A Dose-Response Study

Cells : A Dose-Response Study -Particle Delivery on Vascular Smooth Muscle Cells and Endothelial β Cellular Effects of Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1999 American Heart Association, Inc. All rights reserved. is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231 Circulation doi: 10.1161/01.CIR.99.11.1477 1999;99:1477-1484 Circulation. http://circ.ahajournals.org/content/99/11/1477 World Wide Web at: The online version of this article, along with updated information and services, is located on the

I t is well established that restenosis after angioplasty is characterized by significant neointimal formation and geometric remodeling of the artery. 1 Since the stent-based strategy is effective in preventing vessel shrinkage, excessive neointimal proliferation through the metallic prothesis continues to reduce lumen size. 1 To prevent neointimal hyperplasia, various therapeutic strategies have been developed on the basis of pharmacological approaches, gene transfer, 2,3 and more recently on ionizing radiation (␤and ␥-emitting sources). 4,5 Several animal studies with either ␤ particleemitting stents 6 -8 or wire-mediated endovascular ␤ 9,10 and ␥ emissions 10 -13 have shown great promise in the field of restenosis prevention. ␤-Radioactive stents allow a long-term delivery of ␤ particles within the arterial wall with a low dose rate, whereas endovascular brachytherapy is characterized by high dose rates over a short-term period. 4,5 Although there is a great interest in developing an effective clinical endovascular radiation strategy to prevent restenosis, 14 -16 several questions pertinent to the use of such strategies remain unanswered. Indeed, the cellular consequences of ionizing radiation leading to the prevention of neointima formation after balloon angioplasty remain to be elucidated.
Accordingly, we developed a strategy combining radiotherapy and gene transfer approaches to effect an in situ delivery of pure ␤ particles in vascular cells in a homogeneous manner. We propose in the present study to investigate in vitro the effect of ␤ emission on (1) proliferation of vascular smooth muscle cells (VSMCs) and endothelial cells (ECs), (2) cell cycle progression, (3) cytotoxicity and apoptosis induction, (4) cell migration, (5) vascular cell phenotype, and morphology features. We demonstrate in vitro the impact of local ␤ irradiation on both proliferation and migration of vascular cells in a dose-dependent manner and describe for the first time the interference of ␤ radiation with VSMC cell cycle progression, phenotype modulation, and absence of cytotoxic effects.

Cell Culture
All experimental protocols were approved by the Institutional Committee for Animal Protection of the Centre Hospitalier de l'Université de Montréal. Porcine VSMCs were isolated from aortic explants of Yorkshire Swine (Charles Rivers, St Constant, Québec, Canada) as previously described. 17 Human VSMCs were harvested from unused specimens of saphenous veins obtained after bypass surgery, with informed consent of patients. Cell purity was estimated to be Ͼ95%, with the use of VSMC ␣-actin immunostaining (Sigma). Human coronary artery ECs (Clonetics Corp, San Diego, Calif) were grown in MCDB 131 medium (Sigma) and U937 human cells (monocyte-like leukemic tumor, ATCC, Rockville, Md) in RPMI medium.

Dosimetry for 32 P-Labeled Oligonucleotides
In vitro-predicted dose was calculated considering both the specific activity of 32 P-labeled oligonucleotides and the exposure time according to the dose-point-kernel (DPK) prediction. The DPK estimation, derived from the Monte Carlo method, was based on the homogeneous distribution of radioactive particles in the medium. 18 On the basis of autoradiograms, ␤-particle delivery was shown to be uniform in the 2-cm 2 well, and no radiation was detected in adjacent wells (data not shown).

Local ␤-Particle Delivery
We used 32 P-labeled DNA oligonucleotides as a molecular delivery mode of ␤ particles to vascular cells. In the present in vitro report, the sense sequence of c-myc was used to ensure that the observed effects were not sequence specific but mainly the result of ␤-particle emission. Previous studies conducted to compare sense and antisense sequences reported that the efficacy of both 32 P-labeled sequences was statistically similar in inhibiting VSMC proliferation (unpublished data, 1996). Sense c-myc phosphorothioate oligonucleotides (5Ј-ATGCCCCTCAACGTG-3Ј) were synthesized on a DNA/RNA automated synthesizer (Applied Biosystems) and purified according to the supplier's protocol (Glen Research). Oligonucleotides were internally labeled with [ 32 ␥ATP] (3000 Ci/mmol, Du Pont Canada Inc) by T4 polynucleotide kinase (Gibco-BRL). The stability of the labeled oligomer was assessed by incubating VSMCs with the probe (0.4 nmol/L) for 1 to 7 days. Extracts were loaded on 20% urea-polyacrylamide gel and revealed by autoradiography. To determine the cellular localization of the radioactive source, VSMCs were transfected with oligomer 5Ј-labeled with fluorescein-isothiocyanate molecule (FITC, 2 mol/L).

External ␥and Electron-Particle Deliveries
To compare the low dose-rate strategy obtained with the use of labeled oligonucleotides to a high dose-rate approach, we exposed VSMCs, ECs, and U937 cells to either 60 Co (Cobalt-60) or electron beam sources, leading to ␥-(1.6 Gy/min) or electron-particle (4 Gy/min) deliveries, respectively (3, 10, 20, and 30 Gy).

VSMC Viability
Cell integrity was evaluated by the trypan blue exclusion test (Gibco-BRL). 19 Increasing ␤ doses were tested (1, 3, 5, 10, 20, and 30 Gy) by incubating VSMCs with labeled-oligonucleotides for 72 hours. Cell viability assay was performed in triplicate on 4 cell preparations and expressed as a percentage of total cell counts.

Cell Proliferation Assay
VSMCs and ECs were synchronized with serum-deprived medium for 48 hours and then stimulated with FBS in the presence of increasing levels of 32 P-labeled oligonucleotides (0.4 to 10 Gy). Nonirradiated cells received equal concentrations of unlabeled oligonucleotides (Ͻ20 nmol/L). After [methyl-3 H]-thymidine incorporation (6.7 Ci/mmol, NEN Life Science Products), cells were treated as previously described. 20 To evaluate the cell regrowth after removal of the ␤ source from the medium, VSMCs were first incubated with labeled oligonucleotides (5.5 Gy), leading to severe inhibition of proliferation (80%), then the ␤ source was removed from the cell layer for 3 and 5 days. The effects of local irradiation on proliferation was calculated as a percentage of that obtained with nonirradiated cells, on the basis of quadruplicates with 5 to 7 cell preparations. Because similar effects of labeled oligonucleotides on porcine and human VSMCs were seen (Figure 3), we then decided to perform all the following experiments on porcine VSMCs because porcine explants are more readily available.

Internucleosomal DNA Fragmentation
VSMCs, ECs, and U937 cells were incubated with labeled oligonucleotides for 72 hours (1, 3, 5, 10, 20, and 30 Gy). The 3 cell types were also exposed to external 60 Co and electron beam sources (3, 10, 20, and 30 Gy) and processed 24 and 72 hours after irradiation. As a positive control of DNA fragmentation, we treated U937 cells with 1 mol/L camptothecin, a DNA topoisomerase I inhibitor (Sigma) for 6 hours. Genomic DNA was extracted from treated cells as described. 21

Transmission Electron Microscopy
VSMCs, ECs, and U937 cells were incubated with labeled oligonucleotides for 72 hours (3 and 30 Gy). Cells were fixed in Millonig's sodium phosphate solution containing 2.5% glutaraldehyde and stored in 0.1 mol/L Millonig's solution until processing (50 to 70 nm thick). Transmission electron microscopy was performed with a Zeiss Em10 CA microscope (JFE Enterprises).

Cell Migration and Immunofluorescence
VSMC and EC migration assays were performed with modified Boyden chambers (Neuroprobe Inc) in which cells were allowed to migrate through a polycarbonate membrane (8-m pore size, Costar Corp). 22 Immunostaining was performed to investigate effects of ␤ irradiation on cell phenotype by incubating cells with either anti-␣actin or anti-von Willebrand factor antibodies (Sigma) for VSMCs and ECs, respectively. 23 After nuclei staining with 0.01% Hoechst solution (Sigma), slides were observed with an epifluorescence microscope. Low ␤-irradiation doses (0.4 and 1 Gy) were used to evaluate the setting of migration and phenotype modulation.

Statistical Analysis
All values are expressed as meanϮSEM. Differences between control and irradiated cells were assessed by ANOVA, with subsequent Student-Newman-Keuls test for multiple comparisons. Statistical significance was established at a value of PϽ0.05.

Local ␤-Particle Delivery
We used a phosphorothioate-modified oligonucleotide because its high stability has been previously demonstrated. 24 The labeled oligomer showed no evidence of radiolysis or degradation within the media even after 7 days (Figure 1), demonstrating the high stability of 32 P-labeled oligonucleotides in vitro. As illustrated in Figure 2A, Ͼ95% of cells were positive for fluorescent label, showing that the transfection was highly effective and uniform in vitro. FITC oligomers were localized mainly in the nucleus in accordance with previous studies. 24,25 In agreement with the DPK prediction, 18 no radiation was reported in adjacent wells ( Figure 2B); the level and the distribution of ␤-particle emission were reinforced by autoradiography analysis (data not shown). Together, these results demonstrated that stable labeled oligonucleotides homogeneously emit ␤ particles.

Vascular Cell Proliferation
The effect of unlabeled oligonucleotides (Ͻ20 nmol/L) on cell proliferation was evaluated to be a reduction of Ͻ10%, in accordance with previous studies. 20,26 For the first time, we report a significant inhibition of VSMC proliferation with ␤ irradiation in a dose-dependent manner (0.4 to 10 Gy, Figure  3A), with an ED 50 at 1.10Ϯ0.07 Gy and 1.08Ϯ0.12 Gy for porcine and human, respectively, an ED 50 statistically similar in both species. Increasing ␤ dose induced comparable patterns of growth inhibition in ECs, with an ED 50 at 2.15Ϯ0.10 Gy ( Figure 3B). However, ED 50 comparison of VSMC and EC demonstrated that the ability of ␤ irradiation to inhibit proliferation was significantly higher in VSMCs compared with that with ECs (PϽ0.01), except at the highest dose (10 Gy). Removal of 32 P-labeled oligonucleotides did not completely abolish the potential inhibitory effect of ␤ irradiation on VSMC proliferation over 3 to 5 days (PϽ0.05 vs nonirradiated cells, Figure 4).

Cell Cycle Progression in VSMCs
As shown in Figure 5, serum-deprived cells were mainly distributed in the G 0 /G 1 phase of the cell cycle (57%Ϯ2%, Figure 1. Stability of 32 P-oligonucleotides in the presence of VSMCs 1 to 7 days after transfection. The 15-mers internally labeled conserved the 32 P isotope within the sequence. At a lower level, labeled oligomers of 14-mers were also detected, corresponding to the ligation of two 7-mers with an internal 32 P within the sequence. panel A) with reduced DNA synthesis activity (32%Ϯ2%) and low level of cell division (8%Ϯ1% in the G 2 /M phase). Between 3 to 20 hours after serum activation, quiescent cells did not progress into the S phase (data not shown). As previously described, 27 24 hours after serum addition, we observed that most of VSMCs progressed into the S phase (85%Ϯ3%, Figure 5B), and to a lesser extent, cells remained in the G 0 /G 1 state. Interestingly, VSMCs remained in the G 0 /G 1 phase after ␤ exposure (2, 10, and 30 Gy), with a concomitant reduction in cell cycle progression into the S phase ( Figures 5 and 6). ␤ irradiation did not modify significantly the G 2 /M phase of the cell cycle ( Figure 6). As illustrated in Figure 5, no apoptotic peak corresponding to DNA degradation was noted at low or high ␤ doses. Investigation of long-term effects of ␤ exposure on cell cycle progression showed that cells remained mainly in the G 0 /G 1 state (57%Ϯ3%) for up to 120 hours.

Comparative Effects of Local and External Irradiation on DNA Fragmentation
Labeled oligonucleotides at low (3 Gy) and elevated dose ranges (10, 20, and 30 Gy) did not induce DNA fragmentation in VSMCs, ECs, and U937 cells ( Figure 7A). Exposure to 60 Co or electron beam sources did not alter DNA integrity in VSMCs and ECs, whereas this type of delivery induced DNA laddering in U937 cells (Figures 7B and 7C). These data suggest that U937 cells may be more sensible than vascular cells after high dose-rate radiation. Camptothecin also led to severe DNA fragmentation in leukemic cells after 6 hours ( Figure 7A, lane 4), as previously reported. 21

Morphological Changes and Apoptosis
Electron micrographs showed that after labeled-oligonucleotide irradiation (3 and 30 Gy), VSMCs, ECs, and U937 cells conserved normal ultrastructural features ( Figure 8). Presence of apoptotic cells was identified by the following features: chromatin compaction, cell surface blebbing, cytoplasmic shrinkage, and condensation. Morphological studies reported that the frequency of apoptotic cells in control and ␤-irradiated VSMCs, ECs, and U937 cells wasϽ3%, whereas after camptothecin treatment, 50% to 60% of leukemic cells met the apoptosis criteria. 21

Cell Migration and Phenotype
Our results reported similar inhibitory effect in migration of VSMCs (43%Ϯ0.4% and 20%Ϯ3.5% of control cell level) and ECs (47.3%Ϯ2.5% and 13%Ϯ2.3% of control cell level) for doses of 0.4 and 1 Gy, respectively (PϽ0.001 vs control cells). Proliferating VSMCs showed a weak immunostaining of ␣-actin proteins, corresponding to the synthetic phenotype ( Figure 9A). Local ␤ irradiation induced an increase in ␣-actin fibers that appeared to be dose dependent ( Figures 9B  and 9C) and occurred in nearly all cells (Ͼ95%, panel D), suggesting the appearance of a contractile phenotype in irradiated VSMCs. With ECs, ␤ irradiation (0.4 and 1 Gy) did not modify von Willebrand factor immunostaining (data not shown).

Discussion
The development of endovascular radiotherapy initiated a series of seminal experiments demonstrating the effectiveness of a ␥ source to prevent restenosis over shortterm 9,10,12 and long-term observation periods. 9,11,28 Verin et al 29 and Waksman et al 10,30 subsequently demonstrated the effectiveness of a ␤ source in reducing intimal hyperplasia in rabbit and pig models, respectively. Stent-based approaches were also proposed as platforms from which ␤ emission could be delivered to the vessel wall. 6 -8 In a landmark trial, Teirstein et al 14 reported the successful reduction of restenosis rates in injured arteries with iridium-192. Although the concept of radiotherapy to prevent restenosis recently reached the clinical arena,  several questions regarding its mechanism of action on vascular cells remain unanswered. Therefore, we proposed in this study to examine the consequences of ␤ emission radiation on various fundamental functions of vascular cells such as cell viability, proliferation, and migration.
We also dissected the effects of ␤ radiation on cell cycle progression, apoptosis induction, phenotype modulation, and morphological features in VSMCs and ECs.
To deliver ␤ particles in a homogenous fashion, we developed a novel approach based on the local delivery of radiolabeled short DNA single strands. We demonstrate that oligomers internally labeled with phosphorus-32 ( 32 P) provided a stable source of radiation and an uniform dose distribution to target cells. Comparison of the effects of sense and antisense sequences demonstrated that both labeled sequences had similar inhibitory effects, supporting the concept that the oligonucleotides only served as a molecular vector to locally deliver ␤ particles. The 32 P-isotope (half-life 14.3 days) has already been adopted by 2 different platforms in coronary radiation clinical trials (wire based and stent based).
To date, few data are available on the biological effects of radiation on vascular cells. Fischell et al 31 observed that 32 P-radioactive wires inhibited the growth and migratory functions of VSMCs by use of a nonuniform dose distribution. A dose-response effect of a cesium-137 source on  VSMC proliferation was reported and showed "reproductive cell death" without, however, any sign of apoptosis. 32 The current study reports for the first time that ␤ irradiation efficiently inhibits human and porcine VSMCs and human ECs in a dose-dependent manner. At a moderate range of ␤-particle delivery (0.4 to 6 Gy), ECs appeared to be more radioresistant than VSMCs. No difference, however, could be seen at doses that induced optimal proliferation inhibition. Protocols used in current clinical trials of brachytherapy 14 -16 thus would not be expected to have a different effect on VSMCs and ECs. VSMC growth inhibition was maintained over at least 5 days after radioactive source removal, suggesting either that the effects of ␤ irradiation are sustained even after source removal or that remaining incorporated intracellular oligonucleotides continued to inhibit proliferation. 24 To understand the mechanisms of the inhibitory effect of radiation, we investigated cell cycle progression in irradiated VSMCs. For the first time, we report that ␤ irradiation caused cells to remain in the G 0 /G 1 phase and reduced cell numbers in the S phase, as leukemic cells after radiotherapy. 33 Moreover, this growth arrest was maintained over 5 days. The mechanisms of ␤ irradiation-induced G 1 arrest in VSMCs remain, however, to be defined. Recent recombinant adenoviral strategies reported that adenovirus containing p21, 27 p53, 34 or retinoblastoma 35 efficiently blocked the cell cycle progression and successfully reduced intimal hyperplasia in animals. It may be hypothesized that p53-induced p21 inhib-its the cyclin-dependent kinases, 36,37 leading to the G 1 arrest in irradiated VSMCs.
Previous studies on the effects of external radiation using high dose-rates reported damage leading to "reproductive cell kill" 32,33 or apoptosis induction. 33 However, "reproductive cell kill" is not synonymous to the loss of physical integrity. The current low dose-rate ␤-emission approach did not induce signs of loss of organelle integrity or DNA fragmentation, as shown by ultrastructural and FACS analyses. To compare this low dose-rate approach with that used in brachytherapy, we exposed cells to external sources of ␥ or electron particles. High dose-rate irradiation did not induce DNA laddering in VSMCs and ECs, whereas leukemic cells appeared to die by radiation-induced DNA fragmentation. This is in accordance with the low level of apoptosis found in rabbit ␤-irradiated arteries 29 and the absence of apoptosis reported in porcine irradiated arteries. 10 The low dose-rate delivery effected by labeled oligonucleotides (2 to 10 cGy/h) is more closely related to that used with radioactive stents 6 -8 than that used in brachytherapy (1 to 5 Gy/min). 9,10,12,13,28 Consequently, such low dose-rate strategies could be defined as cytostatic, preserving cell integrity and viability and acting as a growth suppressor. 37 Also, no apoptosis could be seen in vascular cells in our high dose-rate experiments. Whether "physical cell kill" is a desirable end point in the objective of preventing clinical restenosis remains unanswered.
In addition to inhibiting cell proliferation, ␤ irradiation significantly reduced migration of VSMCs and ECs, at much lower ␤ doses (1 Gy) than those affecting the proliferative properties, suggesting that the migratory function may be  , ECs (D, E, and F), and U937 cells (G, H, and I) were exposed to labeled oligonucleotide (0, 3, and 30 Gy) for 72 hours. For comparison, arrow represents typical apoptotic cell in H and I (Ͻ3%), characterized by chromatine condensation within intact nuclear membrane (arrowhead), whereas normal cells had homogeneous chromatine (n) (magnification ϫ6000). more radiosensible than cell growth. ␤ irradiation inhibited EC migration but failed to alter von Willebrand factor immunostaining, suggesting that reduced migratory function of irradiated ECs may be responsible for potential endothelium-related dysfunction by altering reendothelialization of injury site after balloon angioplasty. VSMC functions are closely related to their phenotype modulation, 38 and in proliferative disorders such as restenosis, contractile VSMCs become dedifferentiated and develop the synthetic phenotype. This immature stage, similar to the in vitro conditions, is correlated with a high migratory capacity of VSMCs. 38 We demonstrate that low ␤ irradiation induced the transition from the synthetic to the contractile phenotype as shown by the higher ␣-actin immunostaining in ␤-irradiated cells.
In conclusion, we report that labeled oligonucleotides inhibited proliferation of VSMCs and ECs in a cell-specific and dose-specific manner by acting as a growth suppressor, resulting in G 1 arrest, with no morphological changes and apoptosis induction. ␤ radiation succeeded in inhibiting cell migration and enhanced the contractile phenotype of VSMCs. By acting on vascular cell function, the ␤-radioactive strategy may contribute to prevent intimal hyperplasia after angioplasty. Figure 9. Immunostaining of ␣-actin in ␤-irradiated VSMCs (left) with 0 (A), 0.4 (B), and 1 Gy (C) during 72 hours. Nuclei were stained with Hoechst assay (corresponding right panels). D represents superposition of ␣-actin and nuclei staining, demonstrating that Ͻ95% of irradiated cells were altered (1 Gy) (magnification ϫ40).