Fibroblast growth factor-2–toxin induced cytotoxicity: differential sensitivity of co-cultured vascular smooth muscle cells and endothelial cells

Atherosclerosis 137 (1998) 277 – 289

Fibroblast growth factor-2 –toxin induced cytotoxicity: differential sensitivity of co-cultured vascular smooth muscle cells and endothelial cells Peter H. Lin a,c,e, Dewei Ren c,e, Mark K. Hirko c,e, Steven S. Kang c,e, Glenn F. Pierce b, Howard P. Greisler c,d,e,* a

Department of Surgery, The Chicago Medical School, North Chicago, IL, USA b PRIZM Pharmaceuticals, San Diego, CA, USA c Department of Surgery, Loyola Uni6ersity Medical Center, 2160 South First A6enue, Maywood, IL 60153, USA d Department of Cell Biology, Neurobiology and Anatomy, Loyola Uni6ersity Medical Center, Maywood, IL, USA e Department of Surgery, Hines Veterans Administration Hospital, Hines, IL, USA Received 9 May 1997; received in revised form 12 November 1997; accepted 14 November 1997

Abstract Recombinant FGF-2-SAP is a mitotoxin consisting of the plant-derived ribosome-inactivating toxin saporin (SAP) fused to basic fibroblast growth factor (FGF-2). FGF-2-SAP targets and kills cells bearing upregulated FGF receptors. In vivo, FGF-2-SAP inhibits smooth muscle cell hyperplasia in models of restenosis. The present study examined the potential for a differential effect of FGF-2-SAP on canine vascular endothelial cells (EC) and smooth muscle cells (SMC) separately as well as in a novel co-culture model. Canine vascular SMC and EC cultures were established separately and made quiescent once cells reached 80% confluence. Following the release from growth arrest, both cell types were treated with FGF-2-SAP, or FGF-2, or SAP alone for 48 h. [3H]TdR incorporation was used to determine the growth response of SMC and EC. The co-culture system was created by plating canine vascular SMC and EC on either side of a microporous 13 mm thick polyester membrane insert. Both cell types were grown to 80% confluence and independently made quiescent. Following the release from growth arrest, cells were treated with FGF-2-SAP, or FGF-2, or SAP alone. Negative and positive control groups were untreated wells containing phosphate buffered saline and complete growth media, respectively. After 48 h, both [3H]TdR incorporation and total DNA content, by fluorometric measurement, were quantitated in SMC and EC independently. FGF-2-SAP showed a concentration-dependent cytotoxicity in both canine SMC and EC but cytotoxicity for EC required substantially higher concentrations. In co-cultured SMC, FGF-2-SAP significantly decreased both [3H]TdR uptake and total DNA content at 0.5, 5, 50, and 500 ng/ml (0.01–10 nM) compared to positive controls. In co-cultured EC, FGF-2-SAP decreased [3H]TdR uptake at 50 and 500 ng/ml and total DNA content at 500 ng/ml compared to positive controls. Neither SAP alone nor FGF-2 alone showed a significant effect on [3H]TdR uptake or DNA content of either SMC or EC. In this unique co-culture model, which better replicates the relationship between SMC and EC in vivo, we demonstrated a dose – response range of FGF-2-SAP at which both the proliferation and total cell number of SMC, but not EC, is significantly reduced. These data suggest that FGF-2-SAP may have therapeutic utility in inhibiting myointimal hyperplasia in the absence of a deleterious effect on regenerating endothelium following vascular reconstructions. © 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Endothelial cells (EC); Fibroblast growth factor (FGF); Intimal hyperplasia; Mitotoxin; Restenosis; Saporin; Smooth muscle cells (SMC)

* Corresponding author. Tel.: +1 708 2168541; fax: +1 708 2166300. 0021-9150/98/$19.00 © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0021-9150(97)00284-0

278

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289

1. Introduction Myointimal hyperplasia represents the most common mode of failure of peripheral vascular reconstructions. The hallmark of myointimal hyperplasia is smooth muscle cell (SMC) migration from the media into the intima followed by proliferation and synthesis of collagen and extracellular matrix. These activities are regulated in part by growth factors, prominent among them being basic fibroblast growth factor (FGF-2). One potentially therapeutic approach to inhibit myointimal proliferation of SMC is to use a chimeric fusion toxin, which fuses the genes for FGF-2 and saporin (SAP), a ribosome inactivating protein derived from the plant Saponaria officinalis [1]. The resulting mitotoxin, FGF2-SAP, binds to FGF receptors (FGFR) with similar affinity as FGF-2 and becomes internalized by receptor-mediated endocytosis [2]. Following lysosomal degradation of the growth factor moiety, the proteaseresistant saporin enzymatically inactivates ribosomes and inhibits protein synthesis resulting in cell death. SMC undergo a significant transient upregulation of FGFR expression compared to either endothelial cells (EC) or uninjured SMC following arterial balloon injury [3,4]. In vivo, injured SMC respond to exogenously delivered FGF-2 with exaggerated proliferation resulting in increased intimal hyperplasia [5]. The abundance of FGFR on SMC following therapeutic intervention (injury) makes these cells a preferential target of an FGF-2-SAP mitotoxin. The purpose of the present experiments was threefold. First, to study the proliferation of canine vascular EC and SMC using a co-culture model which simulates the vascular wall interaction in an in vitro environment; second, to examine the differential effects of FGF-2SAP on proliferation of canine vascular EC or SMC cultured alone; and third, to examine the effect of FGF-2-SAP on proliferation of both vascular EC and SMC in the co-culture system.

2. Materials and methods

2.1. Cell har6est Animal care complied with the ‘Principles of Laboratory Animal Care’ and ‘The Guide for the Care and Use of Laboratory Animals’ (NIH Publication No. 85-23, revised 1985). Adult mongrel dogs were anaesthetized with 30 mg/kg intravenous thiamylal sodium (Parke-Davis, Morris Plains, NJ), intubated, and ventilated. Anesthesia was maintained with nitrous oxide and halothane. Exposure was obtained by bilateral neck incisions. Canine carotid arteries were resected, the adventitia and intima removed, and SMC cultured in standard explant methods [6]. EC were enzymatically

harvested from canine jugular veins with sequential digestion using EDTA/trypsin and collagenase as previously described [7]. Cells were maintained in Dulbecco’s modified Eagle’s medium (Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS; Hyclone Laboratories, Logan, Utah), 10 ng of FGF-1 (a generous gift from W.H. Burgess, American Red Cross Holland Laboratories, Rockville, MD), 10 U of heparin (Lyphomed, Melrose Park, IL), 10 nM non-essential amino acids, 100 mM sodium pyruvate, 100 U of penicillin, 100 mg of streptomycin, and 50 mg of gentamicin (Gibco) per ml. Cells were incubated at 37°C in a humidified atmosphere of 5% CO2. Immunoperoxidase staining with a primary mouse monoclonal antibody to a-actin (Sigma Chemical Company, St. Louis, MO) and a secondary biotinylated horse anti-mouse antibody (Vector Laboratories, Burlingame, CA) was performed for SMC identification. Primary SMC migrating from explants were used in all experiments. EC were identified by immunoperoxidase staining using primary rabbit anti-human-von Willebrand factor antibody (Sigma) and a secondary goat anti-rabbit IgG (Vector). EC in passages 2–5 were used in all experiments.

2.2. Co-culture system Co-cultures were established by plating cells on either side of a microporous 13 mm thick polyester membrane insert (Costar Scientific Corp, MA) with a 0.5 mm pore size. The membrane insert was coated with fibronectin 1.5 mg/cm2 (American Red Cross) for 10 min each on both surfaces prior to cell seeding. Co-culture inserts were placed in sterile tissue culture plates in an inverted position such that the outer side of the membrane faced upward. Either EC or SMC were first plated on this outer membrane surface for 20 min during which the cells became adherent. The inserts were then placed in 24-well tissue culture plates in the upright position such that the first cell line was on the outer membrane. The second cell line was then seeded on the inner membrane surface for 20 min to allow cellular adherence. Growth medium was placed in both upper and lower compartments and replaced every 48 h. SMC were seeded at a density of 104 cells/cm2 and EC were seeded at a density of 2.5 × 103 cells/cm2 on the membrane surface based on pilot studies which indicated that these seeding densities will allow both cell lines to reach 80% confluence following 72 h of growth.

2.3. Proliferation assays of co-cultured EC and SMC The proliferation of SMC was selectively studied in co-culture opposite EC (SMC/EC) and compared to the control group of SMC cultured opposite SMC (SMC/ SMC). After co-cultures were established for 72 h, the

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289

279

Fig. 1. Tritiated thymidine uptake of co-cultured canine SMC. Experimental group (SMC in SMC/EC co-cultures) is compared with control (SMC in SMC/SMC co-cultures).

co-cultured cells were placed in serum-free medium for 48 h, based on the pilot studies that indicated such a medium would fully quiesce SMC without significantly affecting the growth rate of EC. Following the quiescence period, the co-cultures were returned to complete growth medium for an additional 6 days of growth during which time the co-cultured SMC were harvested daily for bioassays. Proliferation assays included tritiated thymidine ([3H]TdR) uptake following 48 h of incubation (n= 5 wells per day per group) and fluorometric measurements of total DNA (n =5 wells per day per group). The proliferation of EC was studied in co-culture opposite SMC (EC/SMC) and compared to the control group of EC cultured opposite EC (EC/EC). After co-cultures were established for 72 h, the co-cultured cells were placed in an FGF-1-free medium for 120 h, based on the pilot studies that indicated such a medium would fully quiesce EC without significantly affecting the growth of SMC. Following the quiescence period, the co-cultures were returned to complete growth medium for an additional 6 days of growth during which time the co-cultured EC were harvested daily for bioassays. Proliferation assays included [3H]TdR uptake following 48 h of incubation (n =5 wells per day per group) and fluorometric DNA measurements (n=5 wells per day per group).

2.4. Tritiated thymidine uptake DNA synthesis was measured by [3H]TdR uptake; 1 mCi of [3H]TdR (New England Nuclear, Boston, MA) was added to each co-culture well daily in replicates of five for 4 consecutive days and incubated for 48 h prior

to scintillation counting. Growth medium was removed and the cells on the bottom surface of membrane inserts were gently scrapped off using a soft-rubber spatula. Co-culture inserts, with cells remaining on the upper surface of the membrane, were gently rinsed with 0.9% sodium chloride. Cells were fixed with methanol and lysed with distilled water. DNA was precipitated with 5% trichloroacetic acid and cell lysates were washed with distilled water. Precipitated DNA was solubilized by adding 200 ml of 0.3 M NaOH for 20 min. This solution was placed in 10 ml of liquid scintillation cocktail (Beckman Ready Protein+) and counted in an automated counter (Beckman LS6800; Beckman Instruments, Fullerton, CA).

2.5. Fluorometric DNA measurement Total DNA content (reflecting total cell numbers) was measured by a filter fluorometer (Ratio-2 Filter Fluorometer; Optical Technology Devices). Calf thymus DNA (Sigma) was used as the control in a fluorescence enhancement assay. The fluorometric quantitation buffer consisted of 2.0 mol/l NaCl, 1 mmol/l ethylenediamine tetraacetic acid, 10 mmol/l trisaminomethane base (Gibco), and 8.0 mg/ml of the fluorescent dye (Hoefer Scientific, San Francisco, CA) Co-culture samples, with undersurfaced cells removed, were placed in a − 80°C freezer overnight then thawed in 100 ml distilled water. Cell samples were added to 900 ml of fluorometric buffer and placed in a spectrophotofluorometer with an excitation wavelength of 260 nm. Test samples were compared with the standard curve generated with calf thymus DNA.

280

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289

Fig. 2. Fluorometric DNA measurement of co-cultured canine SMC. Experimental group (SMC in SMC/EC co-cultures) is compared with control (SMC in SMC/SMC co-cultures). (A) Total DNA content by fluorometric measurement from day 1 to day 6. (B) Net increase of total DNA content per 48 h.

2.6. Effects of FGF-2 -SAP on EC and on SMC single-cell-type cultures Effects of FGF-2-SAP on either SMC or EC were evaluated independently in tissue culture plates. The mitotoxin FGF-2-SAP (PRIZM Pharmaceuticals, San Diego, CA) is a hybrid molecule containing FGF-2 (18 kDa) fused to the ribosome inactivating protein saporin (30 kDa). The chimeric protein was expressed in Escherichia coli from a single gene and was purified using heparin affinity and size exclusion chromatography [8]. FGF-2-SAP, a 46 kDa protein, was formulated at 0.5 mg/ml in 10 mM sodium citrate, 140 mM sodium chloride, and 0.1 mM EDTA. It binds with high affinity to FGFR [9,10]. SMC cultures were established in 96-well tissue culture plates at a seeding density of 5000 cells/well, and were fed with growth medium for 120 h

as they reached 80% confluence. SMC were made quiescent by addition of FBS-free medium for 48 h. Following the quiescence, SMC were returned to growth medium for 48 h, and at which time FGF-2-SAP (0.5, 5, 50, or 500 ng/ml), SAP alone (25, 250, 2500 ng/ml) or FGF-2 alone (1, 10, 100 ng/ml) was added to each well. Negative and positive control groups were wells containing PBS and complete growth media, respectively. Proliferation assays were performed using [3H]TdR uptake following 48 h of incubation (n=5 wells per group). EC cultures utilized 96-well tissue culture plates coated with 1.5 mg/cm2 of fibronectin (American Red Cross) for 10 min prior to EC seeding. EC were seeded at a density of 2000 cells/well and fed with growth medium for 72 h as they reached 80% confluence. EC were made quiescent by placing in FGF-free medium

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289

281

Fig. 3. Tritiated thymidine uptake of co-cultured canine EC. Experimental group (EC in EC/SMC co-cultures) is compared with control (EC in EC/EC co-cultures).

for 120 h. Following quiescence, EC were returned to growth medium for 48 h, and at which time FGF-2SAP (0.5, 5, 50, or 500 ng/ml), SAP alone (25, 250, 2500 ng/ml) or FGF-2 alone (1, 10, 100 ng/ml) was added to each well. Negative and positive control groups were wells containing PBS and complete growth media, respectively. Proliferation assay was performed using [3H]TdR uptake following 48 h of incubation (n= 5 wells per group).

2.7. Effects of FGF-2 -SAP on co-cultured EC and SMC The effect of mitotoxin FGF-2-SAP as well as unconjugated forms of FGF-2 and SAP (PRIZM) on cell growth were studied in co-cultured EC and SMC. EC/SMC and SMC/EC co-cultures were established in 24-well tissue culture plates. Once co-cultured cells reached 80% confluence, culture medium was replaced with either FGF-1-devoid medium for 120 h or FBSdevoid medium for 72 h to quiesce co-cultured EC or SMC, respectively. Following the quiescence period, co-cultures were returned to complete growth medium plus FGF-2-SAP (0.5, 5, 50 or 500 ng/ml), SAP alone (25, 250, 2500 ng/ml) or FGF-2 alone (1, 10, or 100 ng/ml). Negative and positive control groups were wells containing PBS and complete growth media respectively. Half of the co-cultured wells also received [3H]TdR (1 mCi/well). Following 48 h of growth, cocultured EC and SMC were harvested independently and their proliferation response was analyzed by either [3H]TdR incorporation (n =5 wells per group per

concentration) or fluorometric measurement (n=5 wells per group per concentration). During the co-culture experiment, cellular morphology was followed daily by phase contrast inverted microscopy.

2.8. Statistical analysis Statistical analysis was performed using two-way ANOVAs and Student’s t-tests utilizing SigmaStat 1.0 microcomputer software (Jandel Scientific, Corte Madera, CA). Results were expressed as counts per minute9 standard deviation (CPM9 S.D.) based on [3H]TdR incorporation, or total DNA content (9 S.D.) based on fluorometric measurement. Proliferation response of co-cultured SMC and EC to FGF-2-SAP mitotoxin or unconjugated FGF-2 or SAP were expressed as percent of CPM relative to positive controls or percent total DNA content relative to positive controls. A P value of less than 0.05 was considered to be significant.

3. Results

3.1. Growth kinetics of co-cultured SMC in the presence of EC Co-cultured EC stimulated SMC DNA synthesis ([3H]TdR uptake) 92% more (PB 0.01) than SMC/ SMC cultures on day 1 after growth quiescence. This stimulatory effect progressively declined as EC be-

282

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289

Fig. 4. Fluorometric DNA measurement of co-cultured canine EC. Experimental group (EC in EC/SMC co-cultures) is compared with control (EC in EC/EC co-cultures). (A) Total DNA content by fluorometric measurement from day 1 to day 6. (B) Net increase of total DNA content per 48 h.

came confluent with confluent EC appearing to inhibit SMC proliferation on day 4 (28% less than SMC/ SMC cultures, PB 0.05) (Fig. 1). With regard to total cell number, DNA contents by fluorometric measurement were significantly higher (PB 0.01) for SMC in the SMC/EC co-culture group than the SMC/SMC group comparing results from each day separately from day 1 to day 6 (Fig. 2A). The total DNA contents were converted to net DNA gain per 48 h to compare with [3H]TdR uptake during the corresponding 48 h incubation period. Similar findings were noted with SMC in SMC/EC co-cultures displaying a rapid DNA increase per 48 h initially (256% in the first 48 h period and 92% in the second 48 h period above the SMC/SMC controls, P B0.01), followed by a progressive decline in the net increase in total DNA per 48 h compared to SMC/SMC controls. (Fig. 2B). These results demonstrate that proliferating EC release soluble growth-promoting substances for SMC.

3.2. Growth kinetics of co-cultured EC in the presence of SMS SMC did not significantly affect the growth of cocultured EC compared to the EC/EC control group. EC in both EC/SMC and EC/EC groups showed a progressive increase in [3H]TdR uptake which peaked on day 3 followed by a decline on day 4 (Fig. 3). Similarly, total DNA content and the rate of DNA synthesis per 48 h did not show any significant difference between the two groups (Fig. 4A and B). These results suggest that proliferating SMC do not release rate limiting growth-promoting substances for proliferating EC.

3.3. Effects of FGF-2 -SAP on SMC or EC single-cell-type cultures When SMC were treated with FGF-2-SAP following growth quiescence, there was a concentration-

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289

283

Fig. 5. Effect of FGF-2-SAP, SAP alone, and FGF-2 alone on proliferating SMCs following quiescence.

Fig. 6. Effect of FGF-2-SAP, SAP alone, and FGF-2 alone on quiescent ECs.

dependent cytotoxicity with 37% (P B 0.05) and 49% (P B0.05) reduction in [3H]TdR uptake at the concentration of 50 and 500 ng/ml, respectively, similar to previously reported findings [4,11]. The addition of unconjugated saporin and FGF-2 to complete growth media containing 10% FBS did not have significant effects on the SMC growth (Fig. 5). By contrast, FGF2 alone induced significant SMC proliferation when added to media containing only 2% FBS (data not shown). These results indicate that potent stimulatory effects of human FGF-2 on canine SMC are similar to previous results using rat SMC [4]. When EC were exposed to FGF-2-SAP for 48 h following the release from growth arrest, there was no significant effect on proliferation. In a separate study, EC were maintained in quiescent phase while treated with FGF-2-SAP, saporin alone, or FGF-2 alone for 48 h without growth recovery in growth medium and

results compared to positive controls re-fed growth media. Under these conditions, there was a concentration-dependent cytotoxicity of FGF-2-SAP (at 0.5, 5, 50, and 500 ng/ml) with reduction of [3H]TdR uptake of 42% (PB 0.01), 52% (PB 0.01), 65% (PB 0.0001), and 87% (PB 0.0001) compared to positive controls, respectively (Fig. 6). As expected, human FGF-2 alone showed stimulatory effects on quiescent canine EC growth compared to negative control. These results suggest that FGF-2-SAP may be inducing proliferation (via the FGF-2), which, at higher concentrations, allows saporin to kill the cells. When quiescent EC were exposed to saporin alone, there was no effect on [3H]TdR uptake compared to untreated cultures.

3.4. Effect of FGF-2 -SAP on co-cultured SMC and EC In co-cultured SMC, FGF-2-SAP (at 0.5, 5, 50, and

284

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289

Fig. 7. Effect of FGF-2-SAP on co-cultured SMC and EC based on tritiated thymidine uptake.

500 ng/ml) showed a concentration-dependent cytotoxicity with [3H]TdR uptake of 84% (ns), 50% (PB 0.01), 25% (PB 0.0001), and 1% (PB 0.0001) compared to positive controls, respectively (Fig. 7). In co-cultured EC, there was a trend toward increasing [3H]TdR uptake by 106% and 113% at FGF-2-SAP concentrations of 0.5 and 5 ng/ml compared to positive controls, respectively. At higher concentrations of FGF-2-SAP of 50 and 500 ng/ml, co-cultured EC showed a decrease in [3H]TdR uptake to 59% (PB 0.01) and 2% (P B 0.0001), respectively (Fig. 7). With regard to total DNA content (Fig. 8), there was a progressive decline in co-cultured SMC DNA to 88% (ns), 66% (ns), 53% (P B0.05) and 36% (P B0.01) at FGF-2-SAP 0.5, 5, 50 and 500 ng/ml compared to positive controls, respectively. Total DNA content in co-cultured EC declined significantly only in response to 500 ng/ml FGF-2-SAP with a decrease to 49% (PB0.01) compared to positive controls. Thus a differential sensitivity to FGF-2-SAP was observed for co-cultured SMC and EC.

3.5. Effect of unconjugated FGF-2 or SAP on co-cultured SMC and EC The presence of unconjugated saporin at 25, 250, and 2500 ng/ml (1–100 nM) resulted in no change in [3H]TdR uptake in both co-cultured EC and SMC compared to their respective positive controls. Similarly SAP alone did not affect the total DNA content of either cell type compared to their respective positive controls. FGF-2 alone at concentrations of 1, 10, and 100 ng/ml (0.1–10 nM) did not have any significant effect on [3H]TdR uptake or DNA content of either SMC or EC compared to the positive controls.

3.6. Morphology The cytotoxicity was also confirmed by inverted phase contrast microscopic observation on co-cultured SMC treated with FGF-2-SAP. The wells in which co-cultured SMC were treated with FGF-2-SAP at 50 and 500 ng/ml showed a progressively reduced cell density compared to untreated wells (Fig. 9A–C).

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289

285

Fig. 8. Effect of FGF-2-SAP on co-cultured SMC and EC based on fluorometric DNA measurement.

4. Discussion Cellular interaction between cultured EC and SMC has been studied an an attempt to elucidate potential in vivo interactions that may contribute to the pathophysiology of myointimal hyperplasia. In vivo studies of the interactions between these cell types has been difficult to perform due to complex poorly understood variables in a dynamic evolving environment. The development of a co-culture model permits the close cellular interaction that exists in vivo while retaining the control of experimental variables associated with tissue culture. The co-culture model used in this study provides the potential for humoral cell – cell interactions while allowing either EC or SMC to be sampled independently for analysis. This advantage is clearly the major disadvantage of the in vitro system since cell – cell and cell–matrix interactions which may be important are not studied. Despite this limitation, useful insights can be obtained.

In our co-culture model, canine jugular veins were the preferred source to harvest EC due to the homogeneous source of endothelium provided by the veins without the added step of removing SMC when using the artery as the endothelial source. The use of venous EC provides a purer cell population in our laboratory as compared to arterial EC. In addition this cell type is of greater clinical relevance for extrapolation of data to vein bypass graft settings. We recognize the importance of additional studies using arterial EC for extrapolation to angioplasty models and using microvascular EC which may be relevant in all clinical interventions. Proliferation assays of venous EC in our laboratory demonstrated a similar growth response when EC were seeded in either single-cell-type culture or co-culture model. Once EC and SMC were harvested from the respective venous and arterial sources, our co-culture model was established by seeding these two cell types at different cell densities to ensure that both cell lines would reach similar degrees of confluence following a

286

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289

Fig. 9. Morphology of co-cultured SMC under inverted phase contrast microscopic observation. (A) Confluent SMC control group (without FGF-2-SAP) displaying the typical ‘hill and valley’ appearance. (B) Co-cultured SMC treated with FGF-2-SAP at 50 ng/ml. (C) Co-cultured SMC treated with FGF-2-SAP at 500 ng/ml showing progressively reduced cell number.

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289

co-culturing period. Following the growth quiescence of each cell type, both EC and SMC were allowed to proliferate in complete growth media. SMC showed a significant transient increase in both [3H]TdR uptake and total DNA content immediately following the release from quiescence as the result of the influence of humoral factors released by subconfluent EC. Several growth factors produced by EC have been identified which stimulate SMC proliferation including FGF-2 and platelet derived growth factor (PDGF) [12–14]. Previous studies have shown that proliferating SMC undergo a significant transient upregulation of their FGFR expression compared to either EC or quiescent SMC in vitro [3,4,11]. Balloon de-endothelialization of rat carotid arteries is followed by an exuberant medial SMC proliferation which can be significantly inhibited by infusion of neutralizing anti-FGF-2 antibody or FGF-2-SAP. Conversely, myointimal proliferation can be increased by infusion of recombinant FGF-2 [5,15]. We can only speculate on the possible mechanisms by which EC and SMC may modulate the growth kinetics in our co-culture model. Davies and Kerr studied a co-culture model using a microcarrier system separating bovine vascular EC and SMC [16]. Under these conditions, they noted a similar stimulation of SMC proliferation by EC. In our co-culture model, the rapid proliferation of SMC was transient as confluent EC exerted a progressively inhibitory effect on SMC growth. This effect has been observed by others [17– 19]. EC may release recognized growth inhibitory substances [20–27] such as active TGF-b1 [26], which in turn may either modulate PDGF and PDGFR expression [17] or inhibit SMC hill and valley formation [19], glycosaminoglycans [24] and nitric oxide [27]. EC may alter the level of glycosaminoglycans in the extracellular matrix via a humoral effect [24]. EC may increase SMC production of cyclic guanosine monophosphate (cGMP) resulting in growth inhibition [26]. EC also produce nitric oxide, via both constitutive and inducible nitric oxide synthetase, which activates soluble guanylate cyclase and elevates the levels of intracellular cGMP to inhibit SMC proliferation [27]. The early stimulatory effect on SMC growth by subconfluent EC in this co-culture model provides a model to study potential means of inhibiting SMC growth following EC injury in vivo and to further our understanding of myointimal hyperplasia. FGF-2, as stated above, has been suggested to play a dominant role in mediating the first wave of SMC proliferation following injury [5,15]. Reduction of intimal hyperplasia by delivery of an anti-FGF-2 antibody has been reported, but there is concern regarding possible systemic toxicity and clinical applicability [4]. The SMC which respond to balloon injury become responsive to the mitogenic effects of FGF-2 in part as a result of a transient increase of high affinity cell surface FGFR

287

compared to EC [3,28]. This selectivity in FGFR expression in SMC following injury remains appealing since FGF-2-SAP selectively kills cells which express large numbers of FGFR [11,29,30]. Internalized FGF2-SAP is trafficked to lysosomes where it is degraded, and free SAP is released into the cytoplasm where it inhibits ribosomal function and protein synthesis resulting in cell death. Previous studies using different cell types with various degrees of FGFR expression have shown a direct relationship between receptor number and the cytotoxicity of the mitotoxin [2,11]. These findings prompted this study in which we examined the potential selective cytotoxicity of FGF2-SAP on EC and SMC in single cell cultures as well as in a co-culture model. In the present study, exponentially growing canine SMC were susceptible to the cytotoxicity of FGF-2SAP in a concentration-dependent fashion via increased FGFR binding. The response of canine EC to FGF-2SAP was less consistent in that the mitotoxin failed to show a significant concentration-dependent cytotoxicity on proliferating EC. Under the experimental conditions when EC were released from growth arrest by restoration of FGF in the growth media, decreased proliferation may have resulted from competition with FGF-2-SAP for receptor occupancy. When low concentrations of FGF-2-SAP were added to quiescent EC the FGF-2 portion of the mitotoxin demonstrated a stimulatory effect on quiescent EC. As the concentration of FGF-2-SAP increased, saporin appears to reach a critical level which causes significant cytotoxic effects on EC. Unconjugated saporin alone, when added to quiescent EC, failed to show any effect on the EC growth indicating that the mitotoxin is entering EC via FGFR. As expected, the addition of FGF-2 alone to quiescent EC resulted in significant proliferation based on [3H]TdR uptake. In our co-culture model, we found a large dose–response range of FGF-2-SAP which was cytotoxic to proliferating SMC but not EC. Higher concentrations of FGF-2-SAP (50 and 500 ng/ml) were lethal to both SMC and EC as both [3H]TdR uptake and total DNA contents were significantly reduced. The direct correlation of data comparing DNA content and [3H]TdR incorporation suggest the response to be a reduction in cell number rather than a simple decrease in cell cycling. The decrease in cell number suggested by total DNA content was consistent with visual observations in the cultures. The resistance of EC to FGF-2-SAP at concentrations which were cytotoxic to SMC may be explained by two possible mechanisms: (1) EC may possess a more efficient detoxifying mechanism than SMC, and (2) EC have been found to have a fourfold lower density of cell surface FGFR than proliferating SMC [11]. Once the mitotoxin is internalized intracellularly, it must undergo lysosomal degradation so that

288

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289

the SAP toxin can reach the ribosomes. EC, with relatively fewer cell surface FGFR than SMC, may not take up enough FGF-2-SAP to permit sufficient intracellular SAP activity. Although the ribosomal inactivation proceeds enzymatically, presumably a threshold number of SAP molecules are required for cell death to occur. Recent studies in our laboratory have also demonstrated that FGF-2-SAP has almost no effect on postconfluent SMC (H.P. Greisler and P.H. Lin, unpublished results, 1995) which contain a low density of FGFR. Taken together, these findings are consistent with the notion that the cytotoxicity of FGF-2-SAP is related to the degree of FGFR expression on different cell types. It is interesting to note that the addition of SAP toxin or FGF-2 alone did not affect the growth of either co-cultured EC or SMC even at the highest concentrations tested (2500 ng/ml SAP and 100 ng/ml FGF-2). This confirms that SAP alone has minimal toxicity because it is not receptor-bound and therefore does not become internalized intracellularly to inhibit ribosomes. The observation that FGF-2 alone displayed no stimulatory effects on either EC or SMC growth may be explained by the composition of the growth medium which resulted in maximal growth stimulation. Further studies are under way in our laboratory to investigate any possible growth stimulatory effect on EC by low concentrations of FGF-2-SAP using growth media inducing only half maximal proliferation response. This selective cytotoxicity of FGF-2-SAP against proliferating SMC compared to EC in our co-culture model raises the possibility of a therapeutic window for the application of FGF-2-SAP in vivo, in that it might be possible to selectively kill SMC without affecting EC. It is important to point out, however, that the in vivo environment is considerably more complex. The prospects for such a therapy via local application of mitotoxin are more promising than systemic administration due to concern over possible toxicity. Clearly, further studies are needed in testing this strategy for inhibiting myointimal hyperplasia to determine whether the success achieved in this co-culture model can also be achieved in vivo.

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12] [13]

[14]

[15]

[16]

[17]

Acknowledgements [18]

This work was funded in part by a grant from the National Institutes of Health, RO1 HL41272. [19]

References [20] [1] Stirpe F, Gasperi-Campani A, Barbieri L, Flalsca A, Abbondaza A, Stevens WA. Ribosome-inactivating proteins from seeds of

Saponaria officinalis L. (soapwart) of Agrostemma githago L. (corn cockle) and of Asparagus officinalis L. (asparagus), and from the latex of Hura crepitans L. (sandbox tree). Biochem J 1983;216:617 – 9. Lappi DA, Maher PA, Martineau D, Baird A. The basic fibroblast growth factor – saporin mitotoxin acts through the basic fibroblast growth factor receptor. J Cell Physiol 1991;147:17–26. Biro S, Siegall CB, Fu M, Speir EH, Pastan I, Epstein SE. In vitro effects of a recombinant toxin targeted to the FGF receptor on rat vascular smooth muscle and endothelial cells. Circ Res 1991;71:640 – 5. Casscells W, Lappi DA, Olwin BB, et al. Elimination of smooth muscle cells in experimental restenosis: targeting of fibroblast growth factor receptors. Proc Natl Acad Sci U S A 1992;89:7159 – 63. Lindner V, Lappi DA, Baird A, Majack RA, Reidy MA. Role of basic fibroblast growth factor in vascular lesion formation. Circ Res 1991;68:106 – 13. Kang SS, Gosselin C, Ren D, Greisler HP. Selective stimulation of endothelial cell proliferation with inhibition of smooth muscle cell proliferation by FGF-1 plus heparin delivered from fibrin glue suspensions. Surgery 1995;117:280 – 7. Greisler HP, Johnson S, Joyce K, et al. The effects of shear stress on endothelial cell retention and function on expanded polytetrafluoroethylene. Arch Surg 1990;125:1622– 5. McDonald JR, Ong M, Shen C, et al. Large scale purification and through characterization of E. Coli expressed basic fibroblast growth factor – saporin mitotoxin. FASEB J 1995;9:410–4. Lappi DA, Martineau D, Baird A. Biological and chemical characterization of basic FGF – saporin mitotoxin. Biochem Biophys Res Commun 1989;160:917 – 23. Lappi DA, Ying W, Barthelemy I, et al. Expression and activities of a recombinant basic fibroblast growth factor–saporin fusion protein. J Biol Chem 1994;269:12552– 8. Biro S, Lappi DA, Yu ZX, Baird A, Casscells W. Stimulation and inhibition of endothelial and smooth muscle cells by basic fibroblast growth factor – saporin in vitro and in vivo. Circulation 1991;84(Suppl. 2):553. Gajdusek C, DiCorleto PE, Ross R, Schwartz SM. An endothelial cell-derived growth factor. J Cell Biol 1980;85:467 –72. Ross R, Masuda J, Raines EW. Cellular interactions, growth factors, and smooth muscle proliferation in atherogenesis. Ann N Y Acad Sci 1990;598:102 – 12. McNeil PL, Muthukrishnan L, Warder E, D’Amore PA. Growth factors are released by mechanically wounded endothelial cells. J Cell Biol 1989;109:811 – 22. Lindner V, Reidy MA. Proliferation of smooth muscle cells after vascular injury is inhibited by an antibody against basic fibroblast growth factor (bFGF). Proc Natl Acad Sci U S A 1991;88:3739 – 43. Davies PF, Kerr C. Cocultivation of vascular endothelium and smooth muscle cells using microcarrier techniques. Exp Cell Res 1982;141:455 – 9. Majack RA, Majesky MW, Goodman L. Role of PDGF-A expression in the control of vascular smooth muscle cell growth by transforming growth factor-b. J Cell Biol 1990;111:239–47. Fillinger MF, O’Connor SE, Wagner RJ, et al. The effect of endothelial cell co-culture on smooth muscle cell proliferation. J Vasc Surg 1993;17:1058 – 67. Powell RJ, Cronenwett JL, Fillinger MF, Wagner RJ. Effect of endothelial cells and transforming growth factor-b1 on cultured vascular smooth muscle cell growth patterns. J Vasc Surg 1994;20:787 – 94. Castellot J, Addonizio ML, Rosenberg R, Karovsky MJ. Cultured endothelial cells produce a heparin like inhibitor of smooth muscle cell growth. J Cell Biol 1981;90:372 – 9.

P.H. Lin et al. / Atherosclerosis 137 (1998) 277–289 [21] Herman IM. Endothelial cell matrices modulate smooth muscle cell growth, contractile phenotype and sensitivity to heparin. Hemostasis 1990;20:166–77. [22] Haudenschild CC, Schwartz SM. Endothelial regeneration: II. Restitution of endothelial continuity. Lab Invest 1979;41:407 – 18. [23] Reidy MA, Schwartz SM. Endothelial regeneration: III. Time course of intima changes after small defined injury of rat aortic endothelium. Lab Invest 1981;44:301–8. [24] Merrilees MJ, Scott L. Interaction of aortic endothelial and smooth muscle cells in culture: effect on glycosaminoglycan levels. Atherosclerosis 1981;39:147–61. [25] Fritze LMS, Reilly CF, Rosenberg RD. An antiproliferative heparan species produced by postconfluent smooth muscle cells. J Cell Biol 1985;100:1041–9.

.

289

[26] McNeil PL, Muthukrishnan L, Warder E, D’Amore PA. Growth factors are released by mechanically wounded endothelial cells. J Cell Biol 1989;109:811 – 22. [27] Schini VB, Busse R, Vanhoutte PM. Inducible nitric oxide synthetase in vascular smooth muscle. Drug Res 1994;44: 432 – 6. [28] Epstein SE, Siegall CB, Biro S, Fu YM, FitzGerald D, Pastan I. Cytotoxic effects of a recombinant chimeric toxin on rapidly proliferating vascular smooth muscle cells. Circulation 1991;84:778 – 87. [29] Casscells W, Lappi DA, Baird A. Molecular atherectomy for restenosis. Trends Cardiovasc Med 1993;3:235 – 43. [30] Lappi DA, Baird A. Mitotoxins: growth factor-targeted cytotoxic molecules. Prog Growth Factor Res 1990;2:223 – 36.