Dependence of Augmentation of Arterial Endothelial Cell Expression of Plasminogen Activator Inhibitor Type 1 by Insulin on Soluble Factors Released From Vascular Smooth Muscle Cells
Background Insulin-resistant states are characterized by accelerated atherosclerosis and are associated with increased plasma concentrations of insulin and plasminogen activator inhibitor type 1 (PAI-1). To determine whether arterial expression of PAI-1 in response to insulin contributes to the increased PAI-1 observed, human and porcine arteries in culture were exposed to insulin, and results were compared with responses of specific arterial cellular constituents maintained in culture and coculture.
Methods and Results Human and porcine arterial segments and cells obtained from arteries were maintained in culture. Insulin increased accumulation of PAI-1 in conditioned medium from arterial segments (ng PAI-1 [1 nmol/L insulin minus control]: human arteries 47±17, porcine arteries 3.1±1.2, P<.05 for each) and from endothelial cells (ECs) cocultured with smooth muscle cells (SMCs, ng PAI-1 [1 nmol/L insulin minus control]: human cells 43±8, porcine cells 0.5±0.1, P<.05 for each). Insulin had no effect on EC expression of PAI-1 when not cocultured with SMCs. Increased accumulation of PAI-1 was seen when ECs, in coculture chambers without SMCs, were cultured with medium previously conditioned by SMCs in the presence of insulin. The increased accumulation of PAI-1 in conditioned medium was secondary to both an increased transport of PAI-1 from the basal to the apical surface of ECs as well as an increased production of PAI-1 by ECs.
Conclusions Insulin augments arterial expression of PAI-1 by stimulating release of a soluble factor(s) from SMCs. Accordingly, increased arterial elaboration of PAI-1 in response to insulin is likely to account, in part, for the elevated PAI-1 observed in the blood of subjects with insulin-resistant states.
Insulin-resistant states, such as those seen with type II diabetes mellitus, obesity, and hypertension, are associated with increased prevalence and mortality associated with atherosclerosis.1 2 3 4 Decreased fibrinolytic capacity is seen as well, secondary to increased expression of the primary physiological inhibitor of plasminogen activators, PAI-1.5 6 7 8 9 10 The increased concentrations of PAI-1 correlate positively with increased concentrations of immunoreactive insulin.5 6 7 8 9 10 We and others have hypothesized that the positive correlation between immunoreactive insulin and PAI-1 in plasma is secondary to direct effects of insulin on cellular expression of PAI-1. Such effects have been seen with HepG2 cells, a highly differentiated human hepatocellular carcinoma cell line.11 12 13 14 15
Vascular wall cells appear to be the primary source of both circulating plasminogen activators and PAI-1.16 17 The clearance of PAI-1 from the circulation is rapid, with a half-life for the first phase of clearance of <9 minutes.18 Thus, local arterial production of both plasminogen activators and PAI-1 is likely to be pivotal in influencing thrombosis and its persistence in response to vascular injury.
Observations of expression of PAI-1 by ECs in response to insulin and insulin precursors have been conflicting.14 19 20 21 We have hypothesized that the inconsistencies may be related to the differential presence of contaminating vascular SMCs in some endothelial cell cultures and potential interactions between diverse cell types.21 The present study was designed to define effects of insulin on the expression of PAI-1 in segments of arteries, isolated arterial cellular constituents, and cocultured cellular constituents of different types exposed to pathophysiological concentrations of insulin.
In protocols approved by the University of Vermont Institutional Review Boards, human arteries were obtained during surgery from tissue removed for clinical indications. Porcine arteries were obtained immediately after death from pigs in a local slaughterhouse. Arteries were placed immediately in Hanks’ buffer with 300 U/mL penicillin and 300 μg/mL streptomycin. Subsequent processing was performed in a laminar flow hood.
Connective tissue was trimmed from the adventitial surface of the vessel. Arteries were then cut into rings 2 to 5 mm wide. Large arteries such as the aorta were cut into full-thickness segments 3 to 5 mm square. The arteries were then cultured in DMEM with Ham’s nutrient mixture F12 (DME/F12, Gibco BRL) with 1% BSA (Sigma Chemical Co), 50 U/mL penicillin, and 50 μg/mL streptomycin at 37°C in atmosphere enriched with 5% CO2. Media were exchanged daily.
Cell Isolation and Culture
ECs were isolated by exposure of the luminal wall of arterial segments to collagenase (Worthington), 1 mg/mL in Hanks’ buffer, for 30 minutes. The intimal surface was then scraped with a scalpel blade in a fibronectin-coated culture plate. Adherent ECs were cultured in M199 (Gibco BRL) with 20% FCS (Gibco BRL), 50 μg/mL ECGS (a mixture containing fibroblast growth factor and endothelial cell growth factors-α and -β, Collaborative Biomedical Research), 80 μg/mL heparin (Sigma), 50 U/mL penicillin, and 50 μg/mL streptomycin in atmosphere enriched with 5% CO2. Experiments were performed with monolayers of cells in DME/F12 with 1% BSA and 50 μg/mL ECGS. Cells were characterized immunohistochemically as endothelial by expression of von Willebrand factor with a primary polyclonal rabbit antibody to von Willebrand factor. All immunohistochemistry was performed by the cell imaging core facility at the University of Vermont. Cultures were >90% endothelial as defined by immunohistochemistry. ECs were used in passages 2 through 6 to minimize the effect of dedifferentiation observed with prolonged culture.22
SMCs were isolated by explantation. The intimal surface was dissected from the artery, and the arterial media was cut into pieces ≈1×1 mm. The tissue was placed on the etched surface of tissue culture dishes, and a microscopic coverslip was used to anchor the tissue in place. Cells were cultured in DMEM (Gibco BRL) with 20% FCS, 50 U/mL penicillin, and 50 μg/mL streptomycin in an atmosphere enriched with 10% CO2. Experiments were performed on cells in a monolayer in DME/F12 with 1% BSA in an atmosphere enriched with 5% CO2. Cells were characterized as smooth muscle by immunostaining with a primary mouse monoclonal antibody directed against smooth muscle actin. Cultures were >90% smooth muscle as defined by immunohistochemistry. SMCs were used in passages 2 through 8 to minimize the effect of dedifferentiation observed with prolonged culture.
Coculture of Cells
Six-well tissue culture plates and culture inserts that allow communication between the upper and lower chambers through 0.45-μm pores on the basal surface (culture membrane) of the insert were purchased from Collaborative Biomedical Research. ECs and SMCs isolated from human and pig arteries were treated with versene and trypsin (Gibco BRL) to yield cell suspensions and were quantified with a hemacytometer. Equal numbers were allowed to adhere to culture plates and insert membranes in endothelial or smooth muscle medium as described above. Cell numbers were expanded in serum-containing medium without intermixing of cell types. Before experiments, both ECs and SMCs were washed with PBS and cultured in DME/F12 with 50 μg/mL ECGS in an atmosphere enriched with 5% CO2 for 16 to 24 hours. Subsequently, the medium was replaced with fresh medium, and combining of cell types was induced by moving the inserts. ECs were cultured on the inserts (upper chamber) and SMCs were cultured on the wells (lower chamber) for all of the coculture experiments described, and the media in both the upper and lower chambers contained ECGS.
To characterize whether postconfluent cells responded differently from newly confluent cells, both ECs and SMCs were grown to confluence in growth medium as defined above. The growth medium was then replaced with DME/F12 with 2% heat-inactivated FCS. The medium to which ECs were exposed had 50 μg/mL ECGS added as well. After 6 and 9 days, medium was removed, the cells were washed with PBS, and DME/F12 with 50 μg/mL ECGS was added to all wells. Experiments were performed the following day in fresh DME/F12 with 50 μg/mL ECGS.
Insulin was purchased from Sigma and reconstituted in DME/F12 with 1% BSA. The concentration was determined by radioimmunoassay. Media and reagents regularly screened for endotoxin contamination with the limulus amoebocyte lysate assay (Associates of Cape Cod) did not exhibit contamination >0.01 ng/mL (100-fold less than concentrations required to stimulate expression of PAI-123 ).
Extraction of Tissue Protein From Arterial Rings
Arterial rings were washed three times in PBS and processed at 4°C. After wet weight was obtained, the tissues were pulverized in liquid nitrogen and homogenized in RIPA buffer (10 mmol/L Tris [pH 7.4], 150 mmol/L NaCl, 1% Nonidet P40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mmol/L iodoacetamide, and 1 mmol/L PMSF [Sigma]). Cellular debris was removed by centrifugation (12 000g for 15 minutes). The supernatant fraction was stored at −20°C until assays were performed. Total protein was quantified conventionally with the method of Bradford.24 Albumin, quantified by colorimetric assay with bromcresol green (Sigma), in tissue lysates accounted for <0.02% of the total protein. The DNA content in tissue lysates was quantified with the use of a DNA-binding fluorochrome H33258.25 Concentrations of DNA were determined by comparing the emission wavelength at 450 nm with that of a known amount of calf thymus DNA (excitation wavelength, 365 nm).
PAI-1 was measured by ELISA as described previously.26 Antibodies were provided by Professor Désiré Collen from the Catholic University of Leuven. The interassay coefficient of variation was 3.6%, and the intra-assay coefficient of variation was 2.5%. Functional activity of PAI-1 protein is labile, particularly in non–serum-containing media. Accordingly, all determinations of fibrinolytic system protein concentrations were made by ELISA. Porcine PAI-1 protein is detected by the ELISA with decreased sensitivity. Despite this decreased sensitivity, a linear response to increasing amounts of porcine PAI-1 was observed. Thus, relative changes in PAI-1 protein can be determined accurately, but the concentration of PAI-1 protein measured in samples from porcine specimens underrepresents the total mass of protein.
ECs or SMCs were grown to confluence, and after 16 hours in DME/F12 with ECGS followed by 30 minutes in DME/F12 with ECGS devoid of methionine, they were exposed to 0.1 mCi/mL of 35S-methionine (Transmet, ICN) added to DME/F12 with ECGS and devoid of methionine. After 1 hour, the medium containing the Transmet was removed, and the cells were washed three times with PBS before the wells and inserts were combined in coculture and exposed to conditions. The media were collected 24 hours later. PAI-1 was immunoprecipitated with a goat polyclonal antibody against human PAI-1 (American Diagnostica Inc). The antibody-antigen complexes were isolated with protein G Sepharose (Pharmacia) and separated from conditioned media with centrifugation (15 000g for 2 minutes). After three washes, the protein G, antibody, and antigen (PAI-1) were separated by addition of reducing buffer and boiled for 3 minutes, and the proteins were separated with SDS-PAGE.27 Radioactivity of PAI-1 was quantified with radioisotopic scanning, and autoradiograms were exposed at −70°C.
PAI-1 mRNA Expression
Total cellular RNA was isolated with the use of phenol-chloroform extraction28 with TRI reagent (Molecular Research Center, Inc). To obtain an adequate amount of RNA, RNA isolated from three wells was combined. RNA was isolated from confluent monolayers of cells exposed to controlled conditions or insulin for specific intervals. The concentration of RNA was quantified with UV spectroscopy (260 nm). PAI-1 mRNA was quantified with dot blot analysis.29 RNA was applied to Gene Screen (DuPont New England Nuclear) under vacuum in a solution containing 6×SSC (0.9 mol/L sodium chloride, 0.09 mol/L sodium citrate) and 7.4% formaldehyde. The total amount of RNA (3 μg) applied to each dot was kept constant by the addition of yeast tRNA (Sigma). Yeast tRNA was applied as a control to characterize nonspecific binding of the cDNA probe. The RNA was cross-linked to the Gene Screen with UV light. Hybridization of membranes was performed as previously described with a 0.9-kb EcoRI, Sal I fragment of the cDNA probe for PAI-1.11 12 Total radioactivity in hybridized dots was quantified by radioisotopic scanning, and autoradiograms were exposed at −70°C.
Determination of significance of differences between paired samples was performed with a paired t test. Differences between groups were evaluated with Student’s t tests and ANOVA. Values are mean±SEM.
Arterial Organ Culture
Arterial rings were viable in culture for at least 4 days, as evidenced by their morphological appearance and continued synthesis of proteins. Protein synthesis rates were determined by the incorporation, over 1 hour, of 35S-methionine into the proteins precipitated by trichloroacetic acid (10% vol/vol). Arterial rings maintain 76±4% of their incorporation of 35S-methionine through 4 days in culture with respect to incorporation determined immediately after procurement of the specimen.
As judged from macroscopic appearance, human arteries were classified as “normal” or atherosclerotic. Arteries were further classified on the basis of whether the subject from whom the arterial specimen was taken had insulin resistance, defined by obesity (body mass index >28 kg/m2), type II diabetes, or hypertension. All of the atherosclerotic specimens studied were obtained from subjects with insulin resistance. Thus, three groups of specimens were analyzed: morphologically normal rings from subjects without insulin resistance (n=3 subjects), morphologically normal rings from subjects with insulin resistance (n=3 subjects), and severely atherosclerotic rings from subjects with insulin resistance (n=10 subjects). In parallel studies, coronary arteries were isolated from porcine hearts (n=7 animals). All of the porcine specimens were normal in macroscopic appearance and were from pigs weighing 100 to 140 kg.
To characterize responses with respect to similar numbers of cells, DNA was quantified in tissue lysates prepared after organ culture. All results presented were normalized to arterial ring content of DNA.
Accumulation of PAI-1 in 24-hour conditioned medium was increased when “normal” human and porcine arteries were exposed to a pathophysiological concentration (1 nmol/L) of insulin (Fig 1⇓). The effect of insulin was similar with morphologically normal arteries from subjects without insulin resistance (ratio of insulin/control for accumulation of PAI-1, 2.1±0.1, P<.001 compared with control) and with porcine coronary arteries (ratio of insulin/control for accumulation of PAI-1, 1.5±0.1, P=.013 compared with control). In the paired analysis, the increased accumulation of PAI-1 in response to 1 nmol/L insulin was 47±17 ng PAI-1/10 ng DNA (P<.02) for the human arteries and 3.1±1.2 ng PAI-1/10 ng DNA (P<.001) for the porcine arteries.
An essential feature of insulin-resistant states is a diminished end-organ response to a given concentration of insulin. Accordingly, we characterized the response of arteries taken from subjects with insulin resistance (Fig 2⇓). The 24-hour accumulation of PAI-1 was increased in response to insulin from macroscopically normal arteries from subjects with insulin resistance (ratio of insulin/control for accumulation of PAI-1, 1.6±0.1, P<.001) and from atherosclerotic arteries from subjects with insulin resistance (ratio of insulin/control for accumulation of PAI-1, 1.5±0.2, P<.03). As can be seen in Fig 2⇓, the increment in PAI-1 in response to insulin was less uniform in atherosclerotic arteries compared with macroscopically normal arteries. The insulin-induced increment in expression of PAI-1 from “normal” arteries from subjects with insulin resistance was 28±11 (P<.03) and 21±7 ng PAI-1/10 ng DNA (P<.01) from atherosclerotic arteries from subjects with insulin resistance.
Arterial Cells in Culture
ECs and vascular SMCs of human and porcine origin were maintained in monolayer culture and exposed to insulin. The 24-hour accumulation of PAI-1 in conditioned media was similar under control conditions and in response to insulin when ECs were cultured without admixture with SMCs. To simulate the architecture of arteries, ECs were cocultured with SMCs. The coculture system is defined as ECs cultured on a porous membrane in the upper chamber and SMCs cultured in the lower chamber. Insulin or vehicle (control) was added only to the upper chamber. Increased accumulation of PAI-1 in the upper chamber was observed in response to insulin when ECs were cocultured with SMCs (Fig 3⇓). The dose response demonstrated a plateau of effect for concentrations of insulin >10 nmol/L. In each case, cell numbers were quantified to determine that differences in cell numbers did not account for the increased accumulation of PAI-1.
The microscopic appearance of ECs and SMCs did not change during their coculture. To further verify that no substantial change in cell number had occurred, the protein and DNA contents were determined in cell lysates from ECs and SMCs before and after 24 hours in coculture. The content of protein and DNA in cell lysates remained stable during the experimental interval (mg protein before/after ECs, 0.52±0.03/0.52±0.03; SMCs, 0.48±0.02/0.48±0.01; ng DNA before/after ECs, 8±1/7.6±0.6; SMCs, 6.4±0.8/6.2±0.2).
All coculture experiments were performed in serum-free medium (DME/F12) with 50 μg/mL ECGS and no further addition so as to minimize confounding variables. To determine whether ECGS modulated the effect of insulin on elaboration of PAI-1, experiments were performed with variable amounts of ECGS. The effect of insulin was not different when the concentration of ECGS was altered (insulin-induced increment in PAI-1 [1 nmol/L insulin minus control]: with 25 μg ECGS/mL, 0.4±0.1; with 50 μg ECGS/mL, 0.4±0.2; with 100 μg ECGS/mL, 0.5±0.3 ng/mL).
To determine whether newly confluent ECs and SMCs responded to insulin in a manner different from postconfluent cells, the cells were exposed to conditions 1, 7, and 10 days after visual determination of confluence. A similar augmentation was observed in porcine cells 1 day postconfluency, 7 days postconfluency, and 10 days postconfluency (1 nmol/L insulin-induced increment in PAI-1 [1 nmol/L insulin minus control]: 1 day postconfluency, 1.3±0.6; 7 days postconfluency, 1.3±0.3; 10 days postconfluency, 1±0.4 ng/mL).
Coculture of ECs With Medium Previously Conditioned by SMCs
Results in experiments with arterial segments in culture and arterial cellular constituents in coculture suggested that SMCs are a necessary component mediating effects of insulin on EC expression of PAI-1. To characterize the interaction between ECs and SMCs, ECs were exposed to medium previously conditioned by SMCs (previously conditioned medium will be referred to as CM hereafter) (without SMCs in the coculture chamber). These experiments were performed by conditioning DME/F12 (without ECGS) for 24 hours with SMCs. After this 24-hour interval, ECGS (50 μg/mL) was added to the CM before addition to the lower chamber.
Insulin did not change the 24-hour accumulation of PAI-1 when a confluent monolayer of porcine ECs was cultured in the upper chamber without SMCs (Table⇓ and Fig 4⇓). The culture of ECs (without SMCs) in coculture chambers with CM added to the lower chamber resulted in an increased 24-hour accumulation of PAI-1 in the upper chamber (ECs alone, 3.1±0.3; ECs cultured with SMC-CM, 4.3±0.2 ng/mL; n=3 for each; P<.05). The addition of insulin to the upper chamber when ECs were cocultured with SMC-CM did not further increase the accumulation of PAI-1 (Table⇓ and Fig 4⇓). By contrast, when ECs were cultured in coculture chambers with medium conditioned by SMCs in the presence of insulin (SMC-ICM), the accumulation of PAI-1 in the upper chamber was augmented (ECs with SMC-CM, 4.3±0.2 ng/mL; ECs with SMC-ICM, 5.3±0.2 ng/mL; n=3 for each; P<.01). This effect was similar to that observed with the coculture of ECs and SMCs with and without insulin (ECs with SMCs, 3.7±0.1; ECs with SMCs with 10 nmol/L insulin, 4.5±0.2 ng/mL; n=3 each; P<.05).
Similar to the effect seen with medium previously conditioned by SMCs, medium previously conditioned by ECs (EC-CM), when added to the lower chamber, augmented the accumulation of PAI-1 in the upper chamber (Table⇑). The addition of insulin to the upper chamber during coculture of ECs (upper chamber) with EC-CM did not further augment accumulation of PAI-1. In contrast to results with medium previously conditioned by SMCs, no additional augmentation of PAI-1 accumulation was seen when medium previously conditioned by ECs in the presence of insulin was added to the lower chamber (Table⇑). Thus, increased expression of PAI-1 by ECs in response to insulin requires a soluble factor(s) produced specifically by SMCs.
ECs are known to transport insulin from the apical to the basal surface.33 To confirm that SMCs are exposed to insulin in our coculture system, insulin and albumin were added to the upper chamber of coculture chambers without cells on the upper membrane, with subconfluent (80% confluent) ECs on the upper membrane, and with confluent ECs on the upper membrane. After addition of 1.6 mg/mL albumin to the upper chamber, the concentration of albumin (mg/mL) in the lower chamber 24 hours later was 0.31±0.01 without cells, 0.27±0.02 with subconfluent ECs, and 0.13±0.02 with confluent ECs. After addition of 265 μIU/mL insulin to the upper chamber, the concentration of insulin (μIU/mL) in the lower chamber 24 hours later was 151±17 without cells, 147±9 with subconfluent ECs, and 80±5 with confluent ECs. The ratio of the concentration of insulin in the lower chamber to the concentration of insulin in the upper chamber was 0.92±0.05 without cells, 0.91±0.07 with subconfluent ECs, and 1.55±0.12 with confluent ECs. The increased concentration of insulin in the lower chamber relative to the upper chamber observed with confluent ECs is consistent with active transport of insulin by ECs from their apical to basal surface.
Distribution of PAI-1 in Coculture Chambers
To define whether increased elaboration of PAI-1 by ECs, the transfer of PAI-1 from the lower to upper chamber, or both were responsible for the increased accumulation of PAI-1 seen in the upper chamber, the distribution of PAI-1 in both the upper and lower chambers was determined. Increased accumulation of PAI-1 was observed in 24-hour medium conditioned by porcine SMCs in the presence of insulin (10 nmol/L insulin minus control, 1±0.3 ng/mL; P<.05; n=4). After subtraction of the amount of PAI-1 present in conditioned media (CM and ICM), an increment in the total amount of PAI-1 in both the upper and lower chambers was seen (increment in total PAI-1 [ng]: ECs with CM in lower chamber, 13.4±0.4; ECs with ICM in lower chamber, 19.2±0.8; P<.001 compared with ECs with CM). Thus, increased elaboration of PAI-1 by ECs accounts for a substantial proportion of the increased PAI-1 observed in the upper chamber after exposure to insulin.
Change in the Concentration of PAI-1 Over Time
Insulin augmented the accumulation of PAI-1 in both the upper and lower chambers when ECs and SMCs were cocultured. Because the effect of insulin is mediated by a direct effect of insulin on SMCs, these experiments were performed by addition of insulin to both the upper and lower chambers. A significant increase in the accumulation of PAI-1 in both upper and lower chambers was observed beginning 4 hours after exposure to 10 nmol/L insulin (Fig 5⇓). Accumulation of PAI-1 in the upper chamber composed 29±2% of the total (upper+lower) PAI-1 present under control conditions. The insulin-induced increment in the accumulation of PAI-1 in the upper chamber was greater after 2 to 16 hours compared with the insulin-induced increment in the lower chamber (insulin/control: upper, 2±0.2; lower, 1.4±0.1; P<.05).
Directional Release of PAI-1 and Transport of PAI-1 by ECs
To characterize the directional elaboration of PAI-1 by ECs and the fate of PAI-1 produced by SMCs, proteins synthesized by cells were labeled before combination in coculture or exposure to insulin. These experiments were designed to characterize the fate of PAI-1 synthesized before ECs had been combined with SMCs and before exposure to selected conditions. After 35S-methionine labeling of proteins produced by ECs, 77±2% of the radiolabeled PAI-1 was detected in the upper chamber under control conditions and 72±2% (n=6 for each, P=NS, Fig 6⇓) of the radiolabeled PAI-1 was detected in the upper chamber after exposure to 10 nmol/L insulin added to the upper chamber only. Thus, insulin does not appear to alter the directional elaboration of PAI-1 from intracellular stores in ECs.
When the proteins produced by SMCs were labeled with 35S-methionine, labeled PAI-1 was detected in the upper chamber after coculture with a confluent monolayer of ECs in the upper chamber. Under control conditions, 18±3% (ratio of cpm upper to cpm lower) of radiolabeled PAI-1 was detected in the upper chamber. By contrast, 30±2% (n=6 for each, P<.01) of radiolabeled PAI-1 was detected in the upper chamber after exposure of the cells to 10 nmol/L insulin (added to the upper chamber only). An increase in the radiolabeled PAI-1 detected in the upper chamber was associated with a decrease in the radiolabeled PAI-1 detected in the lower chamber (Fig 6⇑). Although absolute quantification is limited with this technique, the results suggest that ECs transport PAI-1 that has been produced by the SMCs and that insulin modulates this transport, resulting in augmented accumulation of PAI-1 in the upper chambers. On the basis of results with the transfer of media previously conditioned by SMCs, this effect appears to play a modest role in the increment in PAI-1 observed in the upper chamber after exposure to insulin.
Effect of Insulin on PAI-1 mRNA Expression
Dot blot analysis of PAI-1 mRNA expression was performed on total cellular RNA isolated from a confluent monolayer of porcine ECs in coculture with a confluent monolayer of porcine SMCs after exposure to control conditions or 10 nmol/L insulin (added to both the upper and lower chambers) for specific intervals (Fig 7⇓). RNA from three wells was combined, and results reflect the average of six wells. Increased expression of PAI-1 mRNA in response to insulin was observed in RNA isolated from ECs (average of 2, 4, and 16 hours: control, 2±0.2 cpm; insulin, 2.8±0.2 cpm; P<.01). In addition, increased expression of PAI-1 mRNA was observed in RNA isolated from SMCs (average of 2, 4, 16, and 24 hours: control, 3.6±0.7 cpm; insulin, 5.1±0.8 cpm; P<.001). Thus, insulin not only modulates the directional transport of PAI-1 by ECs but also increases the production of PAI-1 protein by ECs and SMCs.
Insulin, in concentrations present in blood in subjects with type II diabetes and other insulin-resistant states,30 31 32 increased the expression of PAI-1 by intact arterial rings in culture. Our results show that insulin augments EC production of PAI-1 and that its effect on ECs requires mediators elaborated from vascular SMCs. Thus, the results help to explain the variability that has led to conflicting observations with diverse preparations of ECs in different studies in which homogeneity of cell types undoubtedly differed. In addition, the results suggest that altered EC expression of PAI-1 contributes to increased PAI-1 in blood in patients with insulin resistance and resultant hyperinsulinemia.
As judged from the results in aggregate, the augmented accumulation of PAI-1 in response to insulin was mediated by both ECs and SMCs. Effects on ECs required interaction involving SMCs. ECs exposed to insulin transport the insulin to their basal surface.33 This effect was confirmed in our coculture system. Thus, increased amounts of insulin in the lumen of arteries is accompanied by increased concentrations of insulin in the interstitial fluid surrounding SMCs. As shown in the present study, the exposure of vascular SMCs to insulin augments production and release of PAI-1 and results also in the release of a soluble factor(s) that augments EC expression of PAI-1. Accordingly, augmented arterial release of PAI-1 into blood in the setting of hyperinsulinemia is likely to account, in part, for the increment of PAI-1 measured in the blood of subjects with insulin-resistant states. Increased elaboration of PAI-1 by hepatocytes11 12 13 14 15 and adipocytes34 35 has been shown to contribute as well.
The increased concentration of PAI-1 measured in the upper chamber when ECs were cocultured with vascular SMCs was secondary to both an increased production of PAI-1 protein by ECs and increased transport of PAI-1 from the basal surface (that produced by other cells) to the apical surface. Thus, enhanced production of PAI-1 by diverse cell types (eg, hepatocytes and adipocytes) as well as by vascular SMCs is likely to be associated with increased transport of PAI-1 by ECs to the lumen of vessels in the setting of hyperinsulinemia.
The results of this study support the hypothesis that the depressed fibrinolytic capacity found in blood in subjects with insulin-resistant states is secondary to direct effects of insulin on elaboration of PAI-1 from cellular elements in vessel walls. The interaction between ECs and SMCs is critical to the effect of insulin on elaboration of PAI-1 by vascular wall constituents. The results in cocultured cells are dependent on (1) a limited duration of culture of the mural cellular constituents (ie, passages 2 through 6 for ECs) and (2) the confluence of both cell types. Differences in the extent to which these conditions defined above have been present in previous work appear likely to account for the disparity in results reported.14 19 20 21
In addition to the variables defined in our experiments, the 4G/5G PAI-1 polymorphism may influence responses of ECs. This polymorphism has been demonstrated in the promoter region of the PAI-1 gene.36 37 38 Subjects with the 4G polymorphism have higher concentrations of PAI-1 in blood.36 37 38 This polymorphism may account for some of the heterogeneity of the arterial expression of PAI-1 observed in organ culture. Further, the polymorphism may potentiate the response to triglycerides and insulin in vivo and in vitro as observed in our results with organ culture.37 38
The prevalence of complications of atherosclerosis1 2 3 4 and the relative inefficiency of mechanical revascularization procedures in patients with type II diabetes39 40 41 underscore the importance of elucidation of mechanisms underlying the development and progression of vascular disease in subjects with insulin-resistant states. Our data indicate that intercellular communication is critical in mediating augmented expression of PAI-1 by human and porcine arteries. Identification and characterization of the specific mediator(s) of the intercellular communication through SMCs modulate EC expression of PAI-1 under basal conditions and, in response to insulin, should provide novel and potentially sensitive targets for therapy aimed at normalizing expression in vivo of fibrinolytic system proteins in the presence of insulin resistance. Anticipated benefits from restoration of fibrinolytic capacity include reduction of the incidence and severity of thrombotic vascular occlusive disease, improved efficacy of revascularization procedures, and possibly a decreased rate of progression of atherosclerosis.
Selected Abbreviations and Acronyms
|CM||=||previously conditioned medium|
|ECGS||=||endothelial cell growth supplement|
|ICM||=||CM in presence of insulin|
|PAI-1||=||plasminogen activator inhibitor type 1|
|SMC||=||smooth muscle cell|
The authors wish to thank Patricia Quinn Baumann for expert technical assistance and Burton E. Sobel, MD, for review of the manuscript.
- Received April 23, 1997.
- Revision received May 22, 1997.
- Accepted June 6, 1997.
- Copyright © 1997 by American Heart Association
Uusitupa MI, Niskanen LK, Siitonen O, Voutilainen E, Pyorala K. 5-Year incidence of atherosclerotic vascular disease in relation to general risk factors, insulin level, and abnormalities in lipoprotein composition in non–insulin-dependent diabetic and nondiabetic subjects. Circulation. 1990;82:27-36.
Pan WH, Cedres LB, Liu K, Dyer A, Schoenberger JA, Shekelle RB, Stamler R, Stamler J. Relationship of clinical diabetes and asymptomatic hyperglycemia to risk of coronary-heart disease mortality in men and women. Am J Epidemiol. 1986;123:504-516.
Modan M, Or J, Karasik A, Drory Y, Fuchs Z, Lusky A, Chetrit A, Halkin H. Hyperinsulinemia, sex, and risk of atherosclerotic cardiovascular disease. Circulation. 1991;84:1165-1175.
McGill JB, Schneider DJ, Arfken CL, Lucore CL, Sobel BE. Factors responsible for impaired fibrinolysis in obese subjects and NIDDM patients. Diabetes. 1994;43:104-109.
Auwerx J, Bouillon R, Collen D, Geboers J. Tissue-type plasminogen activator antigen and plasminogen activator inhibitor in diabetes mellitus. Arteriosclerosis. 1988;8:68-72.
Schneider DJ, Sobel BE. Augmentation of synthesis of plasminogen activator inhibitor type-1 by insulin and insulin-like growth factor type-1: implications for vascular disease in hyperinsulinemic states. Proc Natl Acad Sci U S A. 1991;88:9959-9963.
Fattal PG, Schneider DJ, Sobel BE, Billadello JJ. Post-transcriptional regulation of expression of plasminogen activator inhibitor type 1 mRNA by insulin and insulin-like growth factor 1. J Biol Chem. 1992;267:12412-12415.
Nordt TK, Schneider DJ, Sobel BE. Augmentation of synthesis of plasminogen activator inhibitor type-1 (PAI-1) by precursors of insulin: a potential risk factor for vascular disease. Circulation. 1994;89:321-330.
Mayer EJ, Fujita T, Gardell SJ, Shebuski RJ, Reilly CF. The pharmacokinetics of plasminogen activator inhibitor-1 in the rabbit. Blood. 1990;76:1514-1520.
Alessi MC, Anfosso F, Peiretti HF, Nalbone G, Juhan-Vague I. Up-regulation of PAI-1 synthesis by insulin and proinsulin in HepG2 cells but not in endothelial cells. Fibrinolysis. 1995;9:237-242.
Schneider DJ, Nordt TK, Sobel BE. Stimulation by proinsulin of expression of plasminogen activator inhibitor type-1 in endothelial cells. Diabetes. 1992;41:890-895.
Van Hinsbergh VWM, Binnema D, Sheffer MA, Sprengers ED, Kooistra T, Rijken DC. Production of plasminogen activators and inhibitors by serially propagated endothelial cells from adult human blood vessels. Arteriosclerosis. 1985;7:389-400.
Sawdey M, Podor TJ, Loskutoff DJ. Regulation of type 1 plasminogen activator inhibitor gene expression in cultured bovine aortic endothelial cells. J Biol Chem. 1989;264:10396-10401.
DeClerck PJ, Alessi MC, Verstreken M, Kruithof EKO, Juhan-Vague I, Collen D. Measurement of plasminogen activator inhibitor 1 in biologic fluids with a murine monoclonal antibody-based enzyme-linked immunosorbent assay. Blood. 1988;71:220-225.
Laemmli UK. Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature. 1970;39:680-682.
White BA, Bancroft,, FC. Cytoplasmic dot hybridization. J Biol Chem. 1982;257:8569-8572.
Defronzo RA. The triumvirate: β-cell, muscle, liver: a collusion responsible for NIDDM. Diabetes. 1988;37:667-687.
Defronzo RA. Insulin secretion, insulin resistance, and obesity. Int J Obesity. 1982;6(S1):73-82.
Modan MH, Halkin S, Almog S, Lusky A, Eshkol A, Shefi M, Shitrit A, Fuchs Z. Hyperinsulinemia: a link between hypertension obesity and glucose intolerance. J Clin Invest. 1980;66:763-772.
Lundgren CH, Brown SL, Nordt TK, Sobel BE, Fujii S. Elaboration of type-1 plasminogen activator inhibitor from adipocytes: a potential pathogenetic link between obesity and cardiovascular disease. Circulation. 1996;93:106-110.
Dawson S, Hamsten A, Wiman B, Henney A, Humphries S. Genetic variation at the plasminogen activator inhibitor-1 locus is associated with altered levels of plasminogen activator inhibitor-1 activity. Arterioscler Thromb. 1991;11:183-190.
Panahloo A, Mohamed-Ali V, Lane A, Green F, Humphries SE, Yudkin JS. Determinants of plasminogen activator inhibitor 1 activity in treated NIDDM and its relation to a polymorphism in the plasminogen activator inhibitor 1 gene. Diabetes. 1995;44:37-42.
Stein B, Weintraub WS, Gebhart SSP, Cohen-Bernstein CL, Grosswald R, Liberman HA, Douglas JS, Morris DC, King SP. Influence of diabetes mellitus on early and late outcome after percutaneous transluminal coronary angioplasty. Circulation. 1995;91:979-989.
Sobel BE. Potentiation of the vasculopathy by insulin: implications from an NHLBI clinical alert. Circulation. 1996;93:1613-1615.