| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Department of Internal Medicine III, University of Freiburg Medical School (T.K.N., K.P., C.B.), 79106 Freiburg, Germany; and Department of Medicine, University of Vermont College of Medicine (B.E.S.), Burlington, Vermont 05446
Address all correspondence and requests for reprints to: Thomas K. Nordt, M.D., Medizinische Universitätsklinik, Hugstetter Strasse 55, 79106 Freiburg, Germany. E-mail: nordt{at}mm31.ukl.uni-freiburg.de
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Human hepatic and vascular cells, the main sources of circulating PAI-1, were studied in cell culture. In human hepatic cells, PAI-1 accumulated in conditioned medium by 23% within 24 h after exposure to 3 µg/mL troglitazone (P = 0.001). The accumulation depended on the concentration of troglitazone, but not that of insulin (known to stimulate PAI-1 synthesis). By contrast, in human aortic smooth muscle cells, 3 µg/mL troglitazone decreased basal PAI-1 expression by 23% (P = 0.037) and decreased transforming growth factor-ß-induced expression by 34% (P = 0.026). Concomitant insulin had no effect. Tissue-type plasminogen activator was decreased by 38% (P = 0.002). In human endothelial cells, PAI-1 was diminished by 32% (P < 0.001), whereas tissue-type plasminogen activator was unaffected.
The results suggest that the reduction in plasma activity of PAI-1 induced by troglitazone in patients may reflect both directly mediated diminution of its elaboration from vessel walls and indirectly mediated reduction of its hepatic synthesis secondary to attenuation of hyperinsulinemia (known to increase the hepatic synthesis of PAI-1).
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Troglitazone is a novel oral insulin sensitizer that increases glucose disposal in adipose tissue and skeletal muscle (11, 12). Besides glucose metabolism, troglitazone also affects other (patho)physiological mechanisms, such as low density lipoprotein receptor activity (13) and platelet aggregation (14). In women with polycystic ovary syndrome, it diminishes not only the hyperinsulinemia, but also the markedly increased concentration and activity of PAI-1 in blood (2). However, the mechanisms responsible have not yet been fully elucidated. If its effects on PAI-1 elaboration or degradation are direct, it may serve as a prototype for the development of drugs designed to normalize increased PAI-1 in other clinical conditions. If its effects are indirect and secondary to a reduction of insulin-mediated stimulation of hepatic PAI-1 synthesis, the importance of a reduction of hyperinsulinemia as a therapeutic objective in the treatment of patients with type 2 diabetes and other insulin-resistant states would be accentuated.
The present study was performed to determine whether one or both of those mechanisms are likely to be operative. Accordingly, Hep G2 cells as well as vascular smooth muscle and endothelial cells in culture were exposed to troglitazone at concentrations consistent with those seen in blood in patients treated with troglitazone to enhance insulin sensitivity (15), and PAI-1 expression was characterized by assay of conditioned medium and cell lysates.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Human recombinant insulin acquired from Sigma (Deisenhofen, Germany) was dissolved in 0.9% sodium chloride with 0.5% bovine albumin (fraction V, low endotoxin, cell culture tested; Sigma, St. Louis, MO). Human recombinant transforming growth factor-ß1 (TGFß1) was acquired from Roche Molecular Biochemicals (Mannheim, Germany). Immediately before use troglitazone (acquired from Parke-Davis, Ann Arbor, MI) was dissolved in dimethylsulfoxide (Roth, Karlsruhe, Germany) and diluted in cell culture medium. The final concentration of dimethylsulfoxide to which cells were exposed was 0.1% (vol/vol), which affected neither cell morphology nor basal or insulin-stimulated expression of PAI-1. Control experiments with vehicle alone were performed for each set of conditions studied.
Hep G2 cell cultures
Hep G2 cells, highly differentiated human hepatoma cells, were acquired from American Type Culture Collection (HB 8065; Manassas, VA). They were selected because they simulate normal human hepatocytes with respect to expression of PAI-1 in response to agonists and antagonists. They were grown to confluence, serum starved as described previously (16), and subsequently exposed to fresh medium constituted with selected concentrations of agonists.
Human aortic smooth muscle cell (HASMC) cultures
HASMCs were prepared from segments of human ascending aorta
obtained fresh from donors of hearts used for cardiac transplantation
in the Department of Cardiovascular Surgery at the University of
Heidelberg (Heidelberg, Germany). After removal of both the endothelium
and subendothelial layer and the adventitial layer, the remaining
tunica media tissue was cut into approximately
1-mm3 pieces and cultured as explants in
conventional cell culture dishes without specific coating. Within 4
weeks HASMCs had migrated out of the explants. Cells were verified to
be smooth muscle cells as judged from typical hill and valley
morphology and their homogeneous staining with mouse monoclonal
anti-
-actin antibody (Roche Molecular Biochemicals) and
goat fluorescein-conjugated antimouse IgG antibody (Sigma,
St. Louis, MO). For all experiments, cells were used in passages
2–7 only.
HASMCs were grown to confluence in RPMI 1640 medium containing 25 mmol/L HEPES and L-glutamine (BioWhittaker, Inc., Verviers, Belgium), supplemented with 10% FBS (Serva, Heidelberg, Germany) and 1% antibiotic antimycotic solution (Sigma; 100 U/mL penicillin, 100 µg/mL streptomycin, and 250 ng/mL amphotericin, final concentrations). Monolayers of confluent cells were serum starved in RPMI 1640 without supplements for 16–24 h. After serum starvation the cells were exposed to fresh medium constituted with selected concentrations of agonists. For all experiments HASMCs from at least three different donors were used and characterized separately.
Human umbilical vein endothelial cell (HUVEC) cultures
HUVEC were prepared from umbilical cords conventionally with the use of collagenase A from Clostridium histolyticum (Roche Molecular Biochemicals). They were grown to confluence in medium 199 with Earle’s salts and L-glutamine (Life Technologies, Inc., Eggenstein, Germany) supplemented with 10% FBS, 50 mg/L endothelial cell growth supplement (Harbor BioProducts, Norwood, MA), and the same 1% antibiotic antimycotic solution as that used for HASMCs. Experiments were performed with medium 199 supplemented with 10% FBS only. The endothelial cells exhibited typical cobblestone morphology and were used in primary culture (passage 1 only).
Preparation of cell lysates
Cells in culture were washed twice with phosphate-buffered saline (BioWhittaker, Inc.), incubated in 0.5% Triton X-100, and transferred to siliconized tubes. The cell suspensions were sonified on ice for 1 min to optimize cell lysis. After centrifugation at 12,000 x g at 4 C for 10 min, the supernatant fractions were supplemented with Tween-80 (final concentration, 0.01%) and stored at -20 C until assay.
Assay of PAI-1, tissue-type plasminogen activator (t-PA), urokinase-type plasminogen activator (u-PA), and total protein
After selected time intervals conditioned media were supplemented with Tween-80 (final concentration, 0.01%) and stored at -20 C until assay. The concentrations of PAI-1 protein in conditioned media and cell lysates were measured by enzyme-linked immunoabsorption assay (ELISA) as described previously (8). Active, latent, and t-PA-complexed forms of PAI-1 are detected with equal sensitivity with the ELISA used.
The concentration of t-PA protein was measured by ELISA with the use of an immobilized goat anti-t-PA and horseradish peroxidase-labeled anti-t-PA Fab fragments (Imubind total t-PA Stripwell ELISA material from American Diagnostica, Greenwich, CT).
The concentration of u-PA protein was measured by ELISA with the use of a sandwich technique (immobilized mouse monoclonal antiurokinase antibody, urokinase, and horseradish peroxidase-labeled goat antiurokinase antibody) and an immunological specificity and accuracy control (TintElize u-PA, Biopool, Umea, Sweden). The limit of sensitivity for detection is approximately 0.1 ng/mL.
Total protein was assayed with the use of a colorimetric microassay procedure after solubilization of the protein with detergent (DC Protein Assay, Bio-Rad Laboratories, Inc., Hercules, CA).
Statistical analysis
Data are the mean ± SEM. The significance of differences between groups was assayed by two-way ANOVA, the statistical significance of the results obtained was tested post-hoc according to Bonferroni. P < 0.05 was considered significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Hep G2 cells were exposed to medium containing 0, 0.3, 1, or 3 µg/mL troglitazone for 24 h in the presence or absence of 10 nmol/L insulin. These concentrations were selected based on the concentrations in blood in vivo after the oral administration of troglitazone in therapeutic doses of 5 mg/kg (15). Concentrations of lactate dehydrogenase, aspartate aminotransferase, and alanine aminotransferase in conditioned medium of the Hep G2 cells we studied were identical in the presence and absence of 3 µg/mL troglitazone over 24 h (data not shown), indicative of a lack of toxic effects on the cells.
In the absence of insulin, troglitazone at all
concentrations significantly increased the accumulation of PAI-1 in
conditioned medium by up to 23% (at 3 µg/mL
troglitazone, P = 0.001; Fig. 1
). In the presence of 10 nmol/L insulin,
the corresponding increases were statistically significant as well,
with 19% as a maximum (at 3 µg/mL troglitazone,
P = 0.004). In the absence of
troglitazone, 10 nmol/L insulin induced a 1.7-fold
increase in PAI-1 secretion, confirming results obtained previously
(8). Of interest, the ratio of PAI-1 accumulation in the presence and
absence of insulin remained constant (1.6- to 1.7-fold) regardless of
the concentration of troglitazone present. Similar results
were obtained with cells exposed to troglitazone and 100
nmol/L insulin as well as to 0.1 and 1 nmol/L insulin (three
experiments, each performed in triplicate; data not shown). Thus,
troglitazone did not increase insulin sensitivity with
respect to PAI-1 expression. No detectable changes in cell number or
morphology were seen.
|
).
|
demonstrates the dependence of
the concentration of PAI-1 in Hep G2 cell lysates on the concentrations
of troglitazone and insulin to which the cells had been
exposed for 24 h. In the absence of insulin,
troglitazone (3 µg/mL) increased the concentration of
PAI-1 in cell lysates by 68% (P = 0.015). In the
presence of 10 nmol/L insulin, the increase was 44% (P
= 0.088). With increasing concentrations of troglitazone,
the ratio of PAI-1 concentrations in the cell lysates with and without
exposure of the cells to concomitant insulin tended to decrease (from
1.7- to 1.4-fold).
|
Troglitazone at any of the concentrations studied did not affect the accumulation of u-PA over 24 h (data not shown). t-PA was not detected in any of the conditioned media.
Effects of troglitazone on accumulation of PAI-1 in conditioned medium of HASMCs
To define the potential effects of troglitazone on the release into conditioned medium of PAI-1 protein by vascular smooth muscle cells, HASMCs were exposed to troglitazone at concentrations of 0, 0.01, 0.03, 0.1, 0.3, 1, 3, and 10 µg/mL for 24 h. Troglitazone elicited a concentration-dependent decrease in accumulation of PAI-1 protein in conditioned medium [significant with 1 µg/mL (P = 0.037) and 3 µg/mL (P = 0.037) troglitazone]. Not surprisingly, at a suprapharmacological concentration of 10 µg/mL, changes in cell morphology and a decrease in total protein in cell lysates were seen, consistent with the cytopathic effects of these suprapharmacological concentrations (data not shown).
TGFß was used as an agonist in experiments with HASMCs, because it is
known to increase PAI-1 synthesis (8) and to play an important role in
atherogenesis. When HASMCs were incubated with selected concentrations
of troglitazone (0, 0.3, 1, and 3 µg/mL) in the presence
and absence of 1 ng/mL TGFß for 24 h, concentration-dependent
diminution of accumulation of PAI-1 in conditioned medium was evident
(Fig. 3). Thus, just as in the absence of
exogenous TGFß, a reduction of PAI-1 accumulation of more than 23%
was seen with 3 µg/mL troglitazone (P =
0.037). Analogous results were obtained with 1 ng/mL TGFß as well
(34% reduction with troglitazone; P =
0.026). The ratio between PAI-1 secretion in the presence and that in
the absence of TGFß was 2.0- to 2.4-fold and was not influenced by
the concentration of troglitazone. The ratio of PAI-1
secretion after exposure of the cells to TGFß (0, 0.1, 1, or 10
ng/mL) in the presence and absence of 3 µg/mL
troglitazone was the same (averaging 0.76–0.72)
regardless of the concentration of TGFß. Furthermore, the diminution
of accumulation of PAI-1 by troglitazone was comparable in
the presence of 10 nmol/L insulin, as judged from the results of three
experiments, each performed in triplicate (data not shown). Insulin
alone tended to induce PAI-1 expression in HASMCs; this, however, did
not reach statistical significance.
|
summarizes the results of
time-course experiments when the HASMCs were exposed to TGFß and
troglitazone for up to 72 h. The stimulating effect
of TGFß remained constant during this period. In contrast, the
inhibiting effect of troglitazone seemed to vanish over
time.
|
). However, in cells exposed to 1 ng/mL
TGFß, PAI-1 in lysates was decreased by 53% when incubations had
included 3 µg/mL troglitazone (P = 0.064). As
was the case with Hep G2 cells the magnitude of change in PAI-1
concentrations induced by troglitazone was greater in cell
lysates compared with that in conditioned medium.
|
HASMC accumulation of t-PA, but not soluble u-PA, occurred in
conditioned medium. As shown in Table 3,
a concentration-dependent reduction in accumulation of t-PA was seen
with troglitazone. The reduction was evident in cells
exposed to 10 ng/mL TGFß as well and was not altered by insulin, as
judged from the results of three experiments, each performed in
triplicate (data not shown). Insulin (10 nmol/L) exerted no apparent
effect on accumulation of t-PA regardless of whether
troglitazone was included in the incubation medium.
|
Incubation of HUVEC with graded concentrations of troglitazone (0, 0.01, 0.03, 0.1, 0.3, 1, 3, and 10 µg/mL) for 24 h led to a reduction in the accumulation of PAI-1 in conditioned medium of 7% with 1 µg/mL troglitazone (P = 0.023) and 32% with 3 µg/mL troglitazone (P = 0.001) with no associated change in the concentration of total protein in the cell lysates. As was the case with HASMCs, 10 nmol/L insulin did not alter the results, as judged from the results in three experiments, each performed in triplicate (data not shown). Insulin alone did not consistently influence PAI-1 expression in HUVECs. With 10 µg/mL troglitazone changes in cell morphology associated with a marked decrease in total protein in cell lysates was evident, consistent with the cytopathic effects of the suprapharmacological concentrations.
Effects of troglitazone on accumulation of t-PA in HUVEC-conditioned medium
t-PA, but not u-PA, was detected in conditioned medium of HUVEC. Its concentration was not significantly affected by exposure of the cells to troglitazone in any of the concentrations studied or to 10 nmol/L insulin, as judged from the results of three experiments, each performed in triplicate (data not shown).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In this study cells were exposed to troglitazone at concentrations consistent with those seen in blood in patients treated with troglitazone (15). In addition, Hep G2 cells were incubated with insulin at concentrations consistent with those to which liver cells in vivo are exposed (17). Thus, with extrapolation of the data from the in vitro to the in vivo situation, it may be expected that the reduction of hepatic PAI-1 synthesis secondary to the decreased concentration of insulin after treatment with troglitazone is sufficient to overcome the direct stimulatory effect of troglitazone on hepatic PAI-1 synthesis. Although Hep G2 cells (an immortal cell line) may not simulate all functional aspects of hepatocytes in vivo, in general, responses to agonists and antagonists of synthesis and elaboration of PAI-1 in vivo have been presaged by results with Hep G2 cells in vitro.
The different effects of troglitazone on PAI-1 synthesis in liver cells, on the one hand, and in vascular smooth muscle as well as vascular endothelial cells, on the other hand, are of specific interest. The discrepant effects of insulin on PAI-1 synthesis in Hep G2 cells (significant induction) and in HASMCs as well as HUVECs (no statistically significant induction and no effect at all, respectively) demonstrate major differences in the regulation of PAI-1 synthesis in various cell types. These may be explained by differences both in the occurrence of receptors on the cell surface and in the pathways of intracellular signal transduction and may be modulated by polymorphisms in the PAI-1 gene and its promoter. Thus, further investigations, including determining the impact of other thiazolidinediones on PAI-1 synthesis, are warranted.
The concept that the effect of troglitazone on PAI-1 expression in blood in vivo is primarily indirect is supported further by the close correlation between decreased concentrations of insulin in blood and corresponding differences in PAI-1 in hyperinsulinemic patients with the polycystic ovary syndrome who have been treated with troglitazone (2). In addition, metformin, an oral agent that decreases endogenous hepatic glucose output and is therefore insulin sparing, decreases concentrations of PAI-1 in such patients, most likely because it decreases concentrations of insulin in blood (18). Troglitazone did not alter Hep G2 cell accumulation of u-PA, the plasminogen activator that predominates in tissue, nor did it alter the concentration of total protein in cell lysates. Thus, the induction of PAI-1 expression appeared to be at least relatively specific.
Troglitazone did not alter the accumulation of t-PA with or without insulin in conditioned medium of HUVEC. By contrast, it reduced accumulation of t-PA in conditioned medium of HASMCs, a finding consistent with a reduction in concentrations of t-PA in blood in patients with the polycystic ovary syndrome who are treated with troglitazone (2). In view of the well recognized importance of elaboration of t-PA by endothelial cells, these results are consistent with a predominant functional effect of reduced PAI-1 release in vessel walls. As both smooth muscle cells and endothelial cells exhibited diminished secretion of PAI-1, but only the smooth muscle cells exhibited diminished t-PA secretion in response to troglitazone, the overall effect of the agent is likely to be increase of vessel wall-dependent fibrinolytic system activity. The troglitazone-induced reduction in t-PA occurred even in the absence of insulin. By contrast, the antihyperglycemic effect of troglitazone is evident only when insulin is present in experimental animals given the insulin sensitizer (19). Troglitazone appears to exert several effects on vascular smooth muscle cells, including reduction of high glucose-induced migration and proliferation (20).
The observations in the present study suggest that thiazolidinediones such as troglitazone will be useful in inducing a favorable balance between fibrinolysis and thrombosis in patients with insulin-resistant states. In addition, they imply that these agents may attenuate derangements in the balance between proteolysis and its inhibition within vessel walls that have been implicated in accumulation of extracellular matrix and augmentation of chemotaxis and cell migration early in atherogenesis (21). By contrast, reductions in PAI-1 in blood in vivo and normalization by troglitazone of fibrinolysis in blood in patients with insulin-resistant states are not likely to depend exclusively on direct effects of troglitazone on PAI-1 expression. Rather, they appear likely to be a consequence as well of a reduction in insulin resistance, consequent diminution of prevailing concentrations of insulin in blood, and secondary diminution of augmented PAI-1 expression by the liver associated with a reduction of hyperinsulinemia in patients with states of insulin resistance.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received August 20, 1999.
Revised December 8, 1999.
Accepted December 21, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. A. Parulkar, M. L. Pendergrass, R. Granda-Ayala, T. R. Lee, and V. A. Fonseca Nonhypoglycemic Effects of Thiazolidinediones Ann Intern Med, January 2, 2001; 134(1): 61 - 71. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
