Ionizing radiation enhances IL-6 and IL-8 production by human endothelial cells

Irradiation exposure is known to induce an inflammatory reaction. Endothelial cells play a crucial role both in the inflammatory process and in radiation damage. Therefore, supernatants and cell lysates of 60Co-irradiated human umbilical vein endothelial cells (HUVEC) have been assessed for the presence of pro-inflammatory cytokines. After gamma irradiation, interleukin (IL)-1α, IL-1β and tumor necrosis factor (TNF)-α remained undetectable in both cell supernatants and cell lysates. However, a dose-dependent increase in the production of IL-6 and IL-8 has been demonstrated up to 6 days after exposure. These data indicate that the pro-inflammatory cytokines IL-6 and IL-8 may be involved in the inflammatory response of vascular endothelium induced by exposure to ionizing radiation.


Introduction
The in ammatory process is a classical pathophysiological response to ionizing radiation. In ammatory lesions have been observed locally in a number of tissues, such as skin, intestine or lung and in a wide range of doses (between 5 and 40 Gy). Recent studies have shown early changes such as an increase in the number of adherent and emigrated leukocytes. [1][2][3] An early and persistent cytokine production following local exposure of rat intestine 4 or lung 5 has been observed and related to the late appearance of brosis.
In addition to these local effects, different investigators have shown the release of proin ammatory cytokines in peripheral blood. In humans, a total body irradiation of 10 Gy results in a rapid increase in the serum levels of TNF-a and IL-6, with a maximum reached 4 h after radiation exposure. 6 These increases were substantial but transient, and returned to baseline levels 24 h after irradiation. Circulating IL-1 and IL-6 have also been measured shortly after exposure to UV radiation. 7,8 However, the cellular source of the cytokines was not determined. Exposure of baboons to mixed neutron ±c eld irradiation resulted in an early and transient increase in the serum level of IL-6, followed in some cases by a secondary release of IL-1, IL-6 and IL-8 a few days before the animals died; 9,10 this suggests a fatal prognosis associated with in ammatory response.
The production of pro-in ammatory cytokines by irradiated cells has also been investigated in vitro. Elevated levels of TNF-a and IL-1 have been found after irradiation of various human or mammalian cells, such as alveolar macrophages or tumour cells. 11,12 X-ray exposure induced an overproduction of IL-6 by broblasts 13 and of both IL-6 and IL-8 by glioma cells. 14 Increase in IL-6 and IL-8 production has also been detected after UV exposure of keratinocytes. 15,16 In all cases, these increases occurred early and the authors did not investigate the effects of irradiation later than 72 h.
Endothelial cells (ECs) play a crucial role in the initiation, development and maintenance of the in ammatory response. During the in ammatory process, leukocytes accumulate in the damaged tissue after transendothelial migration mediated by a cascade of events involving pro-in ammatory cytokines, chemokines and adhesion molecules. The major mediators of these processes are IL-1 and TNF-a. They appear to be produced mainly by activated monocytes/macrophages and, in turn, activate both these cells and vascular endothelial cells. This activation involves different steps where adhesion molecules are expressed, and where chemoattractants are released so allowing the transendothelial migration of leukocytes into the underlying tissue. 17 -19 Both IL-6 and IL-8 are involved in the in ammatory response. They have been detected in the serum of patients with in ammatory disorders 20 -22 and can be produced in response to various damaging stimuli such as bacterial or viral infections, 23,24 or burns. 25 IL-8 is an important in ammatory mediator. It is a potent chemoattractant primarily for neutrophils but also eosinophils and a subset of lymphocytes. 26,27 IL-6 is a pleiotropic cytokine and substantial evidence exists about its in ammatory effects such as induction of acute phase proteins by the liver. 28 However, IL-6 has also been considered as an anti-in ammatory cytokine because of its ability to induce IL-1 and TNF-a antagonists. 29 Evidence exists for the involvement of ECs in the radio-induced in ammatory response. An increase in leukocyte adhesion to irradiated endothelium has been observed in vitro, and is at least partly explained by the induction of some adhesion molecules, such as ICAM-1, in ECs from dermal microvasculature. 30,31 Data obtained by other investigators have shown the release of chemoattractants by irradiated bovine endothelial cells, 32,33 and suggest a biphasic response involving a lipid chemoattractant shortly after irradiation followed by the release of a protein attracting neutrophils. The chronic in ammatory lesions, commonly found after irradiation, could involve dysregulation of the vascular endothelium. 34 The objective of this study was to investigate the in vitro radio-induced release of the proin ammatory cytokines IL-1a, IL-1b, TNF-a, IL-6 and IL-8 by HUVEC, to establish if these proin ammatory cytokines might be linked to radio-induced endothelial cell damage. Both time-course and dose ±response studies have been carried out. Furthermore, since TNF-a has been shown to be released after radiation exposure, HUVEC have been irradiated both in the presence and in the absence of TNF-a. In addition, to advance the understanding of the mechanism of action of ionizing radiation, we investigated the possible role of intracellular oxidants in the regulation of cytokine production by HUVEC, and by using Northern blot analysis we determined the effect of ionizing radiation on cytokine gene expression.

Materials and Methods
Cell culture HUVEC were obtained from the ATCC and routinely cultured in gelatin-coated asks in F12K medium (Sigma, France) supplemented with 20% fetal calf serum (Gibco-BRL, France), endothelial cell growth supplement 60 mg/ ml (Sigma, France), glutamine (Gibco-BRL, France), heparin 100 mg/ ml (Sigma, France) and antibiotics. HUVEC are near-diploid non-transformed cells. In our culture conditions, cells reached con uency within 6 days. Cells were used between the sixth and the ninth passage after ATCC freezing.
For irradiation experiments, 1·2 3 10 5 cells were plated on 35 mm diameter Petri dishes (Falcon, Becton-Dickinson, France) in the same medium as described above. Cells were activated with TNF-a 10 U/ ml (speci c activity 5 3 10 7 U/ mg; R&D Systems, UK) and simultaneously irradiated 24 h after plating (control cells were neither activated with TNF-a nor irradiated). This concentration of TNF-a was chosen because it does not induce maximal production of IL-6 and IL-8. Supernatants were harvested and centrifuged to remove cell debris at various times after irradiation and frozen at -808 C until use. At each time point, the number of living adherent cells was evaluated using nigrosine exclusion after extensive washing of the cell monolayer. To evaluate the metabolic activity of irradiated cells, TNF-a was re-added to the 10 Gy-irradiated cells 6 days post-exposure, and cytokines were measured in the supernatant 48 h later.
To determine if the effect of c irradiation could occur through the production of a soluble mediator by the irradiated cells, the following protocol was established: cells were irradiated 24 h after plating, and the supernatants were collected 72 h after irradiation. In parallel, cells were seeded according to the standard protocol, and 24 h after plating the medium was replaced by a 50:50 mixture of fresh medium/ medium from the irradiated cells or the nonirradiated control cells. Concentrations of IL-6 and IL-8 were determined by ELISA of the cell supernatant 72 h later.
Cell lysates were performed as follows: after trypsinization, cell pellets were resuspended in 0.5% Triton X100, centrifuged 14 000 3 g for 30 min, and then the supernatant was assessed for cytokine content.

Irradiation
Proliferating cells were irradiated with a 60 Co source (ICO 4000) at a dose rate of approximately 1 Gy/ min. The dose range tested was 1 ± 20 Gy. The sham irradiated controls were treated under the same conditions.

Antioxidant effect
We explored the role of oxidative stress in the response of HUVEC to ionizing radiation by the use of the antioxidant N-acetyl-L-cysteine (NAC, Sigma, France). Cells were treated with NAC (20 mM) 1 h before irradiation or TNF-a activation. After 72 h cell supernatants were collected and assessed for the presence of IL-6 and IL-8.

Cytokine immunoassays
A double-antibody sandwich ELISA for the quantitative determination of the cytokines in cultured media was performed according to the manufacturer's recommendations (Amersham, France). The limits of detection were 3·9 pg/ ml for IL-1a and b, 15·7 pg/ ml for TNF-a, 3·13 pg/ ml for IL-6 and 31·3 pg/ ml for IL-8.

Northern blot analysis
Total RNA was extracted according to the method of Chomczynski and Sacchi, 35 sizefractionated on a 1% formaldehyde ±agarose gel (15 mg/ lane), transferred to a Gene Screen nylon membrane (Dupont de Nemours, France) and xed by heating (808 C, 3 h). Membranes were prehybridized for 3 h at 428 C and then hybridized overnight in hybridization buffer containing the human cDNA probes (33 10 6 cpm/ lane) labeled with 32 P-adCTP using the Megaprime DNA labelling system (Amersham, France). After hybridization, the membrane was washed. The quanti cation of radioactivity was made using the automated system InstantImager TM (Packard, France) after which the membranes were exposed to Hyperlm-MP lm (Amersham, France) at -808 C. IL-6 RNA expression was assessed using a 700 bp Xmn I ±B an I fragment and IL-8 using the R&D systems probe. Equal loading was analyzed by probing with a 1 kb HindIII fragment of the human glyceraldehyde-3-phosphate dehydrogenase (GAPDH, ATCC) cDNA cloned in the pBR322 vector.

IL-6 bioassay
IL-6 activity was assessed using the 7TD1 cell line 36 obtained from ATCC. Brie y, cells were cultured in a 96 well-dish (1000 cells/ well) in Dulbecco's modi ed medium (Gibco-BRL, France) supplemented with 10% fetal calf serum, glutamine, penicillin and streptomycin in the presence of serial dilutions of the supernatants to be tested or of human recombinant IL-6 (rhIL-6, R&D systems, UK; speci c activity 1·5 3 10 6 U/ mg). The proliferative response of 7TD1 cells was assessed after 72 h of culture using the (3-4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide colorimetric assay as described elsewhere. 37 One unit of IL-6 activity was de ned as 50% of maximum proliferation obtained with rhIL-6. The detection limit of the test was 0·1 U/ ml. An IL-6 antibody (R&D Systems, England) was added to the culture medium containing the HUVEC supernatants at saturating concentration, in order to determine the speci city of the mitogenic activity.

Statistical analysis
Statistical signi cance was calculated using the Mann ±Whitney rank sum test. P values , 0.05 were considered as signi cant.

Effect of c irradiation on endothelial cell growth
In order to assess the effect of c irradiation on HUVEC growth, the viability and the number of cells were determined at various times after radiation exposure in the presence or absence of TNF-a (data not shown). Viability of the adherent cells was always found to be . 90%. However, a decrease in the number of cells was observed as early as 24 h post-exposure for the 5 and 10 Gy irradiated cells and 48 h for the 2 Gy irradiated cells. For example, 2 days and 6 days after irradiation, there were about 3-fold and 4.5-fold less cells respectively in the 10 Gy irradiated asks than in the non-irradiated control asks. The same observations were made with or without TNF-a. In addition, a noticeable change of the adherent cell morphology was observed; hypertrophic cells emerged by 2 days post-exposure.
Effect of irradiation on pro-in¯ammatory cytokine production by HUVEC IL-1a, IL-1b and TNF-a were detectable in neither the cell supernatant nor the cell lysate of control or c-irradiated HUVEC cultures at any of the times tested.
Irradiation induced a higher production of IL-6 in both TNF-a stimulated and non-stimulated cells. This increase in IL-6 was both dose-and time-dependent (Fig. 1). The enhanced IL-6 production increased with the dose of irradiation, and the maximal effect was obtained later with increasing doses: the plateau of IL-6 was reached within the rst 24 h for the 2 Gy irradiated cells (no statistical difference was detected after that time) and 4 days postexposure for the 5 Gy irradiated cells. No plateau was observed following a 10 Gy irradiation. Similar results were obtained when the cells were irradiated in the presence of TNF-a (Fig. 1B) except that the maximal production of IL-6 was detected 3 and 4 days after, respectively, a 5 Gy and 10 Gy exposure. Gamma exposure of HUVEC induced changes in the production of IL-8 similar to that noted for IL-6 ( Fig. 2A and 2B). However, the effect of irradiation on IL-8 production was more pronounced than the effect on IL-6 production. The mean increase in IL-8 was 33.5-fold 6 days after irradiation whereas it was only 9.5-fold for IL-6. It is interesting to note that the effects of c irradiation and TNF-a are synergistic. HUVEC irradiated at 10 Gy in the presence of TNF-a produced 6 days post-exposure, 136-and 36fold more IL-8 and IL-6 respectively than the control cells.
To verify that the increased levels of IL-6 and IL-8 found in the cell supernatant of irradiated cells did not result from the release of the cytokine by damaged cells, we determined the amount of IL-6 and IL-8 present in the cell lysate for various times after irradiation. Table 1 shows the results obtained when cells were irradiated in the presence of TNF-a. Similar results were obtained in the absence of TNF-a except that IL-8 remained undetectable in the lysate of both irradiated and non-irradiated cells at all time periods tested. In the presence of TNF-a the amount of IL-6 and IL-8 present in the cell lysates remained at a level much lower than that in the supernatants. However, increases in the IL-6 and IL-8 levels were found in the cell lysates after c irradiation, suggesting a de novo synthesis rather than a release of preformed protein. To determine if the effect of c irradiation was a direct effect or occurred through induced production of a soluble mediator by the irradiated cells, we cultured non-irradiated cells in the presence of the supernatant from irradiated cells. No induction of IL-6 or IL-8 was observed after 72 h of culture as compared to the cells incubated with the supernatant of non-irradiated cells, suggesting that c irradiation does not act through the release of a molecule into the cell supernatant (data not shown).

Effect of N-Acetyl-cystein e (NAC) on IL-6 and IL-8 production
The use of an oxygen radical scavenger in HUVEC cultures had no effect on the radioinduced increase in production of IL-6 and IL-8, whereas the up-regulation of the two cytokines following TNF-a activation of irradiated or nonirradiated cells was partially inhibited ( Table 2). A 2-fold decrease in IL-6 and IL-8 production was observed after NAC treatment of TNF-a activated cells. However, the antioxidant was more effective when the cells were both activated with TNF-a and irradiated with three times less IL-6 and IL-8 in the supernatants of NAC treated cells as compared with control cells.

Northern blot analysis
In order to determine if the irradiation effect occurred at the transcriptional level, we performed Northern blot analysis. HUVEC were either activated with TNF-a (10 U/ ml) or irradiated (10 Gy). Although IL-6 and IL-8 transcripts were seen 4 h after TNF-a stimulation, they remained at a basal level up to 24 h after a 10 Gy irradiation (Fig. 3).

Biological assays for IL-6 determination
The biological activity of IL-6 produced by HUVEC after irradiation was assessed using the IL-6 dependent cell line 7TD1. Fig. 4B represents the dose-response obtained after culture of 7TD1 cells for 72 h in the presence of supernatant obtained as described in Material and Methods; a dose dependent increase in IL-6 activity was observed. Furthermore, we ob-tained a strong correlation (r = 0.98) between biological activity determined by 7TD1 proliferation assay and the quantitative determination of IL-6 by ELISA (Fig. 4A). The speci city of the assay was determined by using monoclonal antibodies, which completely blocked the mitogenic activity of the HUVEC supernatants (data not shown). Thus, IL-6 produced by irradiated cells seems to be as active as the protein produced by non-irradiated cells.

Discussion
In the present study, we investigated the involvement of in ammatory cytokines in the response of ECs in vitro to irradiation exposure. Our ndings showed that IL-1a, IL-1b and TNFa were not present in the supernatant of HUVEC irrespective of the culture and irradiation conditions, whereas IL-6 and IL-8 levels were increased up to 6 days after c irradiation in a dose-and time-dependent manner. A 2 Gy irradiation is suf cient to induce a signi cant increase in cytokine production under some conditions. After a 10 Gy-irradiation, the enhancement of IL-6 and IL-8 production is progressive and persistent. It appears that this longlasting effect is speci c to c exposure and is not a result of cell damage because the level of cytokines found in the cell lysates remains very low; this suggests that the release of cytokines from the cells could not contribute to the increased levels found after irradiation in the cell supernatant. As shown by others, TNF-a stimulated IL-6 and IL-8 production by HUVEC; 19 this up-regulation occurred early and reached a plateau within 2 days. In contrast, the cytokine-induced production following irradiation was progressive, and the maximum effect was not reached until 6 days after exposure to a dose of 10 Gy. Furthermore, our results demonstrate that not only IL-6 and IL-8 are induced by irradiation in the presence of TNF-a but also that the effects of TNF-a and c irradiation are synergistic. Since TNF-a is likely to be present after irradiation, the synergistic phenomenon should also occur in vivo.
It has recently been shown that ionizing radiation can induce the transcription factor NFkB through a reactive oxygen intermediate signaling pathway. 38 Activation of NF-kB has also been involved in the overexpression of IL-6 observed after radiation exposure. 13 In addition, it has been demonstrated that treatment with hydroxyl radical scavengers decreased IL-8 production in different cellular models whereas it had no effect on IL-6. 39 TNF-a is known to stimulate the release of radical oxygen intermediates from a variety of cell lines. 40 Therefore, we tested the hypothesis that the increase in IL-6 and IL-8 production following TNF-a activation or radiation exposure of HUVEC would result from the increase in intracellular oxidants. Our results showed that treatment with NAC did not have any signi cant effect on the radio-induced production of IL-6 and IL-8 although it did result in a partial inhibition of the effect of TNF-a alone or associated with a 10 Gy-irradiation. This suggests that the synergy between TNF-a and c rays involves an oxidant dependent pathway.
By Northern blot analysis we showed that TNF-a activation led to IL-6 and IL-8 up regulation whereas a 10 Gy exposure did not; this suggests that transcriptional activation is not the main regulatory mechanism of c irradiation in HUVEC.
Different explanations of the radio-induced increase in IL-6 and IL-8 production may be proposed. Firstly, the increased production of these cytokines in the supernatants of irradiated cells could be the result of leakage from damaged cells. However, we demonstrated that the IL-6 and IL-8 content are signi cantly lower in the cell lysates compared to the supernatant. Secondly, the increase in IL-6 and IL-8 production could result from an indirect effect of ionizing radiation through the release of a molecule in the supernatant of irradiated cells. As discussed earlier, TNF-a and IL-1 are not responsible for such an effect, since none of these factors were found in either the cell supernatant or in the cell lysate. Furthermore, when non-irradiated cells were fed with supernatants obtained from irradiated cells, no increase in the production of IL-6 or IL-8 was observed. These data suggest a direct effect of c irradiation on cytokine production. Finally, a decrease in the number of receptors present at the cell surface, a decrease in the release of the soluble form of the receptor, or a change in the af nity of the receptors could be induced by irradiation. This could explain the radio-induced increase in IL-6 production but not the increase in IL-8 because a recent study has shown the absence of IL-8 receptors at the surface of HUVEC 41 and, to our knowledge, no soluble form of the IL-8 receptor has yet been described. We determined that the soluble form of IL-6 receptor (IL-6sR) was not spontaneously released into the supernatant of HUVEC (data not shown), therefore, the increase in IL-6 cannot be explained by a decrease in IL-6sR. Further studies are necessary to evaluate the effect of irradiation on the expression of the membranebound component of the IL-6 receptor. IL-6 and IL-8 may constitute the key effectors of the radio-induced response of endothelial cells. Through their multiple biological activities and in particular their respective induction of ICAM-1 42 and CD11b/CD18, 43 they may participate in the in ammatory reaction following ionizing radiation as described by Panès et al. 3 We have also observed the upregulation of ICAM-1 expression in HUVEC by c irradiation in correlation with an increased adhesion of neutrophils on irradiated HUVEC (data not shown).
In conclusion, we have shown that IL-6 and IL-8 are involved in the in ammatory reaction of endothelial cells following c irradiation, whereas TNF-a and IL-1 do not seem to be responsible for the in ammatory response. The increase in cytokine production was observed for doses as low as 2 Gy, and up to 6 days after exposure. Our results suggest the early and persistent effect of c irradiation on the endothelium and the involvement of IL-6 and IL-8 in the development of the late phase of the radioinduced in ammatory response.