice. At 20 weeks there were significant, dose related, reductions in ALP after all doses and by 40 weeks the ALP levels in vessels of irradiated hearts were only 55 to 36 % of controls.
There was a significant increase in capillary vWF expression (prothrombotic) at 20 to 40 weeks after high doses (8 to 16 Gy) in both strains. This response was slightly less marked than was seen in C57Bl6 mice.
To investigate whether the structural and functional changes in the microvasculature were associated with vascular leakage, albumin deposition in the myocardium was examined. At 40 weeks, half of the C57Bl6 hearts irradiated with 2 Gy and almost all hearts irradiated with 8 to 16 Gy showed albumin in the myocardium; this was absent or very mild in controls. After 16 Gy, myocardial albumin was extensive in 5 of 11 hearts and all these animals also had diffuse amyloidosis, which was confirmed with a Congo red staining. Amyloidosis was not seen in the irradiated ApoE-/- hearts.
Structural effects:
Epicardial thickness more than doubled at 20 weeks after 8 to 16 Gy and remained increased at 40 weeks. This was associated with the presence of iron containing macrophages (indicative of previous haemorrhage).
Increased interstitial collagen deposition (indicative of fibrosis) was present in the myocardium at 40 weeks after 8 to 16 Gy and at 60 weeks after two to eight Gy. In ApoE-/- hearts, these changes were seen earlier and at lower doses.
Electron microscopy studies showed dilated mitochondria with lysis of the cristae in cardiomyocytes at 10 weeks after eight Gy. Myofibrils showed focal lysis and disrupted structure and endothelial cells were dilated. No relevant structural disturbance were seen at two Gy or lower doses.
At 40 weeks, half of the hearts irradiated with two Gy and almost all hearts irradiated with 8 to 16 Gy showed extravascular albumin in the myocardium; this was absent or mild in controls. After 16 Gy, myocardial albumin was extensive in 5 of 11 hearts and all these animals also had diffuse amyloidosis. The remaining six animals from this group exhibited a more focal amyloidosis.
There was an increase in interstitial collagen deposition from 40 weeks after 8 to 16 Gy and 60 weeks after 2 to 8 Gy in C57Bl6 mice. In ApoE-/- mice this collagen deposition occurred earlier (from 20 weeks).
Stress response and immune function in irradiated hearts, partners TUM, MSCCI, HELMUC:
Membrane Hsp70 expression only predicts short term radiation-induced stress response. The cell surface expression and secretion of hsp70 on mouse heart ECs was analysed by determining membrane Hsp70 expression density on viable mouse heart ECs. A maximum membrane expression of Hsp70 was observed already at eight weeks after irradiation which then normalised at 10 weeks. Reduced cell surface expression levels were observed thereafter until 60 weeks after irradiation.
Serum Hsp70 expression indicates long-term persistence of radiation-induced stress response. Hsp70 protein levels were determined in the serum of mice after punction of the heart. Serum analysis of Hsp70 protein levels after irradiation revealed a significant increase only 60 weeks after irradiation.
Hsp70 expression in whole heart preparations: Changes in the expression of the major stress-responsive gene Hsp70i were assessed in both cardiac endothelial cells and in a whole heart tissue up to 40 weeks after irradiation with 0.2 2, 8 and 16 Gy doses. Mice treated with heat shock were treated as a positive control.
Gene expression analysis: CEC isolated from juvenile mice were irradiated in vitro with 2 and 8 Gy, and then 24 hours after the exposure cells were collected and RNA isolated. In addition, animals were irradiated in vivo with two and eight Gy and CECs were isolated 12 and 36 hours, 6 days, 20 and 40 weeks after the exposure, and then RNAs were purified directly after the cell isolation. Only at eight Gy up-regulation of VCAM1 gene was observed early after radiation exposure. At 20 and 40 weeks of follow-up no significant difference between irradiated and sham-irradiated animals for VCAM1, E-Selectin, Bnip and Glut1 (hypoxia-related genes) expression was observed. Radiation caused up-regulation of Hsp70i gene expression, which was visible in a dose-dependent manner within 24 hours after irradiation in vitro and in vivo in CEC isolated from irradiated animals at 20 and 40 weeks after the exposure.
Proteom analysis in endothelial cells in vitro: The human endothelial cell line EA.hy926 and the human coronary artery endothelial cell line (HCAEC) were irradiated in vitro with 0.2 Gy gamma (Co-60) and analysed for proteomic changes 4 and 24 h later. In Ea.hy926 15 significantly differentially expressed proteins were identified, of which 10 were up-regulated and five down-regulated, with more than + 1.5-fold difference compared to unexposed cells. Pathways influenced by the low-dose exposure included the Ran and RhoA pathways, fatty acid metabolism, and stress response. Radiation of 0.2 and 1 Gy did not cause phosphorylation changes in the proteome of EA.hy926 cells when examined 10 min, 30 min, or 4 h later using the total cell lysate or cytosolic fraction. In HCAEC, the total number of deregulated spots after 4 and 24 h was 29 and 26, respectively. The identification of the proteins represented as deregulated spots by ESI LC-MS/MS elucidated several pathways affected by radiation including cell death, cell morphology, RhoA signalling and actin-based motility by Rho.
Inflammatory response and thrombotic changes in irradiated hearts, partners CARIM, NKI, UL, UROS, IRSN:
ApoE-/- and C57/Bl6 mice sacrificed 20, 40 and 60 weeks after irradiation with 0, 0.2 2, 8, and 16 Gy (eight animals per dose) were analysed for CD31, thrombomodulin, VCAM-endocard and CD45. Increased mean area of CD31 staining 20 and 40 weeks post irradiation may indicate endothelial swelling at high doses (8 and 16 Gy) and was found in ApoE-/- mice only. Similarly CD31 fluorescence intensity increased in ApoE-/- mice only, indicating an increased extravasation potential already after irradiation at two Gy. At high doses (8, 16 Gy) a time depend immigration of leukocytes was detected, by increase of CD45 positive cells, in both endocardium and myocardium at 40 weeks post irradiation. This behaviour is not due to the onset of spontaneous atherosclerosis in the ApoE-/- genotype because it was found a similar reaction in the C57/Bl6 wild type model. Also VCAM expression on the endocard showed a trend to increase after irradiation at 16 Gy at all time points in both mice strains. Changes in all four markers (CD31, Thrombomodulin, VCAM-endocard and CD45) indicate a proinflammatory response mostly at high doses (8 and 16 Gy). At low dose (0.2 Gy) only small changes are seen at a few time points with no significances.
A reduction of Thrombomodulin (TM) positive events was found 20 and 40 weeks (C57/Bl6), and 60 weeks (ApoE-/-) after irradiation at 16 Gy. TM exhibits anti-inflammatory, antithrombotic properties, partly by being a sink for thrombin but also by reducing the expression of adhesion molecules. Reduction of TM positive capillaries might indicate a proinflammatory response in the background of unchanged overall capillary count. The ICAM-1 staining was weak and did not appear to be influenced by irradiation.
Analyses of inflammatory gene expression
Microrarray data are only available for C57 Bl6 mice at present. There were no large changes in inflammatory gene expression at 20 weeks, but VCAM1, ICAM 1 and KFL2 were slightly increased at 20 weeks after two Gy. Pathway analysis scored high for cell-to-cell signalling pathway including CXCL1, S100A8 and S100A9. Gene expression data at 20 weeks after 16 Gy showed modest increases in endoglin and decreases in VCAM1. Pathway analysis scored high in cell death/development disorder-pathways, including MYL4, myosin, connective tissue growth factor (CTGF), PAI-1, TIMP3 and MMP.
By 40 weeks after 2 Gy, the expression levels of PAI-1, TM, ICAM1 and KLF2 were increased and pathway analysis scored high for cellular movement-immune cell trafficking pathway and molecular transport/cell cycle - pathway. These pathways include upregulation of LYZ, CCL2 and CCL4. Gene expression data from 16 Gy showed increases in Ki67, PAI-1, P-selectin and endothelin and significant down-regulation of Smad5 and PDGFb. Pathway analysis indicate high scores for cell movement/immune cell trafficking pathway including SPP1, CXCL13 and ESM1.
Using real-time PCR expression of the inflammatory genes (VCAM-1, MCP-1, E-Selectin) was investigated. Neither samples of ApoE-/- mice sacrificed 20 weeks after irradiation nor mice sacrificed 60 weeks after irradiation showed any change in expression of inflammatory marker gene VCAM or E-selectin normalised to reference genes GAPDH and ß-actin. Due to very few native material MCP-1 could not be amplified for any of the samples.
Functional integrity
Studies in EC
Experiments at USFD focused on analysing angiogenic responses in the heart tissue following irradiation with the aim to determine whether radiation damage modulates the capacity of cardiac tissues to repair and re-vascularise. Since isolation and subculture of enough endothelial cells from older mice proved challenging, alternative assays/approaches to determine 'angiogenic' activity of cells derived from the mouse heart were developed.
The results demonstrate that radiation decreases the capacity of reparative angiogenesis in heart tissues. Severe impairment of this process was statistically significant at high radiation doses while a trend for a reduced angiogenic activity is evident at moderate to low doses. Radiation at high doses (8 and 16 Gy) inhibited capillary-like tube formation in ApoE-/- mouse heart cultures at 20 weeks post-irradiation. Capillary-like structures were affected both qualitatively and quantitatively. In particular, in the irradiated groups these were less prominent than the controls and were characteristically shorter and narrower, fewer in numbers and had fewer branching points. There were no significant differences in the total number of viable cells initially extracted from irradiated and non-irradiated hearts. However, once in culture, cells from the 8 and 16 Gy irradiated groups showed a reduction in growth and viability correlating with the profound inhibition of capillary-like network formation seen at these radiation doses. A similar response was seen in C57BL/6 hearts at 20 and 40 weeks post-irradiation. A dose dependent reduction in 'angiogenic index' was observed 20 weeks post-irradiation. A similar pattern was observed at 40 and 60 weeks post irradiation. Statistical significance was achieved only for the 8 and 16 Gy dose groups.
Work at MSCCI focused on analysing irradiation-induced stress responses in cardiac endothelial cells (CEC). These included analysis of expression of stress-response genes described above as well the analysis of cytosceletal structures. Analysis of cytoskeletal structures (actin stress fibres) was analysed in CEC isolated from animals irradiated with two and eight Gy, 20 weeks after irradiation (cells isolated 40 weeks after irradiation were not suitable for analysis). Stress fibre analysis was also performed on CECs isolated from animals irradiated in Gliwice at shorter times after irradiation and on CECs from juvenile animals irradiated in vitro. Filamentous structures of actin were detected upon staining with fluorescently-labelled phalloidin.
Significant changes were observed in CECs irradiated in vitro with eight Gy. Mice were irradiated with two and eight Gy, and at 12 hours, 36 hours, 6 days and 20 weeks after irradiation endothelial cells were extracted. Irradiation with two Gy increased the number of cells with high number of stress fibres (to about 40 %); such changes persisted for 6 days after exposure but were not seen after 20 weeks. Irradiation with eight Gy resulted in a highly significant increase in numbers of stress fibres. We conclude that radiation induced dose-dependent changes in actin cytoskeleton (i.e. formation of contractile stress fibres) in cardiac endothelial cells irradiated in vivo and in vitro. Irradiation with 16Gy did not increased permeability of monolayers formed by CECs to dextran (MW 40 000 Da), neither 3 nor 24 hours after exposure. Therefore, this part of the experimental program was not pursued further.
Work at QUB focused on studying intercellular communication in endothelial cell cultures. Originally, freshly isolated endothelial cells were to be used, but this proved difficult, therefore QUB used established mouse endothelial cell lines for these studies. Previously, the characterisation of various endothelial cell lines was completed in collaboration with UL and the decision was made to consolidate work using a mouse cardiac endothelial cell line H5V thereafter shared by all members of the consortium working on endothelial cell lines. Other mouse endothelial lines used included lymph node large vessel (mlEDN1); brain microvascular (bEND3); heart microvascular (H5V) and previously heart microvascular (MHEC5-T) and lymph node large vessel (SVEC4-10). In particular, QUB focussed on the characterisation of the cytokine profile of endothelial cell lines after irradiation at 0.1 and 2 Gy both in the presence and absence of TNFa. The aim was to understand mechanisms related to the focal nature of cardiovascular disease associated with response of endothelial cells to low and targeted doses of radiation.
In addition, the impact of irradiated monocytes on adhesion properties was determined, and changes in adhesion properties, both static and dynamic, were evaluated along with cytokine changes in endothelial cell monolayers. This was extended to situations where cells are irradiated both directly and under bystander conditions.
The brain derived bEND3 cells responded to irradiation with doses between 0.05 Gy and 5 Gy with a significant increase in adhesion properties under both static and dynamic conditions. Statistically significant changes in adhesion properties were not observed in the other two cells lines While pro-inflammatory effects of radiation were observed only in bEND3 cells, in all three cell lines evidence for radiation-induced anti-inflammatory effect 24 h after irradiation of pre-activated with TNFa cells was seen. Analysis of inflammatory cytokine expression profile of H5V, bEND3 and mlEND1 cells suggests that radiation induced pro-inflammatory response can occur in all three endothelial cell lines as early as 30 minutes after irradiation in both microvascular and lymph node derived endothelial cells. Elevated levels of MCP-1 and decreased expression of transforming growth factor-beta (TGFß) persisted for more than 24 hours in pre-activated cells. An anti-inflammatory effect of radiation on pre-activated endothelial cells was also observed as early as 30 min after irradiation.
X-ray exposure with partially shielded cultures indicates a role for bystander signalling for both cytokine expression and monocyte adhesion. This is observed at 0.1 and 2 Gy and suggests intercellular communication plays an important role. Antibody blocking cell surface adhesion molecules led to significant decrease in monocyte adhesion in mlEND1 and bEND3 cells with the effects being more pronounced in pre-stimulated cells. The important role of the Nf?ß pathway in radiation-induced inflammation was confirmed by adding to the media Bay11-7085 prior to treatment with radiation or TNFa, which led to a significant decrease in adhesion.
Irradiation of the monocyte population with low and moderate doses (0.1 Gy and 2 Gy) causes decrease in the ability of monocytes to adhere to endothelial cells, suggesting potential radiation-induced anti-inflammatory effect. The effect was observed in all three cell lines - bEND3, mlEND1 and H5V.
Studies in heart tissue and cardiomyocytes, partners HELMUC, IRSN:
Mitochondrial proteom and function
The radiation-induced in vivo effects on cardiac mitochondrial proteome and function were investigated at 20 and 40 weeks after local irradiation of the heart. Myocardial mitochondria were isolated from whole heart tissue and tested for proteomic and functional alterations. The proteomic analysis was done using peptide (ICPL) and protein quantification (2D-DIGE). Altogether 18 proteins were found deregulated 20 weeks after 2 Gy-irradiation, of which 12 were up-regulated and six downregulated. Deregulated proteins belonged to Oxphos complexes, carbohydrate metabolism or mitochondria-associated cytoskeleton.
Mitochondrial respiration was measured using succinate as the substrate to analyse the intactness of mitochondria and to determine the efficiency of oxygen consumption. Succinate-stimulated respiration decreased significantly in mitochondria from mouse hearts irradiated with two Gy after 20 w and 40 w, respectively. No significant alteration was found in the succinate-driven respiration after irradiation with 0.2 Gy of C57BL/6 mice or with any dose or time point in ApoE-/- mutant. No changes in mitochondrial membrane potential were registered in mitochondria isolated from either C57BL/6 or ApoE-/- at 40 weeks using different stimuli.
Rho/ROCK pathway
Acute radiation effects on primary cardiomyocytes isolated from 12 week old C57BL6 and irradiated in vitro and long-term radiation effects in primary CM isolated from sham-irradiated and irradiated C57BL6 and ApoE-/- mice at 0.2 2 and 16 Gy; 20, 40 and 60 weeks post-irradiation were performed. Investigation of acute and delayed radiation-induced Rho/ROCK and Smad pathway activation along with remodelling of the actin cytoskeleton were performed in isolated CM using biochemical assays and immunohistochemistry. The activation of the Rho/ROCK pathway by monitoring Rho protein isoprenylation by immunoblotting. ROCK kinase was assessed by monitoring phosphorylation of its target myosin-light chain (MLC) and upregulation of CTGF, a known downstream target protein of the pathway. CTGF is also a target of the TGFß pathway together with PAI-1. Activation of the following members of the TGFß cascade was studied: TGFß1, TGFß RII, Smad2/3/4, PAI-1. Immunostaining was performed to assess radiation-induced remodelling of the alpha-actin-sarcomeric that constitutes the main cytoskeletal component of cardiomyocytes. The pharmacological inhibitor of ROCK, Y-27632, was used to modulate the phenotypical characteristics of CM isolated from animals 20 and 40 weeks post-irradiation. Pharmacological experiments (with Y-27632) could not be performed 60 weeks post-irradiation due to the fragility of the CM isolated at this time-point. In addition, planned knock down short interfering RNA (siRNA) approaches were withdrawn due to the small yield of CM isolated from irradiated animals and to their fragility.
Radiation-induced sequential activation of Smad and Rho pathways and cytoskeletal remodelling in primary cardiomyocytes:
The balance between TGFß1/Smad and Rho/ROCK signalling in response to low and high doses of irradiation was investigated in CM, 4 h and 24 h post-irradiation. Irradiation at 0.2 2 and 16 Gy induced early phosphorylation of Smad2/3 and a dose dependent increased expression of Smad4, indicative of Smad pathway activation 4 h post-irradiation. This was followed by radiation-induced increase of the isoprenylated Rho 24 h after irradiation, indicative of Rho activation. A trend for Rho activation was observed after exposure at 0.2 Gy but only reached significance for doses above 2 Gy. Consistently, a subsequent production of CTGF occurred 24h after exposure of CM to 16 Gy. Immunofluorescence studies showed a radiation-induced remodelling of the central network of a-sarcomeric actin, which is the main structural protein of cardiac muscle. Actin remodelling was observed after low dose (0.2 Gy), a dose at which Rho proteins were not significantly activated suggesting involvement of additional mechanisms. Such alteration of actin network was also visible in irradiated CM at high dose as early as 4 h after irradiation and persisted at 24h, consistent with the alteration of Rho isoprenylation observed after exposure to high doses.
Rho/ROCK pathway is involved in control of actin remodelling in primary cardiomyocytes
Actin remodelling is thought to depend upon modulation of the Rho/ROCK pathway. To confirm this assessment primary CM were exposed to the pharmacological inhibitor of ROCK, Y-27632. First, the inhibition of ROCK kinase activity upon Y-27632 incubation was monitored by studying myosin light chain phosphorylation (MLC). MLC phosphorylation decreased significantly 1 h after exposure to Y-27632 and was sustained at, 4, 6 and 1h after exposure. Consistently, alteration of the actin network associated with striation disappearance was obvious 4 h after exposure to Y-27632 and persisted at 6 and 16 h.
Primary cardiomyocytes can be isolated from irradiated mice. In order to investigate the TGFßeta1/Smad and Rho/ROCK signalling pathway as a late effect in irradiated cardiomyocytes in vivo, the isolation of CM from irradiated mice was optimised. CM was successfully isolated from non-irradiated and irradiated C57Bl6 and ApoE -/- mice 20, 40 and 60 weeks post-irradiation. 20 weeks post-irradiation, the yield, quality and shape of CM was good with good adhesion capability indicative of cell's quality. It allowed sub-culturing and pharmacological modulation. 40 and 60 weeks post-irradiation, quality and shape of CM was insufficient to provide good adhesion and perform long-term culture, therefore cells were lysed immediately after isolation.
Modulation of the TGFß cascade: In CM no alteration of the TGFß cascade was observed 20 weeks post-irradiation, regulations occurred at later time point. Immunoblot analysis showed a global activation of the TGFß cascade in CM isolated from C57BL6 that mainly occurred at late time points. TGFß increased 40 and 60 weeks post-irradiation in animals irradiated with 0.2 Gy, 2Gy and 16 Gy (40 weeks only available). Interestingly, Smad 2/3 and smad4 inductions were dose dependant 40 weeks post-irradiation. 60 week post-irradiation Smad 2/3 level remained high whereas Smad 4 dropped. Finally, the two TGFß targets studied (CTGF and PAI-1) were only moderately stimulated, mostly 40 weeks post-irradiation. Similarly to what was observed in cardiac tissue, Smad7 expression dropped 40 and 60 weeks post-irradiation. Western-blot analysis showed increased TGFß1 levels in CM isolated from ApoE -/- irradiated at low and intermediate doses. However, the protein level of canonical members of the TGFß cascade, Smad 2/3 remained stable. A trend to induction was observed for Smad4. CTGF expression was not modulated at 0,2 Gy whereas PAI-1 significantly picked at this dose and doses above. At higher doses, CTGF protein level moderately but significantly increased. Smad7 expression dropped 60 weeks post-irradiation.
In conclusion, the results show that cardiomyocyte physiology and molecular responses are altered even by low radiation doses. Differences were found between CM responses between irradiation in vitro or after cell isolation from the irradiated tissue. PAI-1 activation was observed in the both mouse strains suggesting that the TGFß/Smad/PAI-1 cascade operated in fibrogenesis. Yet amyloidosis development which was observed in C57Bl6 but not in ApoE deficient mice, was dependent of ApoE-/-, and the activation of another mechanism must be considered. The current results show that the molecular imprint at low dose is different from that obtained at moderate/high dose. Activation of the Smad2/3 pathway occurs immediately after irradiation but neither Rho pathway nor CTGF were activated after exposure at 0.2Gy. The TGFß cascade is differentially activated in both mouse strains. Smad 2/3/4 seemed to operate in C57Bl6 whereas other protein seemed involved in ApoE. This difference may explain the differential pathogenic picture observed in histology. However, the Rho pathway was not found activated in CM despite immunohistochemical pattern.
Final conclusions
The results of the CARDIORISK project permit the following conclusions which may impact profoundly future research in the area of low-dose radiation risk assessment in the cardiovascular system:
1. radiation at low doses does not per se induce atherosclerotic changes in medium to large arteries
2. inflammation does not play a major role in the development of radiation induced cardiovascular disease (CVD); stress responses can be regarded as a consequence of the effects of radiation induced CVD
3. the reduction in microvessel density and the inhibition of neoangiogenesis even at low to intermediate doses does not directly translate into the clinical picture of CVD, but it may well compromise the ability of the heart to respond to subsequent stress
4. these major conclusions support the hypothesis that low-dose radiation exposure and related subtle subclinical changes are permissive factors for the development of CVD which reduce the capacity of the heart to recover if a clinically relevant event/ insult occurs.
Outlook and future research
The CARDIORISK identified the progressive reduction of the capillary network at intermediate and low radiation doses (less than two to eight Gy) as the key mechanism leading to radiation-induced heart disease. This microvascular effect does not lead to measurable ischemic cardiomyopathy but is likely to reduce the compensatory capacity of the heart to insults from other sources such as coronary heart disease etc. In addition, there are also strong suggestions of direct myocardial injury by low radiation doses after similar latencies. Despite the wide range of research methods and biological targets used in CARDIORISK, some key molecular and cellular processes of this microvascular effect are still largely unknown. Future research should focus on resolving these open questions, in particular:
1. the molecular targets in capillary endothelial cells which lead to functional damage, e.g. loss of alkaline phosphatase. Since CARDIORISK provided evidence that endothelial cells isolated from different organs and different endothelial cell lines show very different responses to low dose radiation exposure those studies should use predominantly endothelial cells isolated directly from the hearts of the investigated animals
2. the role of radiation-induced reduction of alkaline phosphatase in capillary endothelial cells and their role for capillary function
3. the molecular mechanisms of low and intermediate dose radiation damage to cardiomyocytes
4. the mechanisms of progression of general and focal loss of alkaline phosphatase and possible consequential random destruction of entire capillarie
5. elucidation of the apparent intercellular communication processes between endothelial cells within a capillary for the development of capillary damage
6. the dependence of reduction of the capillary network on dose (less than two Gy) and time (more than one year) in longer living animals
7. the relationship of radiation-induced microvascular radiation damage on the topographical dose distribution in the hearts, on age at exposure with regard to different clinical manifestations of radiation-induced heart disease, in particular ischaemic heart disease, valvular disease and conduction defects.
For this research, an integrated programme bringing together cardiological and radiobiological scientists is obligatory. The experience of CARDIORSIK has demonstrated the great potential of this cooperation.
Potential impact:
Socioeconomic impact and the wider societal implications of the project
The CARDIORISK consortium has developed a series of experimental systems to study the radiation biology of cardiovascular effects at low doses in vivo and in vitro. The major goal was to provide the impetus necessary to place radiation protection considerations of late cardiovascular effects in a biological context.
The estimation of radiation risk is primarily based on the assumption that, except for heritable, genetic radiation damage, the only somatic radiation damage to be considered in radiation protection is radiation-induced cancer. This dogma, which went unchallenged for several decades, has been challenged by the epidemiological evidence showing radiation-induced mortality from cardiovascular and cerebrovascular diseases of similar magnitude as from radiation-induced cancer. These epidemiological data demands a reassessment of the concepts, quantities and methods of how risk is defined in radiation protection. In order to incorporate cardiovascular and cerebrovascular radiation risks into the overall system of radiation protection, knowledge is required on pathogenic mechanisms, which only radiobiological experiments can provide.
The main problems for the incorporation of the new epidemiological evidence of radiation-induced late cardiovascular and cerebrovascular damage arise in the definition of dose and the shape of the dose response relationship. In the present system of radiation protection, the effective dose is defined as the sum of the mean radiation doses in the respective organs at risk, which is multiplied with the tissue weighting factor. While the tissue weighting factor may be relatively straightforward to define from epidemiological studies - and it is obvious that the tissue weighting factor for the heart and the brain will have to be increased significantly from their present value of 0.025 the question of dose definition cannot be resolved by epidemiological data. The answer to this problem has to be based on scientific evidence which requires full knowledge on the pathogenesis of the respective radiation-induced fatal diseases including in particular information on the critical structures which trigger the pathogenic pathways and their anatomical distribution in the organ. Only radiobiological experiments such as those performed in the CARDIORISK project were seen to put the methods of how to define radiation dose in the heart in situations of inhomogeneous radiation exposure -which are the rule in particular in diagnostic radiology - on a sound basis of scientific evidence. The present system of radiation protection assumes proportionality between mean organ dose and the risk of radiation-induced cancer since cancer is likely to be a clonal disease, arising from a single stem cell that has been transformed and clonally expanded through many additional steps of progressing malignancy. In contrast to this, cardiovascular diseases and cerebrovascular diseases are examples of obvious multicellular origin. All scientific bodies have consistently argued that the distinction between stochastic and deterministic radiation effects can be based on the distinction of single cell clonal effects, justifying a linear dose risk relationship, and multicellular radiation effects, which is observed only after radiation exposure has exceeded a threshold. This distinction can no longer be upheld in view of the epidemiological evidence on cardiovascular radiation mortality of the A-bomb survivors. The resolution of this problem is of great importance for radiation protection since the entire concept of no-threshold linear dose risk relationships and the concept of stochastic radiation effects has been put into the limbo.
The CARDIORISK project has produced an enormous amount of biological data which relate to the pathogenesis of macrovascular and microvascular radiation damage. There is evidence in the mouse cardiovascular system that that large vessels such as carotid artery, arteria saphena and probably coronary arteries follow different pathogenic pathways but which are all associated with inflammatory processes which, however, are only triggered by high radiation doses which are characteristic of radiotherapy. In contrast, microvascular radiation damage is elicited already by lower radiation doses and develops independent of pro-inflammatory changes which lead to structural changes in the capillary network in the myocardium. This damage starts with functional changes in the capillary endothelial cells causing progressive disappearance of entire capillaries. There is some evidence for a simultaneous stress response in the irradiated myocardial cells although it remains open whether both effects develop independently or one is the consequence of the other.
For macrovascular radiation effects in the heart the results of the CARDIORISK project suggest a threshold-type dose response relationship which depends on the radiosensitivity of several well investigated processes involved in the activation of the inflammatory conversion of age-related atherosclerosis which occur after shorter latencies than the microvascular changes. The target structure in the heart where these processes take place and for which, therefore, the specification of local dose has to be defined are the coronary arteries, and in particular the left anterior descending coronary artery as has also been suggested by the findings of the RACE project.
The CARDIORISK project produced evidence that microvascular radiation damage is different following different pathogenic kinetics in different parts of the ventricular myocardium. These topographical differences may be due to pathophysiological factors or to the different response criteria used by the different CARDIORISK partners. More studies need to be performed on the mechanisms and modifying factors which influence the reduction of the microvascular network. The data produced by CARDIORISK suggest that the reduction and rarefication of the capillary network is progressing over much of the life-span and a linear dose response relationship cannot be excluded. The linear dose dependence, though only validated in the dose range of two to eight Gy, is difficult to reconcile with the observation that functional alterations in endothelial cells and rarefication occur throughout much of the myocardium of the ventricles, and that severity of this effect progresses with dose and time. This puts a big question mark on the concept of linear dose response relationships in general and clonal origin of 'stochastic' radiation damage and may have implications for the entire system of radiation protection in all areas of low dose radiation exposure including the out-of-field exposures in radiation oncology. These findings also suggest that the whole myocardium should be delineated as the sub-volume for which dose should be defined and reported which would be, for microvascular damage, the mean myocardial dose.
In conclusion, different doses to the heart may have to be considered in radiation protection, i.e. the mean dose to the myocardium, and in radiation oncology where in addition, the local doses to the coronary arteries have to be determined and minimised as well.
Impact on other Euratom projects
Close cooperation with the integrated Seventh Framework Programme (FP7) project NOTE which includes studies of cardiovascular damage after low dose and protracted total body irradiation had been agreed. The leader of work package two (WP2) of NOTE was a member of our consortium. He used irradiated hearts, arteries and endothelial cells from the CARDIORISK project and investigate their response with the same methods he used in the hearts of the NOTE project in order to compare the cardiac effects of total body irradiation with those of localised heart irradiation. Moreover, the leader of WP4 of NOTE, who aimed at using the data produced in the NOTE project to develop a bio mathematical model of radiation-induced cardiovascular risk, was invited to all CARDIORISK meetings and had direct access to all results produced by the CARDIORISK project to incorporate them into the bio mathematical model of radiation-induced cardiovascular.
Main dissemination activities and exploitation of results
All results will finally be published in the scientific open literature. However, equally important is the direct input of the results of the described studies and our conclusions from the new data into the discussions of the international committees responsible to adjust the radiation protection concepts, rules and regulations to the progress of knowledge. So far, results from the CARDIORISK project have been presented at 16 national and 55 international meetings and published in 30 scientific articles.
As this project addressed long-term cardiovascular effects, final pivotal publication can only now be prepared after all follow-up times have been analysed. From every partner at least one seminal publication is planned for publication. To specifically address the radiation biology and therapy community, a dedicated issue of 'Radiotherapy and oncology' about the CARDIORISK project, the results and conclusions is planned. In a peer-review process one paper from each partner is being considered for submission together with an introductory editorial. A review paper which described the overall design of the CARDIORISK project was written by the scientific secretary and published in 2010 in 'Radiotherapy and oncology'.
By far the most important dissemination activities during the lifetime of the project were:
1. the 'DoReMi' exploratory workshop on 'Radiation-induced cardiovascular disease from low dose exposure' in November 2010 in Bombon, France
2. the CARDIORISK symposium in June 2011 which heralded the successful conclusion of the project.
The other five talks were given by the leading experts in the field of cardiovascular research in the context of radiation exposure, radiation therapy and chemotherapy coming from specialties like radiation oncology, radiation biology, oncology, cardiology and epidemiology presenting complimentary information from experimental and in particular clinical data. All talks highlighted the role of the CARDIORISK project for planning future research in radiation-induced cardiovascular diseases.
In addition, other knowledge dissemination activities aimed to ensure dissemination and training to graduate and undergraduate audiences, other Euratom contracts and the public and stakeholders.
Project website: http://www.cardiorisk.eu
The aim of the CARDIORISK collaborative research project is to elucidate the pathogenesis of early and late alterations in the microcirculation of the heart and of atherosclerotic lesions in arteries after exposure to low radiation doses in comparison to high radiation doses. A major goal was investigation of early molecular, proinflammatory and prothrombotic changes as well as alterations of myocardial perfusion, cardiac cell integrity and immunological processes.
CARDIORISK aimed to achieve the following objectives:
1. elucidation of the pathogenic mechanisms of radiation-induced heart disease and of radiation-induced vascular damage after low and moderate radiation doses at the tissue, cell and molecular level (the primary objective of this project). In the context of this experimental study, low radiation doses are defined as average heart doses between 0.01 and 0.5 Gy, while moderate doses as average heart doses between 0.5 and 5 Gy
2. determination of the radiation dose dependence of the severity, latency, and rate of progression of cardiovascular radiation damage
3. elucidation of the interaction between radiation-induced cardiovascular damage with other risk factors e.g. elevated cholesterol levels for atherosclerotic vascular damage and for microvascular damage
4. clarification of the histopathological and biochemical (proteomic) development of cardiovascular radiation damage, in particular elucidation of differences in responses at high and low radiation doses
5. development of a dose specification system for inhomogeneous dose distribution in the heart and the cardiovascular system, which will be based on the identification of critical subvolumes in the heart and their anatomical distribution in the organ.
Project results:
The CARDIORISK project addressed both macrovascular and microvascular radiation damage after local irradiation. Different groups were developing and implementing various experimental methods and models making use of the same cells and tissues most of which were centrally prepared and provided to all members of the research consortium. This way, results obtained in this research programme apply to identical biological material, treated identically with radiation. In vitro assays and in vivo experiments were all based on cells, tissue or animals from two genetically different mouse strains: Apoliprotein E (ApoE)-/- mice deficient in ApoE and prone to the development of atherosclerosis and C57Bl6 mice as normal counterpart. Molecular and cellular responses at different times can thus be related directly to histopathological and functional changes of the irradiated cardiac microvasculature or the irradiated arteries.
Development of experimental methods
The major tasks of the CARDIORISK consortium within the first two years were the development and establishment of experimental systems to study the radiation biology of cardiovascular effects. Both in vivo and in vitro model systems were deployed to describe and quantify the effects of ionising radiation occurring at low doses.
Local irradiation of the heart, the a. carotis and the A. saphena, partners TUD, NKI
Small animal irradiation setups were developed to reproducibly irradiate either the heart, the a. carotis with the aortic arch or the peripheral A. saphena. Irradiation consisted of doses ranging from 0.2 to 16 Gy. Animals or heart tissue were then distributed to all consortium members for further analysis on comparable biological material. As radiation-induced cardiovascular disease represents a late occurring event, animals were followed up to 60 weeks and investigated at different time points after irradiation (20, 40 and 60 weeks).
In Dresden, a total of 1648 animals were irradiated locally to the heart at doses between 0 (sham irradiation) and 16 Gy. Before irradiation, the correct position of the hearts was verified by digital radiographs, resulting in a total body dose of 4 mGy. Subsequently, single local heart doses of 0.2 2, 8 or 16 Gy were applied. A control group received sham irradiation with 0 Gy. Live animals or heart preparations were then distributed between the partners for the respective experiments. An additional 291 animals were irradiated locally to the heart at the NKI (doses of 0, 2, 8 and 16 Gy). These animals were used for assessment of cardiac function, prior to evaluation of morphological damage and changes in gene and protein expression, by NKI and CARIM.
Preparation of cardiovascular endothelial cells, cardiomyocytes and tissue for functional assays, partners USFD, MSCCI, TUM
As endothelial cells (EC) are among the putative target cells, methods were developed to extract primary EC from non-irradiated and irradiated hearts after different radiation doses (0.2-16 Gy) and different follow-up times.
Cardiac endothelial cell extraction methods were developed and optimised by three partners. They consistently yielded about 100 000 cells / heart after different radiation doses and at different time points after irradiation up to 68 weeks.
Development of ex vivo angiogenesis assays, partner USFD
Cardiac injury triggers repair processes in the myocardium that are dependent on neo-vascularisation. Angiogenesis is therefore central to cardiac repair. To determine whether radiation damage modulates the capacity of cardiac tissues to repair and re-vascularise, novel assays/approaches to determine 'angiogenic' activity of cells derived from the mouse heart at predetermined time points after in vivo irradiation were developed by USFD:
1. In the first approach, explants of both atrial tissue and ventricular myocardium were established in three-dimensional fibrin gels and outgrowth of endothelial sprouts from these explants were used to assess endothelial sprouting after irradiation Each heart section was dissected and around 50 tissue explants about 1 mm3 in size were embedded in a fibrin-gel and overlaid with growth medium, incubated for 10 days and then assessed for extent of sprouting using a semi-quantitative approach. Explants with no sprouts were scored as zero. Explants with a few sprouts were scored as one or two, while explants with extensive sprouting were assigned scores of three or four.
In a second approach, a 'novel' fibroblast-endothelial self-assembling angiogenesis assay was developed and used to monitor cardiac endothelial cell migration, and remodelling into capillary-like structures among fibroblasts/ myofibroblasts/ pericytes. Hearts of irradiated animals were enzymatically digested into a single cell suspension, counted and equal numbers of viable cells plated in 24-well cluster plates. After 7 to 10 days, cultures were fixed and stained with lectin to visualise endothelial cells forming capillary like structures surrounded by fibroblasts/pericytes.
In a third approach, a co-culture model for cardiac EC (CEC) and cardiac fibroblast extracted from (sham-)irradiated mouse hearts was developed to investigate the influence of fibroblast on the angiogenic properties of CEC.
Fibroblasts were plated, allowed to grow for 7 - 10 days to establish post-confluent conditions. Fibroblasts were then irradiated. Endothelial cells were plated on these irradiated fibroblast beds and allowed to grow for 10 days in cultures after which cultures were fixed and stained with lectin to visualise endothelial cells forming capillary like structures.
In vivo imaging of cardiovascular function, partners TUD, NKI, CARIM
To study cardiovascular changes in vivo over the whole follow-up period of 60 weeks within the living animals, non-invasive imaging methods were implemented.
For imaging of the peripheral A. saphena, a new imaging technology was validated by TUD called optical coherence tomography (OCT) which allows structural as well as functional vessel imaging in vivo at specific time points. With OCT imaging, arterial diameter of A. saphena under normal conditions, after vasoconstriction and vasodilatation at day one and three months after irradiation with 0-16 Gy were investigated. The left A. saphena was locally irradiated with 2 to 10 Gy and investigated 3, 6, 9, and 12 months after irradiation to validate the suitability of the method for investigations of late functional changes in arteries exposed to different radiation doses. The arteria of the right leg served as an individual control. OCT after local application of saline served as a control as well. Vasodilation and vasoconstriction was pharmacologically induced. The resulting two dimensional (2D) OCT cross sections of the A. saphena are used to define the respective inner arterial diameters under the different conditions.
For imaging of the heart, single photon emission computed tomography (SPECT-CT) was validated and used as well as ultrasound imaging. With these imaging modalities cardiac blood volume, end diastolic and systolic volumes, ejection fraction and fractional shortening were measured, as relevant cardiac function parameters.
Development of dynamic adhesion assay, partner QUB
The dynamic adhesion assay was further developed by creating a sealed system that allows to carry out the whole experiment under sterile conditions. A preconditioning approach is now used under flow conditions, which leads to re-organisation of the cellular cytoskeleton and pre-orientation of the cells in the direction of the flow. Significant challenges still remain with non-uniform adherence of monocytes and currently work is continuing to improve this.
Investigation of macrovascular effects
Structural changes, partners TUD, NKI
A. carotis: Atherosclerotic changes are observed at high doses only and are mouse strain dependent. Wild type C57Bl6 mice did not develop atherosclerotic lesions within a 30-week follow-up period after carotid artery irradiation (below, at and above the bifurcation) irrespective of dose. Around 20 % of mice developed early fatty streaks at 30 weeks after 14 Gy, but no mature atherosclerosis was observed.
Local high dose (8 and 14 Gy) irradiation to the carotid region of ApoE-/- mice resulted in more atherosclerosis and an inflammatory lesion phenotype in the long term ( around 30 wks follow-up). At 22-30 weeks after carotid artery irradiation of ApoE-/- mice there was a two-fold increase in the number of atherosclerotic lesions and total plaque burden after 8-14 Gy, compared with age and sex-matched controls. The majority of lesions in the irradiated carotid arteries of ApoE-/- mice irradiated with 14 Gy were granulocyte rich, with thrombotic features (iron containing macrophages, thrombin deposits, endothelial cell bands indicative of stimulated angiogenesis), whereas these features were much less common in age-matched controls or after irradiation with 8 Gy. In the short term (1-4 wks), 14 Gy also resulted in a higher incidence of fatty streaks. On the other hand, a lower dose of 2 Gy did not affect lesion size or phenotype.
Heart
There were no obvious atherosclerotic changes in after irradiation of hearts of C57Bl6 mice. However, at 20 weeks after 16 Gy to hearts of ApoE-/- mice there was foam cell accumulation in the endocardium of 17/20 hearts examined, versus 1/14 in age matched controls. There were also atherosclerotic lesions in the coronary arteries of 3/15 of the hearts irradiated with 16 Gy, compared with 1/14 in controls. Atherosclerotic changes at longer follow-up times are currently being evaluated.
A. saphena: No atherosclerotic lesions regardless of radiation dose and time of follow-up. In contrast, no atherosclerotic lesions in A. saphena of C57Bl6 or ApoE-/- mice were observed irrespective of dose and interval after irradiation. For detection of fatty streaks, fat uptake and foam cells oil red and CD45 stainings were conducted in all 10 weeks animals. Some fat vacuoles were detected in Tunica intima in four animals (two of the 2 Gy and the 16 Gy group). CD45 staining showed no monocytes in any layer of the A. saphena. Hematoxylin - Eosin staining showed no pathological changes after irradiation (e.g. edema, plaques). Circumference of A. saphena was measured in 18 animals and varied by 10 % without effect of irradiation. CD45 stainings and circumference measurements will therefore not be followed up in the main cohort.
In vivo optical imaging of vascular function, partner TUD
For imaging of the peripheral A. saphena , the OCT imaging technology was validated which allows structural as well as functional vessel imaging in vivo at specific time points.
Radiation does not alter the functional compliance of the A. saphena
Independent of post-irradiation interval and dose, the diameter at maximum vasoconstriciton was about 50 % of the initial control diameter. After vasodilation, the diameter increased to 1.5x the control diameter, again independent of time-interval and dose. Therefore, no radiation induced changes in functional compliance of the A. saphena were observed. Clearly higher vasodilatatory values are found for ApoE-/--mice at 3 months, compared to C57BL/6-mice. No leg contracture was observed.
The role of inflammatory, adhesive and thrombogenic responses, partners TUD, NKI, UL, UROS
Investigation of A. carotis, NKI:
To investigate whether the increased incidence of fatty streaks (short term) and inflammatory lesion phenotype (long term) is caused by upregulation of inflammatory molecules directly after irradiation, expression of several inflammatory and thrombotic molecules were analysed up to 60 weeks after irradiation using immunohistochemistry and micro-array analysis.
Inflammatory markers were examined, namely P-selectin; intercellular adhesion molecule one (ICAM1); vascular cell adhesion molecule one (VCAM1); thrombomodulin (TM); endothelial oxide synthase (eNOS); monocyte chemoattractant protein one (MCP1); tissue factor (TF); endoglin; von Willebrand factor (vWF) and plasminogen activator inhibitor-1 (PAI-1). The amount (percentage area positive for signal) and intensity of the inflammatory proteins in the endothelium of the carotid arteries were quantitatively analysed at three positions: above, around and below the bifurcation on the carotid arteries.
Expression pattern of inflammatory and prothrombotic factors is independent of structural changes
There was significantly less VCAM1 expression in unirradiated carotid arteries of C57Bl6 mice than in ApoE-/- mice. For the other markers, there were no base line differences between the strains.
One day after 14 Gy irradiation CD31 expression was significantly decreased in the ApoE-/- mice. There were no significant differences in the percentage of EC expressing ICAM1, VCAM1 or TM, and no differences in TF expression in the arterial wall. 1 week after irradiation with 14 Gy, but not lower doses, the percentage of EC expressing ICAM1 was reduced in both strains of mice and VCAM1 was reduced in ApoE-/- mice only. Four weeks after irradiation with 14 Gy the expression of TF and TM was increased in ApoE-/- mice but was reduced in C57Bl6 mice. Irradiation did not cause significant changes in expression of CD31, eNOS, MCP1 or endoglin in the carotid artery at one or four weeks. Eight Gy irradiation did not result in changes in expression of ICAM1, VCAM1, TM, TF, one and four weeks after irradiation in the C57BL6 mice.
Messenger ribonucleic acid (mRNA) analyses:
The first analyses have been performed on the ApoE-/- carotid arteries at one and four weeks after irradiation with 2, 8 and 14 Gy. TM and MCP1 were the only genes that were significantly upregulated at four weeks after 8 Gy.
Investigations of A. saphena, partners TUD, UL, UROS:
Sections of irradiated A. saphena were prepared for immunohistochemical analysis of thrombotic and inflammatory changes in C57/Bl6 and ApoE-/- mice 3, 6, 9, 12, 18 months after irradiation. Twelve different markers were tested for suitability and six (VCAM, E-Selectin, MCP1, TM, CD31 and ICAM) were finally chosen for evaluation. Six markers were not included in the final studies (CD45, Thy-1, Alkaline phospahtase), as the tunica intima is negative for these markers, vWF, KC and TF stained too weak for quantitative evaluation despite of testing different antibodies and protocols.
Strain specific expression of inflammatory and prothrombotic factors without structural atherosclerotic changes
Local increases of CD31 mean fluorescence in C57/Bl6 mice six months after irradiation were observed. CD31 showed only local significances and a tendency to decrease three months after irradiation in Bl6.
ICAM1 mean fluorescence intensity showed increases at early time points (three and six months Bl6; three months ApoE) already at intermediate doses higher than 2 Gy, whereas after nine months no changes were found.
In ApoE-/- mice VCAM showed a partial increase at three months post radiation mainly at 16 Gy. C57/Bl6 mice showed a higher and dose- dependent increase also at three months and additionally at six months. The total area of VCAM-expression showed a tendency to expand at later time points and higher doses.
In ApoE deficient mice local increases in total area of TM 6 and 12 months after radiation exposure were found. In C57/Bl6 fluorescence intensity partially increased 6 months post irradiation. In ApoE-/- a dose-dependent three months post radiation could be revealed. Irradiation with 2 Gy increased the fluorescence intensity whereas 8 and 10 Gy resulted in decreased fluorescence intensities.
MCP-1 positive area showed high variation but no significant and consistent changes. No significant effects in E-Selectin expression in both ApoE and Bl6 were detected.
Investigation of microvascular effects in irradiated heart
In vivo functional imaging of microvascular perfusion, partners NKI, CARIM
Novel in vivo imaging methods (micro SPECT-CT, ultrasound) were used for evaluation of cardiac blood volume and function. One of the tracers used (myoview) can be used to measure changes in cardiac microvascular perfusion in humans undergoing SPECT-CT. We had hoped to use this tracer quantify focal and/or overall reduction of blood perfusion in the irradiated mouse heart. However, this turned out not to be possible due to the extremely high heart beat rate in the mouse. We therefore had to rely on invasive methods, i.e. immunohiostochemistry after injection of fluoresceine-isothyocynate (FITC)/lectin, for evaluation of perfusion in the mouse heart.
SPECT-CT using the blood pool tracer Tc99m-HSA, performed at NKI, showed that cardiac blood volume of C57Bl6 mice decreased significantly at 20 weeks after 2 to 16 Gy (18 to 22 %), with a further decrease (20 to 36 %) at 60 weeks after two to eight Gy.
Gated SPECT-CT using the Tc99m-myoview (at NKI) demonstrated a significant reduction in end diastolic volume (EDV) and end systolic volume (ESV) and increased ejection fraction (EF) at 20 and 40 weeks after irradiation.
Ultrasound measurements (at CARIM) on C57Bl6 mice also showed significantly decreased EDV and ESV volumes and increased EF at 20 weeks after 16 Gy, but these parameters, as measured by US, had largely normalised by 40 weeks. Cardiac function of ApoE-/- mice was measured at 20 weeks after 0 and 16 Gy, by both ultrasound and gated SPECT (myoview). Both analyses indicated very similar reductions in EDV and ESV and increased EF in ApoE-/- mice as in C57Bl6 mice at 20 weeks after irradiation.
Similar reductions in EDV and ESV were seen in mice irradiated in Dresden and assessed by ultrasound investigations (at TUD). In these studies there were no consistent changes in EF in irradiated mice, however, there was a trend for increased EF at 20 weeks after irradiation in all dose groups.
Morphometry of microvascular density, partners TUD, NKI, UL:
Three groups investigated capillary density of irradiated hearts using different methods and criteria of quantification looking at different parts of the left ventricle.
1. NKI: Microvascular density (MVD) was determined in transverse frozen sections of mouse hearts (male C57Bl6 and ApoE-/-) at 20, 40, and 60 weeks after 0, 2, 8 and 16 Gy (uniform, whole heart irradiation). An anti-CD31 antibody with diaminobenzidine (DAB) staining was used to visualize cardiac vasculature of the central or apical part of the heart. To quantify the percentage of perfused microvessels, FITC-lectin was injected intravenously (i.v.) five minutes before tissue harvesting. For quantification of microvessels, five random fields (40x objective) were photographed and a computerised morphometry system was used to quantify MVD. Vessels beneath a size of 1.5 or above 200 µm2 were automatically excluded from the measurements.
2. UL: Apical parts of transverse frozen sections from male ApoE-/- and C57/Bl6 mice were evaluated at 20, 40 and 60 weeks after 0, 0.2 2, 8, and 16 Gy. A fluorescent CD31 antibody was used to detect the vasculature and results are expressed as CD31+ events normalised per 4'-6 diamidino-2-phenolindole (DAPI) positive nuclei.
3. TUD: Formalin fixed, paraffin-embedded sections were used to evaluate the number of microvessels per cardiomyocyte in the papillary muscle. This was based on a staining of adjacent sections with CD31/DAB, to detect endothelial cells, and laminin, to distinguish the cell boundaries of the cardiomyocytes. The number of microvessels per cardiomyocyte was counted manually, at 20 and 40 weeks after 0, 8 and 16 Gy (male ApoE-/- and C57/Bl6 mice).
Dose and time dependent decrease in MVD is accompanied by decrease in perfusion, but not hypoxia. In C57Bl6 hearts of mice irradiated at the NKI, there was a transient increase in microvascular density in the central part of the left ventricle (16-20%) at 20 weeks after two to eight Gy, and a trend for a decrease after 16 Gy (no significant changes in apical regions). A significant decrease in microvascular density was observed at 40 weeks after 16 Gy and 60 weeks after 8 Gy, indicative of progressive loss of microvessels after mid to high doses. In ApoE-/- mice, capillary loss occurred earlier and at lower doses (from 20 weeks after 8 Gy).
Analysis of CD31 and FITC-lectin stained sections showed that 92 % of microvessels in hearts of control C57Bl6 mice were perfused at 20 weeks after treatment. There was no significant change in the number of perfused vessels after irradiation with 2 to 16 Gy. At 40 weeks after 16 Gy, the reduction in MVD was accompanied by a very small, but significant, decrease in perfusion of those vessels (84 % of the remaining microvessels were perfused after 16 Gy, versus 87 % in controls). In ApoE-/- mice, the reduction in MVD at 40 weeks after 8-16 Gy was also accompanied by a small reduction in the perfusion of remaining vessels (86% versus 91 % in age matched controls). However, these small reductions in perfusion did not lead to any detectable hypoxia (EF5 staining).
In C57/Bl6 mice irradiated at TUD and analysed at UL, there were no changes in CD31+ events, normalised per DAPI nuclei, in the apical part of hearts at 20 weeks after irradiation, but at 40 to 60 weeks after 8 to 16 Gy there were decreases in MVD (only significant at 40 weeks after 16 Gy). In ApoE-/- mice, there were no consistent changes in MVD.
The morphometric analysis of the microvascular density in the papillary muscle by TUD, however did not reveal any consistent dose or time dependent changes.
Loss of alkaline phosphatase already occurs at low radiation doses and precedes reduction in micro vessel density, partner NKI
To investigate whether structural changes in MVD were associated with functional changes, changes in expression of alkaline phosphatase (ALP) vWF in microvessels, evidence of vascular leakage (staining for albumin) and collagen deposition in interstitial areas was investigated by NKI.
At 20 to 60 weeks after irradiation with 2 to 16 Gy, the amount of capillary staining for ALP was reduced by up to 50 % in both strains of mice, with more pronounced response in ApoE-/- mice.
In C57Bl6 mice, there was a significant decrease (30 to 44 %) in percentage tissue stained for ALP at 20 weeks after irradiation with 8 to 16 Gy. By 40 weeks, the 2 Gy dose group also had significantly less ALP expression, indicative of further progression of endothelial damage in small blood vessel. At 60 weeks the ALP expression in irradiated groups (2 and 8 Gy) was 50 % of the mean control value, but these differences were borderline significant (p=0.05).
In ApoE-/- mice, there was an earlier reduction in microvascular density than in C57Bl6 mice. At 20 weeks after 8 to 16 Gy MVD was already significantly less than in the controls and this persisted until at least 40 weeks after 8 to 16 Gy (longer times were not evaluated). There was also an earlier, and more pronounced, decrease in ALP expression in microvessels of irradiated ApoE-/- hearts than in C57Bl6 mice. At 20 weeks there were significant, dose related, reductions in ALP after all doses and by 40 weeks the ALP levels in vessels of irradiated hearts were only 55 to 36 % of controls.
There was a significant increase in capillary vWF expression (prothrombotic) at 20 to 40 weeks after high doses (8 to 16 Gy) in both strains. This response was slightly less marked than was seen in C57Bl6 mice.
To investigate whether the structural and functional changes in the microvasculature were associated with vascular leakage, albumin deposition in the myocardium was examined. At 40 weeks, half of the C57Bl6 hearts irradiated with 2 Gy and almost all hearts irradiated with 8 to 16 Gy showed albumin in the myocardium; this was absent or very mild in controls. After 16 Gy, myocardial albumin was extensive in 5 of 11 hearts and all these animals also had diffuse amyloidosis, which was confirmed with a Congo red staining. Amyloidosis was not seen in the irradiated ApoE-/- hearts.
Structural effects:
Epicardial thickness more than doubled at 20 weeks after 8 to 16 Gy and remained increased at 40 weeks. This was associated with the presence of iron containing macrophages (indicative of previous haemorrhage).
Increased interstitial collagen deposition (indicative of fibrosis) was present in the myocardium at 40 weeks after 8 to 16 Gy and at 60 weeks after two to eight Gy. In ApoE-/- hearts, these changes were seen earlier and at lower doses.
Electron microscopy studies showed dilated mitochondria with lysis of the cristae in cardiomyocytes at 10 weeks after eight Gy. Myofibrils showed focal lysis and disrupted structure and endothelial cells were dilated. No relevant structural disturbance were seen at two Gy or lower doses.
At 40 weeks, half of the hearts irradiated with two Gy and almost all hearts irradiated with 8 to 16 Gy showed extravascular albumin in the myocardium; this was absent or mild in controls. After 16 Gy, myocardial albumin was extensive in 5 of 11 hearts and all these animals also had diffuse amyloidosis. The remaining six animals from this group exhibited a more focal amyloidosis.
There was an increase in interstitial collagen deposition from 40 weeks after 8 to 16 Gy and 60 weeks after 2 to 8 Gy in C57Bl6 mice. In ApoE-/- mice this collagen deposition occurred earlier (from 20 weeks).
Stress response and immune function in irradiated hearts, partners TUM, MSCCI, HELMUC:
Membrane Hsp70 expression only predicts short term radiation-induced stress response. The cell surface expression and secretion of hsp70 on mouse heart ECs was analysed by determining membrane Hsp70 expression density on viable mouse heart ECs. A maximum membrane expression of Hsp70 was observed already at eight weeks after irradiation which then normalised at 10 weeks. Reduced cell surface expression levels were observed thereafter until 60 weeks after irradiation.
Serum Hsp70 expression indicates long-term persistence of radiation-induced stress response. Hsp70 protein levels were determined in the serum of mice after punction of the heart. Serum analysis of Hsp70 protein levels after irradiation revealed a significant increase only 60 weeks after irradiation.
Hsp70 expression in whole heart preparations: Changes in the expression of the major stress-responsive gene Hsp70i were assessed in both cardiac endothelial cells and in a whole heart tissue up to 40 weeks after irradiation with 0.2 2, 8 and 16 Gy doses. Mice treated with heat shock were treated as a positive control.
Gene expression analysis: CEC isolated from juvenile mice were irradiated in vitro with 2 and 8 Gy, and then 24 hours after the exposure cells were collected and RNA isolated. In addition, animals were irradiated in vivo with two and eight Gy and CECs were isolated 12 and 36 hours, 6 days, 20 and 40 weeks after the exposure, and then RNAs were purified directly after the cell isolation. Only at eight Gy up-regulation of VCAM1 gene was observed early after radiation exposure. At 20 and 40 weeks of follow-up no significant difference between irradiated and sham-irradiated animals for VCAM1, E-Selectin, Bnip and Glut1 (hypoxia-related genes) expression was observed. Radiation caused up-regulation of Hsp70i gene expression, which was visible in a dose-dependent manner within 24 hours after irradiation in vitro and in vivo in CEC isolated from irradiated animals at 20 and 40 weeks after the exposure.
Proteom analysis in endothelial cells in vitro: The human endothelial cell line EA.hy926 and the human coronary artery endothelial cell line (HCAEC) were irradiated in vitro with 0.2 Gy gamma (Co-60) and analysed for proteomic changes 4 and 24 h later. In Ea.hy926 15 significantly differentially expressed proteins were identified, of which 10 were up-regulated and five down-regulated, with more than + 1.5-fold difference compared to unexposed cells. Pathways influenced by the low-dose exposure included the Ran and RhoA pathways, fatty acid metabolism, and stress response. Radiation of 0.2 and 1 Gy did not cause phosphorylation changes in the proteome of EA.hy926 cells when examined 10 min, 30 min, or 4 h later using the total cell lysate or cytosolic fraction. In HCAEC, the total number of deregulated spots after 4 and 24 h was 29 and 26, respectively. The identification of the proteins represented as deregulated spots by ESI LC-MS/MS elucidated several pathways affected by radiation including cell death, cell morphology, RhoA signalling and actin-based motility by Rho.
Inflammatory response and thrombotic changes in irradiated hearts, partners CARIM, NKI, UL, UROS, IRSN:
ApoE-/- and C57/Bl6 mice sacrificed 20, 40 and 60 weeks after irradiation with 0, 0.2 2, 8, and 16 Gy (eight animals per dose) were analysed for CD31, thrombomodulin, VCAM-endocard and CD45. Increased mean area of CD31 staining 20 and 40 weeks post irradiation may indicate endothelial swelling at high doses (8 and 16 Gy) and was found in ApoE-/- mice only. Similarly CD31 fluorescence intensity increased in ApoE-/- mice only, indicating an increased extravasation potential already after irradiation at two Gy. At high doses (8, 16 Gy) a time depend immigration of leukocytes was detected, by increase of CD45 positive cells, in both endocardium and myocardium at 40 weeks post irradiation. This behaviour is not due to the onset of spontaneous atherosclerosis in the ApoE-/- genotype because it was found a similar reaction in the C57/Bl6 wild type model. Also VCAM expression on the endocard showed a trend to increase after irradiation at 16 Gy at all time points in both mice strains. Changes in all four markers (CD31, Thrombomodulin, VCAM-endocard and CD45) indicate a proinflammatory response mostly at high doses (8 and 16 Gy). At low dose (0.2 Gy) only small changes are seen at a few time points with no significances.
A reduction of Thrombomodulin (TM) positive events was found 20 and 40 weeks (C57/Bl6), and 60 weeks (ApoE-/-) after irradiation at 16 Gy. TM exhibits anti-inflammatory, antithrombotic properties, partly by being a sink for thrombin but also by reducing the expression of adhesion molecules. Reduction of TM positive capillaries might indicate a proinflammatory response in the background of unchanged overall capillary count. The ICAM-1 staining was weak and did not appear to be influenced by irradiation.
Analyses of inflammatory gene expression
Microrarray data are only available for C57 Bl6 mice at present. There were no large changes in inflammatory gene expression at 20 weeks, but VCAM1, ICAM 1 and KFL2 were slightly increased at 20 weeks after two Gy. Pathway analysis scored high for cell-to-cell signalling pathway including CXCL1, S100A8 and S100A9. Gene expression data at 20 weeks after 16 Gy showed modest increases in endoglin and decreases in VCAM1. Pathway analysis scored high in cell death/development disorder-pathways, including MYL4, myosin, connective tissue growth factor (CTGF), PAI-1, TIMP3 and MMP.
By 40 weeks after 2 Gy, the expression levels of PAI-1, TM, ICAM1 and KLF2 were increased and pathway analysis scored high for cellular movement-immune cell trafficking pathway and molecular transport/cell cycle - pathway. These pathways include upregulation of LYZ, CCL2 and CCL4. Gene expression data from 16 Gy showed increases in Ki67, PAI-1, P-selectin and endothelin and significant down-regulation of Smad5 and PDGFb. Pathway analysis indicate high scores for cell movement/immune cell trafficking pathway including SPP1, CXCL13 and ESM1.
Using real-time PCR expression of the inflammatory genes (VCAM-1, MCP-1, E-Selectin) was investigated. Neither samples of ApoE-/- mice sacrificed 20 weeks after irradiation nor mice sacrificed 60 weeks after irradiation showed any change in expression of inflammatory marker gene VCAM or E-selectin normalised to reference genes GAPDH and ß-actin. Due to very few native material MCP-1 could not be amplified for any of the samples.
Functional integrity
Studies in EC
Experiments at USFD focused on analysing angiogenic responses in the heart tissue following irradiation with the aim to determine whether radiation damage modulates the capacity of cardiac tissues to repair and re-vascularise. Since isolation and subculture of enough endothelial cells from older mice proved challenging, alternative assays/approaches to determine 'angiogenic' activity of cells derived from the mouse heart were developed.
The results demonstrate that radiation decreases the capacity of reparative angiogenesis in heart tissues. Severe impairment of this process was statistically significant at high radiation doses while a trend for a reduced angiogenic activity is evident at moderate to low doses. Radiation at high doses (8 and 16 Gy) inhibited capillary-like tube formation in ApoE-/- mouse heart cultures at 20 weeks post-irradiation. Capillary-like structures were affected both qualitatively and quantitatively. In particular, in the irradiated groups these were less prominent than the controls and were characteristically shorter and narrower, fewer in numbers and had fewer branching points. There were no significant differences in the total number of viable cells initially extracted from irradiated and non-irradiated hearts. However, once in culture, cells from the 8 and 16 Gy irradiated groups showed a reduction in growth and viability correlating with the profound inhibition of capillary-like network formation seen at these radiation doses. A similar response was seen in C57BL/6 hearts at 20 and 40 weeks post-irradiation. A dose dependent reduction in 'angiogenic index' was observed 20 weeks post-irradiation. A similar pattern was observed at 40 and 60 weeks post irradiation. Statistical significance was achieved only for the 8 and 16 Gy dose groups.
Work at MSCCI focused on analysing irradiation-induced stress responses in cardiac endothelial cells (CEC). These included analysis of expression of stress-response genes described above as well the analysis of cytosceletal structures. Analysis of cytoskeletal structures (actin stress fibres) was analysed in CEC isolated from animals irradiated with two and eight Gy, 20 weeks after irradiation (cells isolated 40 weeks after irradiation were not suitable for analysis). Stress fibre analysis was also performed on CECs isolated from animals irradiated in Gliwice at shorter times after irradiation and on CECs from juvenile animals irradiated in vitro. Filamentous structures of actin were detected upon staining with fluorescently-labelled phalloidin.
Significant changes were observed in CECs irradiated in vitro with eight Gy. Mice were irradiated with two and eight Gy, and at 12 hours, 36 hours, 6 days and 20 weeks after irradiation endothelial cells were extracted. Irradiation with two Gy increased the number of cells with high number of stress fibres (to about 40 %); such changes persisted for 6 days after exposure but were not seen after 20 weeks. Irradiation with eight Gy resulted in a highly significant increase in numbers of stress fibres. We conclude that radiation induced dose-dependent changes in actin cytoskeleton (i.e. formation of contractile stress fibres) in cardiac endothelial cells irradiated in vivo and in vitro. Irradiation with 16Gy did not increased permeability of monolayers formed by CECs to dextran (MW 40 000 Da), neither 3 nor 24 hours after exposure. Therefore, this part of the experimental program was not pursued further.
Work at QUB focused on studying intercellular communication in endothelial cell cultures. Originally, freshly isolated endothelial cells were to be used, but this proved difficult, therefore QUB used established mouse endothelial cell lines for these studies. Previously, the characterisation of various endothelial cell lines was completed in collaboration with UL and the decision was made to consolidate work using a mouse cardiac endothelial cell line H5V thereafter shared by all members of the consortium working on endothelial cell lines. Other mouse endothelial lines used included lymph node large vessel (mlEDN1); brain microvascular (bEND3); heart microvascular (H5V) and previously heart microvascular (MHEC5-T) and lymph node large vessel (SVEC4-10). In particular, QUB focussed on the characterisation of the cytokine profile of endothelial cell lines after irradiation at 0.1 and 2 Gy both in the presence and absence of TNFa. The aim was to understand mechanisms related to the focal nature of cardiovascular disease associated with response of endothelial cells to low and targeted doses of radiation.
In addition, the impact of irradiated monocytes on adhesion properties was determined, and changes in adhesion properties, both static and dynamic, were evaluated along with cytokine changes in endothelial cell monolayers. This was extended to situations where cells are irradiated both directly and under bystander conditions.
The brain derived bEND3 cells responded to irradiation with doses between 0.05 Gy and 5 Gy with a significant increase in adhesion properties under both static and dynamic conditions. Statistically significant changes in adhesion properties were not observed in the other two cells lines While pro-inflammatory effects of radiation were observed only in bEND3 cells, in all three cell lines evidence for radiation-induced anti-inflammatory effect 24 h after irradiation of pre-activated with TNFa cells was seen. Analysis of inflammatory cytokine expression profile of H5V, bEND3 and mlEND1 cells suggests that radiation induced pro-inflammatory response can occur in all three endothelial cell lines as early as 30 minutes after irradiation in both microvascular and lymph node derived endothelial cells. Elevated levels of MCP-1 and decreased expression of transforming growth factor-beta (TGFß) persisted for more than 24 hours in pre-activated cells. An anti-inflammatory effect of radiation on pre-activated endothelial cells was also observed as early as 30 min after irradiation.
X-ray exposure with partially shielded cultures indicates a role for bystander signalling for both cytokine expression and monocyte adhesion. This is observed at 0.1 and 2 Gy and suggests intercellular communication plays an important role. Antibody blocking cell surface adhesion molecules led to significant decrease in monocyte adhesion in mlEND1 and bEND3 cells with the effects being more pronounced in pre-stimulated cells. The important role of the Nf?ß pathway in radiation-induced inflammation was confirmed by adding to the media Bay11-7085 prior to treatment with radiation or TNFa, which led to a significant decrease in adhesion.
Irradiation of the monocyte population with low and moderate doses (0.1 Gy and 2 Gy) causes decrease in the ability of monocytes to adhere to endothelial cells, suggesting potential radiation-induced anti-inflammatory effect. The effect was observed in all three cell lines - bEND3, mlEND1 and H5V.
Studies in heart tissue and cardiomyocytes, partners HELMUC, IRSN:
Mitochondrial proteom and function
The radiation-induced in vivo effects on cardiac mitochondrial proteome and function were investigated at 20 and 40 weeks after local irradiation of the heart. Myocardial mitochondria were isolated from whole heart tissue and tested for proteomic and functional alterations. The proteomic analysis was done using peptide (ICPL) and protein quantification (2D-DIGE). Altogether 18 proteins were found deregulated 20 weeks after 2 Gy-irradiation, of which 12 were up-regulated and six downregulated. Deregulated proteins belonged to Oxphos complexes, carbohydrate metabolism or mitochondria-associated cytoskeleton.
Mitochondrial respiration was measured using succinate as the substrate to analyse the intactness of mitochondria and to determine the efficiency of oxygen consumption. Succinate-stimulated respiration decreased significantly in mitochondria from mouse hearts irradiated with two Gy after 20 w and 40 w, respectively. No significant alteration was found in the succinate-driven respiration after irradiation with 0.2 Gy of C57BL/6 mice or with any dose or time point in ApoE-/- mutant. No changes in mitochondrial membrane potential were registered in mitochondria isolated from either C57BL/6 or ApoE-/- at 40 weeks using different stimuli.
Rho/ROCK pathway
Acute radiation effects on primary cardiomyocytes isolated from 12 week old C57BL6 and irradiated in vitro and long-term radiation effects in primary CM isolated from sham-irradiated and irradiated C57BL6 and ApoE-/- mice at 0.2 2 and 16 Gy; 20, 40 and 60 weeks post-irradiation were performed. Investigation of acute and delayed radiation-induced Rho/ROCK and Smad pathway activation along with remodelling of the actin cytoskeleton were performed in isolated CM using biochemical assays and immunohistochemistry. The activation of the Rho/ROCK pathway by monitoring Rho protein isoprenylation by immunoblotting. ROCK kinase was assessed by monitoring phosphorylation of its target myosin-light chain (MLC) and upregulation of CTGF, a known downstream target protein of the pathway. CTGF is also a target of the TGFß pathway together with PAI-1. Activation of the following members of the TGFß cascade was studied: TGFß1, TGFß RII, Smad2/3/4, PAI-1. Immunostaining was performed to assess radiation-induced remodelling of the alpha-actin-sarcomeric that constitutes the main cytoskeletal component of cardiomyocytes. The pharmacological inhibitor of ROCK, Y-27632, was used to modulate the phenotypical characteristics of CM isolated from animals 20 and 40 weeks post-irradiation. Pharmacological experiments (with Y-27632) could not be performed 60 weeks post-irradiation due to the fragility of the CM isolated at this time-point. In addition, planned knock down short interfering RNA (siRNA) approaches were withdrawn due to the small yield of CM isolated from irradiated animals and to their fragility.
Radiation-induced sequential activation of Smad and Rho pathways and cytoskeletal remodelling in primary cardiomyocytes:
The balance between TGFß1/Smad and Rho/ROCK signalling in response to low and high doses of irradiation was investigated in CM, 4 h and 24 h post-irradiation. Irradiation at 0.2 2 and 16 Gy induced early phosphorylation of Smad2/3 and a dose dependent increased expression of Smad4, indicative of Smad pathway activation 4 h post-irradiation. This was followed by radiation-induced increase of the isoprenylated Rho 24 h after irradiation, indicative of Rho activation. A trend for Rho activation was observed after exposure at 0.2 Gy but only reached significance for doses above 2 Gy. Consistently, a subsequent production of CTGF occurred 24h after exposure of CM to 16 Gy. Immunofluorescence studies showed a radiation-induced remodelling of the central network of a-sarcomeric actin, which is the main structural protein of cardiac muscle. Actin remodelling was observed after low dose (0.2 Gy), a dose at which Rho proteins were not significantly activated suggesting involvement of additional mechanisms. Such alteration of actin network was also visible in irradiated CM at high dose as early as 4 h after irradiation and persisted at 24h, consistent with the alteration of Rho isoprenylation observed after exposure to high doses.
Rho/ROCK pathway is involved in control of actin remodelling in primary cardiomyocytes
Actin remodelling is thought to depend upon modulation of the Rho/ROCK pathway. To confirm this assessment primary CM were exposed to the pharmacological inhibitor of ROCK, Y-27632. First, the inhibition of ROCK kinase activity upon Y-27632 incubation was monitored by studying myosin light chain phosphorylation (MLC). MLC phosphorylation decreased significantly 1 h after exposure to Y-27632 and was sustained at, 4, 6 and 1h after exposure. Consistently, alteration of the actin network associated with striation disappearance was obvious 4 h after exposure to Y-27632 and persisted at 6 and 16 h.
Primary cardiomyocytes can be isolated from irradiated mice. In order to investigate the TGFßeta1/Smad and Rho/ROCK signalling pathway as a late effect in irradiated cardiomyocytes in vivo, the isolation of CM from irradiated mice was optimised. CM was successfully isolated from non-irradiated and irradiated C57Bl6 and ApoE -/- mice 20, 40 and 60 weeks post-irradiation. 20 weeks post-irradiation, the yield, quality and shape of CM was good with good adhesion capability indicative of cell's quality. It allowed sub-culturing and pharmacological modulation. 40 and 60 weeks post-irradiation, quality and shape of CM was insufficient to provide good adhesion and perform long-term culture, therefore cells were lysed immediately after isolation.
Modulation of the TGFß cascade: In CM no alteration of the TGFß cascade was observed 20 weeks post-irradiation, regulations occurred at later time point. Immunoblot analysis showed a global activation of the TGFß cascade in CM isolated from C57BL6 that mainly occurred at late time points. TGFß increased 40 and 60 weeks post-irradiation in animals irradiated with 0.2 Gy, 2Gy and 16 Gy (40 weeks only available). Interestingly, Smad 2/3 and smad4 inductions were dose dependant 40 weeks post-irradiation. 60 week post-irradiation Smad 2/3 level remained high whereas Smad 4 dropped. Finally, the two TGFß targets studied (CTGF and PAI-1) were only moderately stimulated, mostly 40 weeks post-irradiation. Similarly to what was observed in cardiac tissue, Smad7 expression dropped 40 and 60 weeks post-irradiation. Western-blot analysis showed increased TGFß1 levels in CM isolated from ApoE -/- irradiated at low and intermediate doses. However, the protein level of canonical members of the TGFß cascade, Smad 2/3 remained stable. A trend to induction was observed for Smad4. CTGF expression was not modulated at 0,2 Gy whereas PAI-1 significantly picked at this dose and doses above. At higher doses, CTGF protein level moderately but significantly increased. Smad7 expression dropped 60 weeks post-irradiation.
In conclusion, the results show that cardiomyocyte physiology and molecular responses are altered even by low radiation doses. Differences were found between CM responses between irradiation in vitro or after cell isolation from the irradiated tissue. PAI-1 activation was observed in the both mouse strains suggesting that the TGFß/Smad/PAI-1 cascade operated in fibrogenesis. Yet amyloidosis development which was observed in C57Bl6 but not in ApoE deficient mice, was dependent of ApoE-/-, and the activation of another mechanism must be considered. The current results show that the molecular imprint at low dose is different from that obtained at moderate/high dose. Activation of the Smad2/3 pathway occurs immediately after irradiation but neither Rho pathway nor CTGF were activated after exposure at 0.2Gy. The TGFß cascade is differentially activated in both mouse strains. Smad 2/3/4 seemed to operate in C57Bl6 whereas other protein seemed involved in ApoE. This difference may explain the differential pathogenic picture observed in histology. However, the Rho pathway was not found activated in CM despite immunohistochemical pattern.
Final conclusions
The results of the CARDIORISK project permit the following conclusions which may impact profoundly future research in the area of low-dose radiation risk assessment in the cardiovascular system:
1. radiation at low doses does not per se induce atherosclerotic changes in medium to large arteries
2. inflammation does not play a major role in the development of radiation induced cardiovascular disease (CVD); stress responses can be regarded as a consequence of the effects of radiation induced CVD
3. the reduction in microvessel density and the inhibition of neoangiogenesis even at low to intermediate doses does not directly translate into the clinical picture of CVD, but it may well compromise the ability of the heart to respond to subsequent stress
4. these major conclusions support the hypothesis that low-dose radiation exposure and related subtle subclinical changes are permissive factors for the development of CVD which reduce the capacity of the heart to recover if a clinically relevant event/ insult occurs.
Outlook and future research
The CARDIORISK identified the progressive reduction of the capillary network at intermediate and low radiation doses (less than two to eight Gy) as the key mechanism leading to radiation-induced heart disease. This microvascular effect does not lead to measurable ischemic cardiomyopathy but is likely to reduce the compensatory capacity of the heart to insults from other sources such as coronary heart disease etc. In addition, there are also strong suggestions of direct myocardial injury by low radiation doses after similar latencies. Despite the wide range of research methods and biological targets used in CARDIORISK, some key molecular and cellular processes of this microvascular effect are still largely unknown. Future research should focus on resolving these open questions, in particular:
1. the molecular targets in capillary endothelial cells which lead to functional damage, e.g. loss of alkaline phosphatase. Since CARDIORISK provided evidence that endothelial cells isolated from different organs and different endothelial cell lines show very different responses to low dose radiation exposure those studies should use predominantly endothelial cells isolated directly from the hearts of the investigated animals
2. the role of radiation-induced reduction of alkaline phosphatase in capillary endothelial cells and their role for capillary function
3. the molecular mechanisms of low and intermediate dose radiation damage to cardiomyocytes
4. the mechanisms of progression of general and focal loss of alkaline phosphatase and possible consequential random destruction of entire capillarie
5. elucidation of the apparent intercellular communication processes between endothelial cells within a capillary for the development of capillary damage
6. the dependence of reduction of the capillary network on dose (less than two Gy) and time (more than one year) in longer living animals
7. the relationship of radiation-induced microvascular radiation damage on the topographical dose distribution in the hearts, on age at exposure with regard to different clinical manifestations of radiation-induced heart disease, in particular ischaemic heart disease, valvular disease and conduction defects.
For this research, an integrated programme bringing together cardiological and radiobiological scientists is obligatory. The experience of CARDIORSIK has demonstrated the great potential of this cooperation.
Potential impact:
Socioeconomic impact and the wider societal implications of the project
The CARDIORISK consortium has developed a series of experimental systems to study the radiation biology of cardiovascular effects at low doses in vivo and in vitro. The major goal was to provide the impetus necessary to place radiation protection considerations of late cardiovascular effects in a biological context.
The estimation of radiation risk is primarily based on the assumption that, except for heritable, genetic radiation damage, the only somatic radiation damage to be considered in radiation protection is radiation-induced cancer. This dogma, which went unchallenged for several decades, has been challenged by the epidemiological evidence showing radiation-induced mortality from cardiovascular and cerebrovascular diseases of similar magnitude as from radiation-induced cancer. These epidemiological data demands a reassessment of the concepts, quantities and methods of how risk is defined in radiation protection. In order to incorporate cardiovascular and cerebrovascular radiation risks into the overall system of radiation protection, knowledge is required on pathogenic mechanisms, which only radiobiological experiments can provide.
The main problems for the incorporation of the new epidemiological evidence of radiation-induced late cardiovascular and cerebrovascular damage arise in the definition of dose and the shape of the dose response relationship. In the present system of radiation protection, the effective dose is defined as the sum of the mean radiation doses in the respective organs at risk, which is multiplied with the tissue weighting factor. While the tissue weighting factor may be relatively straightforward to define from epidemiological studies - and it is obvious that the tissue weighting factor for the heart and the brain will have to be increased significantly from their present value of 0.025 the question of dose definition cannot be resolved by epidemiological data. The answer to this problem has to be based on scientific evidence which requires full knowledge on the pathogenesis of the respective radiation-induced fatal diseases including in particular information on the critical structures which trigger the pathogenic pathways and their anatomical distribution in the organ. Only radiobiological experiments such as those performed in the CARDIORISK project were seen to put the methods of how to define radiation dose in the heart in situations of inhomogeneous radiation exposure -which are the rule in particular in diagnostic radiology - on a sound basis of scientific evidence. The present system of radiation protection assumes proportionality between mean organ dose and the risk of radiation-induced cancer since cancer is likely to be a clonal disease, arising from a single stem cell that has been transformed and clonally expanded through many additional steps of progressing malignancy. In contrast to this, cardiovascular diseases and cerebrovascular diseases are examples of obvious multicellular origin. All scientific bodies have consistently argued that the distinction between stochastic and deterministic radiation effects can be based on the distinction of single cell clonal effects, justifying a linear dose risk relationship, and multicellular radiation effects, which is observed only after radiation exposure has exceeded a threshold. This distinction can no longer be upheld in view of the epidemiological evidence on cardiovascular radiation mortality of the A-bomb survivors. The resolution of this problem is of great importance for radiation protection since the entire concept of no-threshold linear dose risk relationships and the concept of stochastic radiation effects has been put into the limbo.
The CARDIORISK project has produced an enormous amount of biological data which relate to the pathogenesis of macrovascular and microvascular radiation damage. There is evidence in the mouse cardiovascular system that that large vessels such as carotid artery, arteria saphena and probably coronary arteries follow different pathogenic pathways but which are all associated with inflammatory processes which, however, are only triggered by high radiation doses which are characteristic of radiotherapy. In contrast, microvascular radiation damage is elicited already by lower radiation doses and develops independent of pro-inflammatory changes which lead to structural changes in the capillary network in the myocardium. This damage starts with functional changes in the capillary endothelial cells causing progressive disappearance of entire capillaries. There is some evidence for a simultaneous stress response in the irradiated myocardial cells although it remains open whether both effects develop independently or one is the consequence of the other.
For macrovascular radiation effects in the heart the results of the CARDIORISK project suggest a threshold-type dose response relationship which depends on the radiosensitivity of several well investigated processes involved in the activation of the inflammatory conversion of age-related atherosclerosis which occur after shorter latencies than the microvascular changes. The target structure in the heart where these processes take place and for which, therefore, the specification of local dose has to be defined are the coronary arteries, and in particular the left anterior descending coronary artery as has also been suggested by the findings of the RACE project.
The CARDIORISK project produced evidence that microvascular radiation damage is different following different pathogenic kinetics in different parts of the ventricular myocardium. These topographical differences may be due to pathophysiological factors or to the different response criteria used by the different CARDIORISK partners. More studies need to be performed on the mechanisms and modifying factors which influence the reduction of the microvascular network. The data produced by CARDIORISK suggest that the reduction and rarefication of the capillary network is progressing over much of the life-span and a linear dose response relationship cannot be excluded. The linear dose dependence, though only validated in the dose range of two to eight Gy, is difficult to reconcile with the observation that functional alterations in endothelial cells and rarefication occur throughout much of the myocardium of the ventricles, and that severity of this effect progresses with dose and time. This puts a big question mark on the concept of linear dose response relationships in general and clonal origin of 'stochastic' radiation damage and may have implications for the entire system of radiation protection in all areas of low dose radiation exposure including the out-of-field exposures in radiation oncology. These findings also suggest that the whole myocardium should be delineated as the sub-volume for which dose should be defined and reported which would be, for microvascular damage, the mean myocardial dose.
In conclusion, different doses to the heart may have to be considered in radiation protection, i.e. the mean dose to the myocardium, and in radiation oncology where in addition, the local doses to the coronary arteries have to be determined and minimised as well.
Impact on other Euratom projects
Close cooperation with the integrated Seventh Framework Programme (FP7) project NOTE which includes studies of cardiovascular damage after low dose and protracted total body irradiation had been agreed. The leader of work package two (WP2) of NOTE was a member of our consortium. He used irradiated hearts, arteries and endothelial cells from the CARDIORISK project and investigate their response with the same methods he used in the hearts of the NOTE project in order to compare the cardiac effects of total body irradiation with those of localised heart irradiation. Moreover, the leader of WP4 of NOTE, who aimed at using the data produced in the NOTE project to develop a bio mathematical model of radiation-induced cardiovascular risk, was invited to all CARDIORISK meetings and had direct access to all results produced by the CARDIORISK project to incorporate them into the bio mathematical model of radiation-induced cardiovascular.
Main dissemination activities and exploitation of results
All results will finally be published in the scientific open literature. However, equally important is the direct input of the results of the described studies and our conclusions from the new data into the discussions of the international committees responsible to adjust the radiation protection concepts, rules and regulations to the progress of knowledge. So far, results from the CARDIORISK project have been presented at 16 national and 55 international meetings and published in 30 scientific articles.
As this project addressed long-term cardiovascular effects, final pivotal publication can only now be prepared after all follow-up times have been analysed. From every partner at least one seminal publication is planned for publication. To specifically address the radiation biology and therapy community, a dedicated issue of 'Radiotherapy and oncology' about the CARDIORISK project, the results and conclusions is planned. In a peer-review process one paper from each partner is being considered for submission together with an introductory editorial. A review paper which described the overall design of the CARDIORISK project was written by the scientific secretary and published in 2010 in 'Radiotherapy and oncology'.
By far the most important dissemination activities during the lifetime of the project were:
1. the 'DoReMi' exploratory workshop on 'Radiation-induced cardiovascular disease from low dose exposure' in November 2010 in Bombon, France
2. the CARDIORISK symposium in June 2011 which heralded the successful conclusion of the project.
The other five talks were given by the leading experts in the field of cardiovascular research in the context of radiation exposure, radiation therapy and chemotherapy coming from specialties like radiation oncology, radiation biology, oncology, cardiology and epidemiology presenting complimentary information from experimental and in particular clinical data. All talks highlighted the role of the CARDIORISK project for planning future research in radiation-induced cardiovascular diseases.
In addition, other knowledge dissemination activities aimed to ensure dissemination and training to graduate and undergraduate audiences, other Euratom contracts and the public and stakeholders.
Project website: http://www.cardiorisk.eu