Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
(CANCER RESEARCH 51, 2552-2558, May 15, 1991] Autocrine Effects of Fibroblast Growth Factor in Repair of Radiation Damage in Endothelial Cells' Adriana Haimovitz-Friedman, Israel Vlodavsky, Alpana Chaudhuri, Larry Witte, and Zvi Fuks2 Departments of Radiation Oncology; Memorial Sloan-Kettering Cancer Center, New York, New York (A. H-F., Z. F.); the Department of Oncology, Hadassah-Hebrew University Hospital, Jerusalem, Israel (I. V.); and the Department of Medicine, Columbia University, College of Physicians and Surgeons, New York, New York (A. C., L. W.) ABSTRACT The study demonstrates that basic fibroblast growth factor (bFGF) serves as an inducer of radiation damage repair in bovine aortic endothelial cells (BAEC). Radiation dose-survival curves were generated with plateau-phase BAEC using culture dishes precoated with HR9-bFGF/ extracellular matrix (ECM) for the postradiation colony formation assay. This natural basement membrane-like ECM is enriched with ECMbound bFGF. Under these conditions the cells exhibited increased repair of radiation damage as compared to cells plated on top of the bFGF-free isotype of this extracellular matrix (the HR9/ECM). While the slopes of the curves did not differ significantly (Do 107 ±6.8 cGy on the HR9/ ECM, compared to 112 ±1.3 cGy on the HR9-bFGF/ECM), there was a nearly complete elimination of the threshold shoulder in the curves generated on the bFGF-free HR9/ECM (Dq 29 ±19 cGy, compared to 174 ±22 cGy on the HR9-bFGF/ECM; P < 0.05). Delayed plating experiments, in which the cells were irradiated under bFGF-free condi tions (while adherent as contact-inhibited monolayers to the HR9/ECM in bFGF-free medium) and maintained after irradiation in the same culture for various periods of time, showed that the cells performed repair of potentially lethal damage (PLDR) and restored clonogenic ability, with a 24 h to immediate postradiation recovery ratio of 3.27. This expression of PLDR was inhibited by neutralizing monoclonal antibodies against bFGF, indicating that the irradiated cells secreted bFGF into their conditioned medium. Northern blot hybridization showed a 5.6-fold increase of the 3.7-kilobase species and a 4.7-fold increase of the 7.0kilobase species of the bFGF-specific mRNA within 6 h after delivery of a single dose of 400 cGy. The data suggest that radiation induces a complete cycle of an autoregulated damage-repair pathway in BAEC, initiated by radiation-induced damage to cellular DNA and followed by stimulation of bFGF synthesis and its secretion into the medium. The newly synthesized bFGF stimulates the PLDR pathway, acting via an extracellular autocrine loop (inhibitable by specific anti-bFGF antibod ies), leading to recovery of cells from radiation lesions and restoration of their clonogenic capacity. INTRODUCTION The current hypothesis concerning the nature of the lethal lesions of ionizing irradiation in eukaryotic cells suggests that the loss of the clonogenic capacity and cell death result from damage to the structure and function of genomic DNA (1, 2). Although mammalian cells are proficient in their capacity to repair radiation damage to DNA (3), not all radiation lesions undergo repair (2). Some lesions are irreparable and are consid ered lethal (4, 5), while others are potentially repairable (6-9), but their fate depends upon competing processes of repair and lethal fixation (10-12). Thus, the proportional loss of the clonogenic capacity at a given radiation dose results from the production of nonrepairable lesions and from potentially re pairable lesions that undergo lethal fixation (10, 12). The net Received 11/30/90; accepted 3/7/91. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1Supported by NIH Grant CA-52462. ; To whom request for reprints should be addressed, at the Department of Radiation Oncology, Memorial Sloan-Kettering Cancer Center. 1275 York Ave., New York, NY 10021. effect of increasing doses of radiation on the clonogenic capacity is classically described by dose-survival curves (13), which typ ically consist of an initial curved component at the low dose range (the threshold shoulder) and of a subsequent exponential component. Several biophysical models simulate the cell sur vival data, and parameters derived from the mathematical func tions that describe these models are used to define the sensitivity of mammalian cells to radiation and their potential for repair of radiation damage (10, 14, 15). Parameters that describe the exponential components are commonly used to define the in herent radiosensitivity of a given cell, while those parameters that describe the curvilinear threshold shoulder reflect the ca pacity to repair radiation lesions and to restore clonogenicity. While there are methods to quantitatively evaluate the repair capacity of radiation-induced damage, there is little information concerning the enzymatic pathways involved in repair of radia tion lesions in eukaryotic cells and the genetic and epigenetic mechanisms that control and regulate this cellular function (16). Potentially lethal radiation damage repair3 has been re ported to be an inducible process (17-20) and sensitive to changes in the microenvironmental conditions of the culture. Conditions leading to reduction of the metabolic rate of cells or in their proliferative activity increase PLDR, while hypertonic media or chemical inhibitors of DNA synthesis promote potentially lethal damage fixation and decreased repair (9, 11). In recent experiments4 we showed that, when BAEC were seeded after irradiation on top of their autologous natural basement membrane-like ECM (the BAEC/ECM), they exhib ited an improved capacity for repair of radiation damage com pared to cells plated on other natural or artificial substrata. Because the BAEC/ECM contains endogenous heparan sulfatebound bFGF (21), these experiments raised the possibility that bFGF may be involved in regulation of radiation damage repair in this cell type. To evaluate this possibility, in the present study, we generated radiation dose-survival curves for BAEC, using culture dishes coated with two isotypes of ECM produced by variants of the PF-HR9 mouse endodermal carcinoma cell lines for the postradiation colony formation assays (22). We recently showed (23) that the wild-type PF-HR9 cells produce an FGF-free ECM (the HR9/ECM), while the HR9-bFGF cell line, which was cloned from PF-HR9 cells transfected with the bovine bFGF gene (23), produces an identical ECM except for the inclusion of bFGF in a bound form to heparan sulfate proteoglycans (the HR9-bFGF/ECM). The ECM-bound bFGF has been shown to confer normal biological activities in terms of mitogenic stim' The abbreviations used are: PLDR, potentially lethal damage repair; BAEC, bovine aortic endothelial cells; ECM, extracellular matrix; BAEC/ECM, ECM produced by BAEC; HR9/ECM, ECM produced by PF-HR9 cells; HR9-bFGF/ ECM, ECM produced by HR9-bFGF cells; FGF, fibroblast growth factor; bFGF, basic fibroblast growth factor; DMEM, Dulbecco's modified Eagle's medium; PBS, phosphate-buffered saline; STV. 0.05% trypsin and 0.02% EDTA in PBS; BCS, bovine calf serum. 4 Z. Fuks. I. Vlodavsky. M. Andreeff. and A. Haimovitz-Friedman. Effects of extracellular matrix on the response of endothelial cells to radiation in vitro, submitted for publication. 2552 Downloaded from cancerres.aacrjournals.org on August 3, 2017. © 1991 American Association for Cancer Research. FGF INDUCTION OF PLDR ulation of endothelial cells and of neuronal differentiation of PC 12 cells (23). Utilizing these matrix systems, we demonstrate in the present study that bFGF is capable of serving as an inducer of PLDR in plateau-phase BAEC. Furthermore, our data also show that the induction of PLDR by bFGF in this cell type represents a component of an autoregulated damagerepair cycle. This cycle is initiated by radiation injury and involves the synthesis and secretion of bFGF by the irradiated cells, and the newly synthesized bFGF activates the PLDR pathway via an extracellular autocrine loop. cytofluorograph was interfaced with a Hewlet-Packard microcomputer and the data were analyzed using the "Cell Fit" software package (Becton-Dickinson, Mountainview, CA) and presented via a Tektronix terminal. Radiation Dose-Survival Curves. Survival after irradiation was de fined as the ability of the cells to maintain their clonogenic ability and to form colonies. Twenty-four h prior to irradiation the culture medium was changed to 10% BCS-DMEM. Immediately prior to irradiation, the cells were dissociated with STV, washed twice in 10% BCS-DMEM, and resuspended in the same medium at a concentration of IO5 cells/ ml. Irradiation was carried out in a gamma cell 40 chamber containing two sources of '"Cs (Atomic Energy of Canada) at a dose rate of 100 cGy/min. Single doses ranging from 200 to 600 cGy were delivered. Immediately after irradiation the cells were seeded for colony formation assays in 35-mm dishes (300-3000 cells/dish) coated with either the Cell Cultures. Cloned populations of BAEC were established from HR9/ECM or the HR9-bFGF/ECM. At designated times as described the intima of bovine aorta as previously described (24). Stock cultures in each experiment, a single dose of partially purified bovine brain were grown in 35-mm culture dishes in DMEM supplemented with bFGF (to a final concentration of 200 ng/ml) was added to the media glucose (1 g/liter), 10% heat-inactivated BCS, penicillin (50 units/ml), and streptomycin (50 Mg/ml). Partially purified bovine brain bFGF [200 of the HR9/ECM cultures, because in the absence of bFGF, BAEC failed to form colonies on the HR9/ECM (23). The cells seeded on top ng/ml; prepared as previously described (25)] or purified recombinant of the HR9-bFGF/ECM did not require exogenous FGF supplements bFGF (10 ng/ml; kindly provided by Dr. J. Abraham, California Biotechnology, Inc.) was added every other day during the phase of to produce colonies, since this ECM was shown to contain biologically active ECM-bound bFGF (23). To prove that the effects observed were exponential growth. After 8-10 days in culture, the cells reached conferred by bFGF and not by a contaminant present in the partially confluence and exhibited features of contact-inhibited monolayers. purified bFGF preparation, recombinant bovine bFGF (to a final con These plateau-phase cells were either used for experiments or further subcultured (up toa maximum of 10 subcultures) at a split ratio of 1:10 centration of 10 ng/ml) was used in some experiments in parallel to 200 ng/ml partially purified bFGF, and the results were similar. After to expand the cell population for other experiments. For subculturing, 7-9 days, the cultures were fixed in 3.7% formaldehyde and stained the monolayers were dissociated with 0.05% trypsin and 0.02% EDTA in PBS (STV) (2-3 min at 22°Cor 4°C),washed twice in 10% BCS- with 1% crystal violet in ethanol. Colonies consisting of 50 cells or more were scored, and 3-6 replicate dishes containing 50-150 colonies/ DMEM, and resuspended in DMEM with supplements as described dish were counted for each radiation dose tested. Survival curves were above. These mild conditions of trypsinization were sufficient to detach generated by a computer-assisted program (27). This program analyzes the cells but did not injure, stimulate, or affect cell functions in a the cellular survival after irradiation from the raw data of colony counts detectable way. Cultures of the PF-HR9 mouse endoderma! carcinoma of each culture dish included in the experiment. Using these data, the cell line and its bFGF-HR9 variant were cultured as previously de scribed (23) in DMEM supplemented with glucose (4.5 g/liter), 10% program calculates the surviving fractions for each dose point with associated weights that are based on a Poisson distribution for the fetal calf serum, and antibiotic supplements as described above. After number of counts and the variation of colony counts for each data reaching confluence, the cells were subcultured weekly at a split ratio point. Using the calculated survival fractions and the associated of 1:10 (after dissociation with STV). All cell cultures were maintained weights, the program then fits the data according to the survival model at 37°Cin 10% CO2 humidified incubators. tested [the single-hit multitarget model (14) or the linear quadratic Preparation of Dishes Coated with ECM. HR9/ECM and HR9model (15)1 D>'iteratively weighted least squares regression analysis of bFGF/ECM were produced as previously described (24). PF-HR9 cells all data points, estimating the covariance of the survival curve param or HR9-DFGF cells were seeded (105/35-mm dish) into culture dishes eters and the corresponding confidence limits. The computer then plots precoated with fibronectin (50 jig/dish) to enforce a firm adhesion of the best fit survival curve and assigns an SE bar to each data point, the secreted ECM onto the plastic substratum. The medium consisted thus permitting statistical comparisons of corresponding points on of DMEM supplemented with glucose (4.5 g/liter), 10% fetal calf different curves. serum, and antibiotic supplements as described above. Ascorbic acid Northern Blot Hybridization for Quantitative Evaluations of bFGF (50 jig/ml) was added on days 2 and 4. After 6-7 days in culture, the mRNA. Plateau-phase BAEC were irradiated while adherent as monocell monolayers were denuded by treatment with a solution containing layers to their original culture dishes. For each experimental point, IO8 0.5% Triton X-100 and 20 mivi NH4OH in PBS (3 min at 22°C), cells were used. For extraction of cellular DNA, the monolayers were followed by 4 washes with PBS. While cellular debris and cytoskeletons dissociated by scraping and the cellular RNA was extracted using the of the lysed cells were effectively removed, the underlying ECM re guanidine isothiocyanate-cesium chloride method (28). The total RNA mained intact and adhered to the entire area of the tissue culture dish. (20 Mg/ml) was sorted out according to size by formaldehyde-1% The ECMs could be stored at 4°Cfor several months before being used agarose gel electrophoresis (29), transferred to nitrocellulose mem branes, and hybridized by standard techniques (30) to the 32P nickin experiments. Cell Cycle Analysis. The cell cycle distribution of BAEC was deter translated PJII-I probe containing a 1.4-kilobase fragment of bovine mined by flow cytometry after acridine orange staining of DNA and bFGF cDNA (31). Each sample was also hybridized with a human yRNA as previously described (26). Prior to staining the monolayers actin probe to normalize for the amounts of mRNA involved in each experiment. The blots were then autoradiographed for 16 h at —85°C were dissociated with STV, washed twice with 10% BCS-DMEM, and resuspended in PBS at a concentration of 2-4 x IO5cells/0.2 ml. The with intensifying screens. Quantitative analysis was performed by film MATERIALS AND METHODS cells were then mixed with 0.4 ml of a solution containing 0.05 N HC1, densitometry measurements. 0.15 M NaCl, and 0.1% (v/v) Triton X-100. After 30 s, 1.2 ml of a Statistical Methods. The statistical analysis of the dose-survival solution containing 1 mM EDTA, 0.15 M NaCl, and 6 mg/ml of curves, parameters, and exponents was performed as previously de chromatographically purified acridine orange in 0.2 M Na2HPO4-0.1 M scribed (27). Mean values were compared by Student's t test. All P values were two sided. citric acid buffer (pH 6.0) was added. The cells were then passed through the focus of an excitation argon-ion laser beam (Lexel model 75), and the red fluorescence (fluorescence >600) and green fluorescence (fluo RESULTS rescence, 530) emissions for each cell were measured by separate Effects of bFGF on Parameters of the Dose-Survival Curves. photomultipilers in a computer-interfaced research cytofluorograph The radiation dose-survival experiments were performed with (modified model FC201; Ortho Instruments, Westwood MA). The 2553 Downloaded from cancerres.aacrjournals.org on August 3, 2017. © 1991 American Association for Cancer Research. FGF INDUCTION plateau-phase BAEC 7-9 days after reaching confluency. Cell cycle analysis by flow cytometry showed that 91 ±1.13% (mean ±SE of 6 experiments) of the cells were in G0-G| phase, 3.66% ±0.71% in S phase, and 5.33 ±0.66% in G2-M phase. Fig. 1 shows the dose-survival curves generated with these cells. Im mediately after irradiation the cells were seeded on either the HR9/ECM or the HR9-bFGF/ECM for colony formation as says. A single dose of the partially purified bovine brain bFGF was added to the media of the HR9/ECM cultures 72 h after irradiation (to a final concentration of 200 ng/ml, equivalent to approximately 10 ng/ml of purified recombinant bFGF), since control unirradiated BAEC failed to form colonies on the HR9/ECM in the absence of bFGF (23). The intent of the relatively long delay in bFGF addition was to permit enough time for either repair or lethal fixation of radiation-induced lesion, while the 72-h period was not considered deleterious, because the cloning efficiencies of control unirradiated BAEC remained essentially unchanged regardless of whether the bFGF was added immediately or up to 96 h after seeding (Fig. 3, top curve). The cells cultured on top of the HR9-bFGF/ECM did not regularly receive exogenous bFGF supplements, because this ECM was shown to contain biologically active ECM-bound bFGF (23). However, in several control experiments partially purified bovine bFGF (200 ng/ml) or recombinant bFGF was added to the culture media of the cells plated on the HR9bFGF/ECM, and the resultant survival curves did not differ from those observed without addition of exogenous bFGF (data not shown). As shown in Fig. 1, the slopes of the dose-survival curves did not differ significantly (the Do was 107 ±6.8 cGy on the HR9/ECM, compared to 112 ±1.3 cGy on the HR9bFGF/ECM). On the other hand, the threshold shoulder was nearly completely eliminated in the cells plated on the HR9/ ECM. The Dq on the HR9/ECM was 29 ±19 cGy, compared to 174 ±22 cGy for the cells plated on the HR9-bFGF/ECM (P < 0.05). Similarly, there was an increase in the n value from 1.3 ±0.7 observed on the HR9/ECM to 3.8 ±1.1 on the HR9bFGF/ECM (P < 0.05). When the data were analyzed accord ing to the linear quadratic model (15), the «exponent was 0.304 ±0.16 Gy-' for the HR9-bFGF/ECM and 0.665 ±0.15 Gy~' for the HR9/ECM-plated cells, while the corresponding ßexponents were 0.069 ± 0.03 and 0.03 ± 0.02 Gy"2, respectively. To evaluate whether the timing of bFGF presentation to the irradiated cells accounted for the differences observed in Fig. 1, exogenous bFGF (200 ng/ml) was added to the HR9/ECM cultures immediately after irradiation, instead of 72 h later. Fig. 2 shows that such an application of bFGF resulted in restoration of the shoulder region. The Dq for the cells plated on the HR9-bFGF/ECM in this experiment was 133 ±45 cGy. When plated on the HR9/ECM and bFGF (200 ng/ml) was added immediately after irradiation, the Dq was 162 ±47 cGy, while when bFGF was added 72 h after irradiation the Dq was 35 ±5.9 cGy. The «value from the linear quadratic model changed from 0.17 ±0.048 Gy"1 when bFGF was added at 0' time to 1.93 ±0.36 Gy~' for bFGF addition at 72 h, and the corresponding change in the ßexponent was from 0.1388 ± 0.008 to 0.0114 ±0.06 Gy~2. With increasing delays in bFGF addition, there was a progressive elimination of the threshold shoulder. Fig. 3 shows that, when plateau-phase BAEC were irradiated with a single dose of 200 cGy and plated for colony formation on HR9/ECM with bFGF (200 ng/ml) added im mediately after irradiation, the surviving fraction was 0.156. When addition of bFGF was delayed, the surviving fractions OF PLDR I I M co 10-2 10 -3 Radiation Dose (Gy) Fig. 1. Radiation dose-survival curves of plateau-phase BAEC seeded for colony formation assays on top of either the HR9-Ã’FGF/ECM (O) or on top of the HR9/ECM (A). The cells seeded on top of the HR9/ECM received a single dose of partially purified bovine brain bFGF (200 ng/ml) 72 h after irradiation, while the cells plated on top of the HR9-bFGF/ECM did not receive exogenous bFGF supplements. The best fit curves were calculated by the least squares regression analysis of all data points, using the single-hit multitarget model (14, 27). The data shown were pooled from 6 different experiments. Point, mean; bar, ±SE. 10 O 00 10 -2 co -3 10 10 Radiation Dose (Gy) Fig. 2. Effect of delay in exposure to FGF on the radiation dose-survival curves of BAEC. Plateau-phase BAEC were irradiated and seeded for colony formation assays on top of the HR9/ECM. Partially purified bFGF (200 ng/ml) was added to the cultures at 0' time (A) or at 72 h after irradiation (O). For comparison, a dose-survival curve of the cells plated on top of the bFGF-HR9/ECM (D) is also shown. Dose survival was computed as described in Fig. 1. Point, mean of 5 dishes; bar, ±SE. decreased progressively, with the maximal loss observed within the first 24 h after irradiation. The capacity of the cells to exhibit recovery from radiation damage at 200 cGy was de pendent not only on the timing of bFGF addition after irradia tion but also on the medium concentration of bFGF. When increasing doses of the partially purified brain bFGF (10-600 2554 Downloaded from cancerres.aacrjournals.org on August 3, 2017. © 1991 American Association for Cancer Research. FGF INDUCTION OF PLDR Delay of bFGF Addition (hours) Fig. 3. Effect of delay in addition to bFGF to BAEC plated after irradiation on the HR9/ECM. Plateau-phase cells were irradiated with a single fraction 200 cGy and plated on the HR9/ECM. bFGF (200 ng/ml) was added at different time points after irradiation (*). Top line, cloning efficiencies of control shamirradiated cells plated on the HR9/ECM and bFGF (200 ng/ml) added at different time points after seeding (O). Point, mean; bar, ±SE. bation to permit firm adhesion of the cells as a monolayer to the HR9/ECM, cell cycle analysis revealed that >90% of the cells showed RNA/DNA profiles characteristic of G0-G, (mostly Go) cells. The cells were then washed and resuspended in the same bFGF-free medium, and while still adherent as a monolayer on the HR9/ECM, the cells were irradiated with a single dose of 400 cGy on ice. Immediately after irradiation, cells of sample dishes were dissociated and tested for colony formation on either HR9/ECM (with a single dose of 200 ng/ ml bFGF given 72 h later) or on HR9-bFGF/ECM. The rest of the irradiated cells were maintained at 37°Cwithout medium change and samples were scored for colony formation at 2-h intervals on either the HR9/ECM or the bFGF-HR9/ECM, as described above. Fig. 5 shows that during the first 2 h after irradiation there was a 2.19-fold difference in the clonal ability of the cells tested on the HR9/ECM and the HR9-bFGF/ECM. However, as the delay before being tested for the clonogenic ability was extended, this difference gradually decreased, reach ing a 1.07-fold difference 8 h after irradiation. These data indicated that the cells were undergoing PLDR. The recovery ratio, defined as the ratio of the surviving fraction for 24 h delayed plating onto the HR9/ECM (with a single dose of 200 ng/ml bFGF given 72 h later) to immediate postradiation plating, was 3.27. To evaluate whether this expression of PLDR was associated with bFGF, the cells were irradiated and main tained postirradiation in the presence of neutralizing monoclo nal antibodies against bFGF. Fig. 6 shows that PLDR was nearly completely inhibited by the addition of 12 /¿g/mlof Fl09 anti-bFGF mouse monoclonal antibody (kindly provided by ImClone Systems, Inc., New York, NY). This concentration of the F1-09 antibody was previously shown to neutralize the mitogenic effects of bFGF on BAEC (32). Northern Blot Hybridization Experiments. Since exogenous bFGF was not added to the culture media during the postradiation maintenance of the cell on HR9/ECM, the PLDR experi ments suggested that the irradiated cells themselves secreted 0.25- 0.0 200 400 FGF (ng/ml) Fig. 4. Effect of bFGF concentration on the capacity of irradiated BAEC to form colonies. Plateau-phase cells were irradiated with a single fraction 200 cGy and plated on the HR9/ECM. bFGF was added at different concentrations immediately after irradiation. The surviving fraction is presented as a function of bFGF concentration. ng/ml) were added immediately after irradiation, a dose-re sponse effect was observed, reaching a nearly maximal effect with 200 ng/ml of the partially purified bFGF (Fig. 4). Involvement of bFGF in PLDR. To further characterize the bFGF-dependent radiation damage repair observed in Fig. 1, the capacity of the cells to carry out PLDR was tested by delayed plating experiments (8, 9). Prior to irradiation, the cells were transferred onto dishes coated with the bFGF-free HR9/ ECM at a density of 1.2 x 10" cells/35-mm dish (a density typical for confluent contact-inhibited BAEC). The medium consisted of DMEM supplemented with glucose (1 g/liter), antibiotics as detailed in "Materials and Methods," and 10% heat-inactivated bovine calf serum (which did not have a mitogenie effect on BAEC), but no FGF. After an overnight incu 4 8 12 16 20 24 Delay Before Plating (hours) Fig. 5. Delayed plating (PLDR) experiments in plateau-phase BAEC main tained after irradiation under bFGF-free conditions. Monolayers of plateau-phase BAEC were irradiated with a single dose of 400 cGy while adherent as contactinhibited monolayers (1.2 x 10* cells/dish) to the HR9/ECM in bFGF-free medium. Immediately after irradiation, and at the specified time intervals after irradiation, cells of sample dishes were assayed for colony formation by plating 2000 cells on either the HR9/ECM (A) (with a single dose of 200 ng/ml partially purified FGF given 72 h later) or on the HR9-bFGF/ECM (O). The surviving fraction is shown as a function of the duration of delay in plating. Similar results were observed in 3 independent experiments. Point, mean of 6-8 dishes; bar, ± SE. 2555 Downloaded from cancerres.aacrjournals.org on August 3, 2017. © 1991 American Association for Cancer Research. FGF INDUCTION OF PLDR thereafter to preirradiation levels (data not shown). Increased levels of mRNA of bFGF (2.2-fold) and secretion of bFGF into the conditioned media were observed even after a radiation dose of 125 cGY, a dose which was not accompanied by detectable acute postirradiation cell lysis (<2% at 24 and 48 h by dye exclusion tests) and which retained the clonogenic capacity in >75% of the irradiated BAEC. 0.20- I O e 0.15- 0.10- DISCUSSION Our data demonstrate that bFGF serves as a potent inducer of potentially lethal radiation damage repair in endothelial cells. In fact, bFGF is the first well-defined physiological agent char acterized as an inducer of PLDR in any type of cells. Our data also suggest that bFGF activates the PLDR pathway via inter 0.00 action of extracellular bFGF with receptors located on the 24 plasma membrane of the cell, since neutralizing monoclonal Delay Before Plating (hours) antibodies against bFGF inhibited PLDR (Fig. 5). This exper Fig. 6. Inhibition of PLDR in irradiated plateau-phase BAEC by monoclonal antibodies against bFGF. PLDR experiments were performed with plateau-phase iment suggested that the newly synthesized intracellular bFGF BAEC as described in Fig. 5. White columns, delayed plating for colony formation was apparently incapable of stimulating PLDR via an internal assays on the HR9-bFGF/ECM; hatched columns, delayed plating for colony loop and that bFGF secretion and stimulation via an extracel formation on the HR9/ECM (with a single dose of 200 ng/ml partially purified bovine brain bFGF given 72 h later); shaded columns, delayed plating conditions lular autocrine loop were necessary to initiate the PLDR path similar to those shown by the hatched bars except that 12 ¿ig/mlof the Fl-09 way. Another basic requirement for the induction of PLDR in anti-bFGF mouse monoclonal antibodies were present in the medium, removed our system was its critical dependence on the timing of bFGF only when the cells were plated for the colony formation assays. Similar data were obtained in 3 independent experiments. Column, mean of 3-5 dishes: bar, presentation to irradiated cells, restricting it to the early pos±SE. tradiation period. A delay resulted in an irreversible lethal fixation of potentially repairable lesions, which progressively developed over the first 24 h after irradiation (Figs. 2 and 3). A B In contrast, a delay in bFGF application of up to 96 h after irradiation did not affect its ability to induce a mitogenic r 7.0response and replication of surviving cells (Fig. 3). Whether the b-FGF mitogenic response and the PLDR pathway share the same bFGF receptor-binding specificities and the same signal trans-3.7 duction pathways in endothelial cells is unknown, but it is currently being investigated. The fact that PLDR is an inducible process was previously demonstrated in several studies, which showed that conditioned media from irradiated cells conferred PLDR in other irradiated cells (17-20). These data suggested that after irradiation cells Actin 2.2 secrete into the culture media substances that serve as autocrine or paracrine inducers of PLDR. One study demonstrated that conditioned media from irradiated cells accumulated an as yet Fig. 7. Northern blot analysis of mRNA extracted from irradiated BAEC unidentified temperature-labile factor which induced PLDR hybridized with the PJII-1 bFGF cDNA probe. Plateau-phase BAEC (10' cells/ and which was both DNase and RNase resistant (20). Our data blot) were scored for the levels of bFGF mRNA 6 h after irradiation with a single dose of 400 cGy. Hybridization was performed with the PJII-I probe containing are consistent with this model, because we have demonstrated a 1.4-kilobase fragment of bovine FGF cDNA (32). The blots were also probed that radiation induced in BAEC a complete cycle of an autorewith a human -. .u-iin probe to normalize for the amounts of mRNA located in gulated damage-repair pathway. This cycle was initiated by each lane. Species of bFGF transcriptional mRNA are shown for control unirradiated BAEC (lane A) and for BAEC irradiated with a single dose of 400 cGy DNA damage and was followed by an increase in bFGF-specific (lane B). mRNA, bFGF synthesis, and its secretion into the conditioned medium. The newly synthesized bFGF induced the PLDR bFGF into the medium. This presumption is consistent with pathway by the bFGF acting via an extracellular autocrine loop, inhibitable by specific anti-bFGF antibodies. The cycle was our previous publication (33), which showed that BAEC surviv completed with recovery of damaged cells from radiation le ing the acute lethal effects of radiation synthesize and secrete biologically active bFGF into their conditioned media (33). To sions and with restoration of their clonogenic capacity. Evi dence for the existence of this autoregulated damage-repair further evaluate this possibility. Northern blot hybridization experiments were performed using the PJII-I probe containing cycle was derived not only from the experiments which tested a 1.4-kilobase fragment of bovine bFGF cDNA (31). Fig. 7 the capacity of the cells to repair potentially lethal damage shows that, following a single dose of 400 cGy, there was a 5.6- (Figs. 5 and 6) but perhaps also indirectly from the radiation dose-survival curves. The minimal threshold shoulder observed fold increase in the 3.7-kilobase species and a 4.7-fold increase in the 7.0-kilobase species of bFGF mRNA. In contrast, the in the survival curves of the HR9/ECM-plated BAEC shown same dose of radiation had no effect on the levels of human y- in Fig. 1 (Dq 29 ±19 cGy) probably represents an autocrineactin mRNA. Increased bFGF mRNA levels were detected as stimulated repair effect of bFGF. The low level of repair activity early as 2 h after irradiation, peaked at 6-8 h, and decreased in this experiment can be attributed to the small numbers of 2556 0.05 - I Downloaded from cancerres.aacrjournals.org on August 3, 2017. © 1991 American Association for Cancer Research. FGF INDUCTION OF PLDR cells used in the colony formation assays (300-3000 cells/35mm dish, compared to 1.2 X 106/35-mm dish used in the PLDR experiments), resulting in low culture media concentrations of bFGF. The demonstration that PLDR is associated with bFGF secreted into the culture medium by the irradiated cells them selves is consistent with our previous data (33), which demon strated the accumulation of increasing amounts of biologically active bFGF in conditioned media of endothelial cells after irradiation. Whereas there is significant information concerning several components of our proposed autoregulated damage-repair cycle in irradiated endothelial cells, others remain essentially un known. In particular, there is little information regarding the molecular basis of potentially lethal damage lesions and the signal transduction and biochemical pathways involved in their repair. PLDR in mammalian cells has been shown to proceed in two distinct phases, with an early fast phase that occurs within the first few minutes after irradiation and a slower phase which proceeds during 4-6 h after treatment, although residual repair has been detected up to 24 h after irradiation (11). Kinetic studies of rejoining of radiation-induced single- and doublestranded DNA breaks in mammalian cells have also demon strated bipasic patterns (34, 35), with a tv, of 2-20 min for the early phase and a tVtof 1-3 h for the late phase (36-39). It has been suggested that the initial fast repair entails rejoining of breaks by enzymes which are constitutively expressed in mam malian cells, such as DNA ligases, exonucleases, and DNA polymerases (35, 40, 41). The late phase, which proceeds at a significantly slower rate, is characteristically blocked by inhib itors of protein, RNA, and DNA synthesis (11), suggesting that it may require the induction of specific genes and gene products to repair more complex types of DNA lesions (41). Our data are consistent with this hypothesis, demonstrating that at the least the induction of bFGF synthesis may be required to initiate the slow phase of PLDR in our system. Whether the induction of specific enzymes is also necessary is as yet unknown, but the well-defined and highly controlled BAEC/HR9 system may be useful in elucidating the biochemical mechanisms of this repair and the nature of potentially lethal damage lesions. The use of neutralizing monoclonal antibodies against bFGF and chemical agents capable of dissociating bFGF from its functional cell surface receptors [e.g., suramin (42), protamine (43)] provides opportunities to turn on and off the stimuli that lead to PLDR in our system and may be useful in the study of individual components of its pathway and in kinetic studies. The radiation-induced activation of bFGF synthesis observed in our study apparently represents a widespread biological radiation-related phenomenon which is not specific to the BAEC, nor is it specific to the bFGF gene. We have found that, in addition to bFGF, irradiation of BAEC resulted in an in crease of platelet-derived growth factor-specific mRNA.5 Boothman et al. (41) recently showed that ionizing irradiation activated in confluence-arrested human malignant melanoma cells a pleiotropic protein synthesis response detected by twodimensional polyacrylamide gel electrophoresis of whole cell extracts. This inducible expression of protein synthesis seemed to be specific to ionizing radiation, since it was not induced by other stresses such as heat shock, hypoxia, or alkylation treat ment with melphalan (41). Ionizing irradiation was also re ported to induce increased tumor necrosis factor-a-specific mRNA transcripts and enhanced production of tumor necrosis 5 L. Witte, A. Haimovitz-Friedman, data. A. Chaudhuri, and Z. Fuks, unpublished factor-«protein in several human tumor cell lines (44), transcriptional and posttranscriptional expressions of the c-jun and c-fos protooncogenes in HL-60 cells (45), transcription of the ß-PKCgene in Syrian hamster embryo fibroblasts (46), tran scription of the chloramphenicol acetyltransferase gene transfected into the genome of NIH-3T3 cells (47), and transcriptional expressions of the interleukin 1 gene in Syrian hamster embryo fibroblasts (48). The pleiotropic gene expressions ob served in mammalian cells after exposure to X-irradiation are believed to represent a stress response to the damage induced by radiation to cellular DNA (41, 45, 47-49). Although the full details of this stress response are unknown, recent studies have suggested that it is different in several of its expressions from the stress responses induced by heat shock or by chemical injury to DNA in mammalian cells (41, 45, 47-50). In analogy to the SOS response to DNA damage in prokaryotic cells (51), a hypothesis has been suggested that relates PLDR after Xirradiation of mammalian cells to the postradiation stress re sponse (41). Our data provide the first direct confirmation for this hypothesis. Furthermore, since many types of cells do not express receptors to bFGF, we suggest that the capacity to stimulate PLDR is not specific to bFGF. Other cell types may respond to other inducers of PLDR, the synthesis of which may also be activated as a stress response to radiation-induced DNA damage, and the stimulation of PLDR by such agents may also be conferred via autocrine/paracrine mechanisms. By analogy to the patterns of the growth factor-mediated mitogenic path ways, it is possible to envision that different cell systems re spond to different autocrine/paracrine-regulated "PLDR-stimulating factors," each functioning through a specific cell surface receptor and through its corresponding signal transduction pathway, but all leading to a common final pathway that terminates with repair of radiation-induced lesions. The demonstration that bFGF is an inducer of radiation damage repair in endothelial cells may also have practical clinical implications. Recent histopathological and ultrastruc tural studies showed that damage to the endothelium of the microvasculature is a predominant lesion leading to radiation injury in slowly proliferating normal tissues (e.g., skin, heart, lung, liver, and the central nervous system) (52,53). In contrast, the endothelium of large vessels and of the endocardium are significantly more resistant to the effects of irradiation in vivo (52, 53). The variations in the sensitivities of different sections of the vascular system to irradiation correlate with our recent observations of the distribution of bFGF in the basement mem branes of various types of blood vessels (32). Our immunohistochemical studies showed that bFGF is ubiquitously detected in the basement membranes of all sizes of blood vessels (32), but while homogeneous and intensive immunoreactivities were observed in large and intermediate size blood vessels, hetero geneity was found in capillaries. The most intense capillary immunoreactivities were observed in the basement membranes of branching points of mid-size capillary vessels, while staining became progressively less intense in nonbranching sections of capillaries and seemed to be absent in some nonbranching regions (32). The latter regions of the the capillary endothelium apparently exhibit the most severe changes of radiation damage (52, 53). Whether the cells of these radiosensitive sections of the capillary intima are capable of responding to irradiation with synthesis and secretion of bFGF is unknown, but the idea of using i.v. applications of bFGF as a mode of radioprotection against radiation damage to the microvasculature is an objective that warrants further consideration. 2557 Downloaded from cancerres.aacrjournals.org on August 3, 2017. © 1991 American Association for Cancer Research. FGF INDUCTION OF PLDR ACKNOWLEDGMENTS We wish to acknowledge with thanks Maureen McLoughlin and Roger S. Persaud for their excellent technical assistance. REFERENCES 1. Painter, R. B. The role of DNA damage and repair in cell killing induced by ionizing radiation. In: R. E. Myen and H. R. Withers (eds.). Radiation Biology in Cancer Research, pp. 59-68. New York: Raven, 1979. 2. Ward, J. F. DNA damage produced by ionizing radiation in mammalian cells: identities, mechanisms of formation and repairability. Prog. Nucleic Acid Res., .ÃŽ.-95-125,1988. 3. Frankenberg-Schwager, M. Review of repair kinetics for DNA damage in duced in eukaryotic cells in vitro by ionizing radiation. Radiother. Oncol.. 14: 307-320, 1989. 4. Bryant, P. E. Enzymatic restriction of mammalian cell DNA: evidence for double-strand breaks as potentially lethal lesions. Int. J. Radiât,liiul.. 48: 55-60, 1988. 5. Radford I. R. Evidence for a general relationship between the induced level of DNA double-strand breakage and cell-killing after X-irradiation of mam malian cells. Int. J. Radiât.Biol., 49: 611-620. 1986. 6. Elkind, M. M., and Sutton, H. A. X-ray damage and recovery of mammalian cells in culture. Nature (Lond.), 184: 1293-1295, 1959. 7. Elkind, M. M. Repair processes in the treatment and induction of cancer with radiation. Cancer (Phila.), 65: 2165-2171, 1990. 8. Philips. E. A., and Tolmach, L. J. Repair of potentially lethal damage in Xirradiated HeLa cells. Radiât.Res., 29: 413-432, 1966. 9. Hahn, G. M., and Little, J. B. Plateau phase cultures of mammalian cells: in vitro model for human cancer. Curr. Top. Radiât.Res., 8: 839-871, 1973. 10. Curtis, S. B. Lethal and potentially lethal lesions induced by radiation—a unified repair model. Radial. Res., 706: 252-270, 1986. 11. lliakis, G. Radiation-induced potentially lethal damage: DNA lesions suscep tible to fixation. Int. J. Radiât.Biol., 53: 541-584, 1988. 12. Steel, G. G.. McMillan. T. J., and Peacock. J. H. The picture has changed in the 1980s. Int. J. Radiât.Biol., 56: 525-537, 1989. 13. Alper, T. Cellular Radiobiology. Cambridge, MA: Cambridge University Press, 1979. 14. Elkind, M. M., and Sutton, H. Radiation response of mammalian cells grown in culture. Radiât.Res., 13: 556-593. 1960. 15. Chadwick, K. H., and Leenhouts, H. P. A molecular theory of cell survival. Phys. Med. Biol., 18: 78-87, 1973. 16. Friedberg, E. C., and Hanawalt, P. C. Mechanisms and Consequences of DNA Damage Processing, pp. 185-405. New York: Alan R. Liss, 1988. 17. Little, J. B. Factors influencing the repair of potentially lethal radiation damage in growth-inhibited human cells. Radial. Res., 56: 320-333, 1973. 18. Horowitz, I. A., Norwint, H., and Hall, E. J. Conditioned media from plateau phase cells: effect on growth of proliferativi cells and on repair of potentially lethal radiation damage. Radiology. 114: 723-726, 1975. 19. Hetzel, F. W., Kruuv. J., and Frey, H. E. Repair of potentially lethal damage in irradiated V79 cells. Radiât.Res.. 68: 308-319, 1976. 20. Little, J. B. Repair of potentially-lethal radiation damage in mammalian cells: enhancement by conditioned medium from stationary cultures. Int. J. Radiât.Biol., 20: 87-92, 1971. 21. Bashkin, P., Klagsburn, M., Doctrow, S., Svahn, C. M., Folkman, J., and Vlodavsky, I. Basic fibroblast growth factor binds to subendothelial extracel lular matrix and is released by heparanase and heparin-like molecules. Biochemistry. 28: 1737-1743, 1989. 22. Chung, A. E., Estes, L. E., Shinozuka, H. B., Braginski, J., and Lorz, C. Morphological and biochemical observations on cells derived from differen tiation of embryonal carcinoma cell line PCC4-F. Cancer Res., 37: 20722081, 1978. 23. Rogelj, S., Klagsbrun. M., Atzmon, R., Kurokawa, M., Haimovitz-Friedman, A., Fuks, Z., and Vlodavsky I. Basic fibroblast growth factor is an extracel lular matrix component required for supporting the proliferation of vascular endothelial cells and the differentiation of PC 12 cells. J. Cell. Biol., /09: 823-831, 1989. 24. Gospodarowicz, D., Moran, J., Brown, D., and Birdwell, C. R. Clonal growth of bovine vascular cells: fibroblast growth factor as a survival agent. Proc. Nati. Acad. Sci. USA, 73: 4120-4124, 1976. 25. Gospodarowicz, D., Bialecki. H., and Greenburg, G. Purification of the fibroblast growth factor activity from bovine brain. J. Biol. Chem., 253: 3736-3743, 1978. 26. Andreeff, M., Darzynklewicz, Z., Sharpless, T. K., Clarkson, B. D., and Melamed, M. R. Discrimination of human leukemia subtypes by flow cytometric analysis of cellular DNA and RNA. Blood, 55: 282-293, 1980. 27. Albright, N. Computer programs for the analysis of cell survival. Radiât. Res., 112: 331-340, 1987. 28. Chirgwin, J. M.. Przybyla, A. E., MacDonald, R. J., and Rutter, W. J. Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry, 18: 5294-5299, 1979. 29. Aviv, H., and Leder, P. Purification of biologically active globin messenger by chromatography on oligothymidylic acid-cellulose. Proc. Nati. Acad. Sci. USA, 69:1408-1412, 1972. 30. Manilas, T., Fritsch, E. F., and Sambrook, J. Molecular Cloning: A Labo ratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1982. 31. Abraham J. A., Mergia, A., Whang, J. L., Turnólo,A., Friedman, J., Hjerrild, K. A., Gospodarowicz. D., and Fiddes, J. C. Nucleotide sequence of a bovine clone encoding the angiogenic protein, basic fibroblast growth factor. Science (Washington DC), 233: 545-548, 1986. 32. Cordon-Cardo, C., Vlodavsky, !.. Haimovitz-Friedman, A., Hicklin, D., and Fuks, Z. Expression of basic fibroblast growth factor in normal human tissues. Lab. Invest., 63: 832-840, 1990. 33. Witte, L., Fuks, Z., Haimovitz-Friedman, A., Vlodavsky, I., Goodman, D. S., and Eldor. A. Effects of irradiation on the release of growth factors from cultured bovine, porcine, and human endothelial cells. Cancer Res., 49:50665072, 1989. 34. Bryant, P. E., Warring, R., and Ahnstrom, G. DNA repair kinetics after low doses of X-rays. A comparison of results obtained by the unwinding and nucleotide sedimentation methods. Mutât.Res., 131: 19-26, 1984. 35. Radford, 1. R. Effect of cell-cycle position and dose on the kinetics of DNA double strand breakage repair in X-irradiated Chinese hamster cells. Int. J. Radiât.Biol., 52: 555-563, 1987. 36. Dikomey, E., and Franzke, J. Three classes of DNA strand breaks induced by X-irradiation and internal fi rays. Int. J. Radial. Biol., 50:893-908, 1986. 37. Wlodek, D., and Hittelman, W. N. The repair of double-stranded breaks correlates with radiosensitivity of L5178-S and L5178-R cells. Radiât.Res., 112: 146-155, 1987. 38. VanAnkeren, S. C., Murray, D., and Meyn, R. E. Induction and rejoining of 7-ray-induced DNA single- and double-strand breaks in Chinese hamster AA8 cells and in two radiosensitive clones. Radiât.Res., 7/6:511-525,1988. 39. Koval, T. M., and Kazmar, E. R. DNA double-strand break repair in eukaryotic cell lines having radically different radiosensitiv ities. Radiât.Res., 7/5:268-277. 1988. 40. Hittelman, W. N., and Pollard, M. A comparison of the DNA and chromo some repair kinetics after -yirradiation. Radiât.Res., 92: 497-509, 1982. 41. Boothman, D. A., Bouvard, I., and Hughes, E. N. Identification and charac terization of X-ray-induced proteins in human cells. Cancer Res., 49: 28712878, 1989. 42. Coffey, R. J., Leof, E. B., Shipley, G. P., and Moses, H. L. Suramin inhibition of growth factor receptor binding and mitogenicity in AKR-2B cells. J. Cell Physiol.. 132: 143-148. 1987. 43. Neufeld, G., and Gospodarowicz, D. Protamine sulfate inhibits mitogenic activities of the extracellular matrix and fibroblast growth factor, but poten tiates that of epidermal growth factor. J. Cell. Physiol., 132: 287-294, 1987. 44. Hallaban, D. E., Spriggs, D. R., Beckett, M. A., Kuff, D. W., and Weichselbaum, R. R. Increased tumor necrosis factor a mRNA after cellular exposure to ionizing radiation. Proc. Nati. Acad. Sci. USA, 86: 10104-10107, 1989. 45. Sherman, M. L., Datta, R., Hallahan, D. E.. Weichselbaum, R. R., and Kuff, D. W. Ionizing irradiation regulates expression of the c-jun protooncogene. Proc. Nati. Acad. Sci. USA, 87: 5663-5666. 1990. 46. Woloschak, G. E.. Chang-Liu, C-M.. and Shearin-Jones, P. Regulation of protein kinase C by ionizing radiation. Cancer Res., 50: 3963-3967, 1990. 47. Lin, C. S., Goldthwait, D. A., and Samols, D. Induction of transcription from the long terminal repeat of Moloney murine sarcoma provirus by UVirradiation, X-irradiation, and phorbol ester. Proc. Nati. Acad. Sci. USA, 87: 36-40, 1990. 48. Woloschak, G. E., Chung-Liu, C-M., Jones, P. S., and Jones, C. A. Modu lation of gene expression in Syrian hamster embryo cells following ionizing radiation. Cancer Res., 50: 339-344, 1990. 49. Herrlich, P., Angel, P., Rahmsdorf. H. J., Mallick, U., Poting, A., Hiber, L., Lucke-Huhle, C., and Schorpp. M. The mammalian genetic stress response. Adv. Enzyme Regul., 25: 458-504, 1986. 50. Foranee, A. J., Zmudzka, B., Hollander, C., and Wilson, S. H. Induction of b-polymerase mRNA by DNA-damaging agents in Chinese hamster ovary cells. Mol. Cell. Biol.. 9: 851-853, 1989. 51. Little, J. W., and Mount, D. W. The SOS regulatory system of Escherichia coli. Cell, 29: 11-22, 1982. 52. Fajardo, L. F., and Berthrong, M. Vascular lesions following radiation. Pathol. Annu., 23: 297-330, 1988. 53. Reinhold, H. S., Fajardo, L. F., and Hopewell, J. W. The vascular system. Adv. Radial. Biol., 14: 177-226, 1990. 2558 Downloaded from cancerres.aacrjournals.org on August 3, 2017. © 1991 American Association for Cancer Research. Autocrine Effects of Fibroblast Growth Factor in Repair of Radiation Damage in Endothelial Cells Adriana Haimovitz-Friedman, Israel Vlodavsky, Alpana Chaudhuri, et al. Cancer Res 1991;51:2552-2558. Updated version E-mail alerts Reprints and Subscriptions Permissions Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/51/10/2552 Sign up to receive free email-alerts related to this article or journal. To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at [email protected]. To request permission to re-use all or part of this article, contact the AACR Publications Department at [email protected]. Downloaded from cancerres.aacrjournals.org on August 3, 2017. © 1991 American Association for Cancer Research.