THROMBOSIS RESEARCH 58; l-12,1990 0049-3848190 $3.00 + .OOPrinted in the USA. Copyright (c) 1990 Pergamon Press pk. All rights reserved.
INSULIN,
CORTISOL AND CATECHOLAMINES DO NOT REGULATE VARIATIONS IN FIBRINOLYTIC ACTIVITY
CIRCADIAN
Wayne L. Chandlerl, Dan Mornin' Richard 0. Whittenl Peter Angletonl, Federico M. Farid Thomas R. Fritsche?, Richard C. Veith3, John RI Strattona. Departments of Laboratory Medicine', Medicinea, and the Geriatric Research, Education and Clinical Center, Department of Psychiatry and Behavioral Sciences3, University of Washington, Seattle, WA, 98195, and the Seattle Veterans Administration Medical Center213, USA. (Received 18.8.1989; accepted in revised form 11.1.1990 by Editor N.U. Bang)
ABSTRACT To evaluate possible hormonal regulators of the diurnal rhythm in fibrinolytic activity, we measured tissue plasminogen activator (t-PA) activity, plasminogen activator inhibitor activity (PAI-l), t-PA antigen, insulin, cortisol, and catecholamines in 6 healthy males (age 34+5) every 2 hours for 24 hours. Fibrinolysis was characterized by a peak in PAI- activity and a trough in t-PA activity at 0600 h. PAI- activity increased 92% and t-PA activity decreased 56% between 2400 h and 0600 h. t-PA antigen (principally a measure of t-PA/PAI-1 complex), peaked at 0800 h. In comparison, insulin levels were lowest at night when PAI- activity was rising. The weak inverse correlation between insulin and PAI- activity (r = -0.28, p < 0.02), was not sufficient to explain the diurnal change in fibrinolysis. While cortisol demonstrated the expected circadian change, the increase in cortisol did not occur until 0400 h, 4-6 hours after the rise in PAIand decrease in t-PA activity started. Supine resting plasma epinephrine and norepinephrine showed no circadian rhythm. From this data, we hypothesize that the increased level of PAIin the morning consumes t-PA, leading to decreased t-PA activity and increased t-PA/PAI-1 complex. Further, we conclude that insulin, cortisol, and catecholamines are not responsible for the circadian rhythm of fibrinolysis.
KEY WORDS: fibrinolysis, catecholamine
circadian
1
rhythm,
insulin,
cortisol,
2
INSULIN, CORTISOL & CATECHOLAMINES...
Vol. 58, No. 1
INTRODUCTION The fibrinolytic system plays an important role in regulating the formation and removal of vascular thrombi. In a study in which samples were obtained at 0800 h and 2000 h, we demonstrated that fibrinolytic activity follows a diurnal rhythm with low activity in the morning and higher activity in the evening in both healthy individuals and patients with unstable angina and prior myocardial infarction (1). This circadian variation in fibrinolytic activity may be clinically significant in that arterial thrombotic disease also follows a circadian rhythm (2). To understand the mechanism underlying the circadian rhythm we need a detailed description of circadian of fibrinolysis, variations in the proteins that control fibrinolysis, tissue plasminogen activator (t-PA) and plasminogen activator inhibitor type 1 (PAI-1). Further, we need to understand the regulatory model would be factors that drive this rhythm. One possible circadian hormonal regulation, similar to a number of other better known diurnal rhythms such as the adrenocorticotropic hormone (ACTH) - cortisol system. To determine possible regulators of the diurnal variation in fibrinolytic activity, we measured t-PA activity, PAI- activity, insulin, t-PA antigen, and four possible hormonal regulators, cortisol and the catecholamines epinephrine and norepinephrine in 6 healthy males every 2 hours for 24 hours. These hormones were selected as they had all previously been reported to show a circadian rhythm and were reported to be regulators of fibrinolysis (3). Elevated insulin levels have been associated with increases in plasma PAI- activity (4). Cortisol has been shown to stimulate PAIsecretion and to inhibit t-PA secretion (5,6), while epinephrine infusions have been reported to increase fibrinolytic activity (7). The results from our study indicate, that while these four hormones may be important in regulating fibrinolysis in other situations, they are not responsible for the circadian rhythm of the fibrinolytic system. MATERIALS AND METHODS Human subiects Studies on human subjects principles of the Declaration obtained from all participants University of Washington Human
were carried out according to the of Helsinki. Informed consent was and the study was approved by the Subjects Review Committee.
Materials Cyanogen bromide (CNBr), bovine albumin (98% pure, salt free) and Triton-X-100 were obtained from Sigma Chemical Co. (St. Louis, MO). Chromogenic substrate D-valyl-phenylalanyllysyl-pnitro-anilide (S-2390) was obtained from Helena Diagnostics (Beaumont, TX). All other materials not described below were reagent or analytical grade. Human glu-plasminogen (product no. 400, lot no. GO187 and G0287) was obtained from American Diagnostica Inc. (Greenwich, CO). The specific activity of the plasminogen preparation was approximately 4.7 U/mg where 1 U = the plasminogen activity in 1 mL of pooled normal plasma (8). Human fi-
Vol. 58, No. 1
INSULIN, CORTISOL & CATECHOIAMINES...
3
brinogen (96% clottable, product no. F4883, lot no. 67F-9461) was obtained from Sigma Chemical Co. (St. Louis, MO). Human fibrinogen was cleaved with CNBr as described by Verheijen et al. (9). Gnechain melanoma derived t-PA (product no. 115, lot no. 7375), was obtained from American Diagnostica Inc. The specific activity of the t-PA preparation was 500,000 IU/mg as compared to the second international t-PA activity standard (86/670) from the National Institute for Biologic Standards and Control, London, UK (lo). Polyclonal goat-anti-human melanoma t-PA and its peroxidase conjugate were obtained from American Diagnostica Inc. Polyclonal anti-cortisol and 1251-cortisol tracer were obtained from Baxter Travenol Diagnostics (Cambridge, MA). Polyclonal anti-human insulin and 1251-insulin tracer were obtained from Radioassay Systems Laboratories, Inc. (Carson, CA). Blood samnlins and samnle orenaration Each subject had an intravenous catheter placed in a forearm vein prior to the start of the study. The catheters were closed with a heparin lock filled with 1 mL of 100 U/mL heparin between sampling periods. Samples were taken every two hours for 24 hours. The first samples were taken at 1000 h. The subjects followed their normal daily activity with the exception that prior to each sample the subjects rested in the supine position for 15 min. After the samples were taken at midnight, the subjects slept until approximately 0600 h except for brief disturbances during sample acquisition at 0200 h and 0400 h. At each sampling period, the following protocol was followed: 1) the heparin lock was flushed with 5 mL of sterile saline, 2) 3 mL of blood were drawn and discarded, 3) 2.5 mL of whole blood were drawn and immediately added to a prechilled tube containing 50 UL of 0.15 mol/L NaCl, 90 gm/L EGTA, and 60 gm/L reduced glutathione (catecholamine sample), 4) 3 mL of blood were drawn and allowed to clot (cortisol and insulin sample), 5) 9 mL of blood were drawn into 1 mL of 130 mmol/L sodium citrate (PAIactivity and total t-PA antigen samples), 6) 1 mL of the citrate anticoagulated whole blood was added to 0.5 mL of 0.5 mol/L sodium acetate buffer, pH 4.2 (t-PA activity sample), and 7) the heparin lock was refilled with 1 mL of 100 U/mL heparin. The catecholamine samples were centrifuged at 4'C for 15 min at 9OOg, the plasma removed and recentrifuged as before, and the supernatant removed and stored in tightly sealed plastic tubes at -8OOC. After clotting for 15 min at room temperature, the unanticoagulated blood samples were centrifuged at 4OC for 10 min at 25OOg, the serum removed and stored at -8OOC. The citrate anticoagulated and acidified blood samples were centrifuged at room temperature for 5 min at 25OOg, and the plasma removed and stored at -8OOC. Activated partial thromboplastin times run on titrated plasma samples indicated they contained less than 0.1 U/mL heparin activity. Measurement of t-PA activitv and PAI- activity t-PA activity and PAI- activity were measured as PreViOUSly described using amidolytic assays (11,12). The final results for both the t-PA activity and PAIactivity assays were corrected for dilution by the anticoagulant and acidification buffer. The
4
INSULIN, CORTISOL & CATECHOLAMINES...
Vol. 58, No. 1
imprecision for replicate analysis of a single sample of acidified plasma within a single run of the t-PA activity assay was 0.96+0.03 [SD] IU/mL (CV = 3%, n = 16). The imprecision of replicate analysis of a single sample run on different days for the t-PA activity assay was 0.92+0.06 [SD] IU/mL (CV = 7%, n = 5). In the PAI-I activity assay, one arbitrary unit (1 AU) of PAIactivity is defined as the amount of PAI- that inhibits 1 IU of assay is one-chain t-PA activity. In our laboratory the PAIdesigned to measure up to four plasma samples per run. The imprecision for replicate analysis of a single sample of plasma within a single run of the PAI- activity assay was 19.5kO.3 [SD] AU/mL of replicate analysis of a (CV = 2%, n = 4). The imprecision single sample run on different days for the PAI- activity assay was 10.220.7 [SD] AU/mL (CV = 7%, n = 20). Total
t-PA antisen assay Total t-PA antigen was determined as described previously using an enzyme-linked immunosorbent assay (13). The final results were corrected for dilution of the plasma by citrate antifor replicate analysis of a single coagulant. The imprecision sample of titrated plasma within a single run of the t-PA antigen assay was 7.9LO.2 [SD] ng/mL (CV = 3%, n = 8). The imprecision of replicate analysis of a single sample run on different days for the t-PA antigen assay was 5.8kO.5 [SD] ng/mL (CV = 9%, n = 9). Other
Assavs Cortisol and immunoreactive insulin were measured using radioimmunoassays (14,15). Epinephrine and norepinephrine were measured using a single isotope radioenzymatic assay (16). Statistics The mean and standard error of the mean (SEM) were determined for each sampling time for each assay type. For a given sampling time, results for t-PA antigen, insulin, cortisol, epinephrine and norepinephrine were normally distributed among subjects. Results for t-PA activity and PAIactivity demonstrated wide variations in values among different subjects (skewed distributions). Simple means for these two assays primarily reflected the results of the subjects with the highest values, not the average circadian rhythm of the group. Therefore, for these two assays, we normalized the results by dividing a given subject's results by the peak value for that subject. These results were expressed in percent, thus the peak value for each subject was 100% with all other values being expressed as a percent of the peak. These percent values were then averaged for each time of day for the six subjects. For each assay type, we first performed a one-way analysis of variance (ANOVA) vs time of day. Those results showing significant differences among different times of day (P c 0.01) were further analyzed using the Tukey Honestly-Significant-Difference IIIUltiCOIIIpariSOn test at P = 0.05 and P = 0.01 levels of significance (17). Original data and copies of the statistical evaluation tables are available from the authors.
Vol. 58, No. 1
INSULIN, CORTISOL & CATECHOLAMINES...
5
RESULTS t-PA Activitv All six subjects demonstrated a trough in t-PA activity in the morning and a peak in the afternoon or evening. There was a wide range in t-PA activity levels among different subjects. Peak activities ranged from 1.0 to 2.6 IU/mL, and trough activities from 0.4 to 1.1 IU/mL. The time of trough t-PA activity ranged from 0600 h to 1200 h, but occurred most commonly (4 of 6 subjects) at 0600 h. The time of peak t-PA activity ranged from 1200 h to 2200 h. While each of the individual subjects showed circadian variations in t-PA activity, there were no significant differences in average absolute t-PA activity (IU/mL) among the times tested (one-way ANOVA, p = 0.9). This was due to the wide range of values (skewed distribution) and variations in peak and trough times (phase shifts) among different subjects. To eliminate the effect of the wide range in values among subjects, we compared the average normalized percent t-PA activity (see methods) vs time of day (Fig. 1). Percent activity data demonstrated significant differences among different times of day (one-way ANOVA, p = 0.008), with lower t-PA activities between 0400 h and 1000 h compared to all other times of day (p = 0.01). If the trough values for each subject were phase shifted to the same time of day (0600 h), the apparent average rhythm became even stronger (one-way ANOVA, p CO.001). It should be noted that the data in Figure 1 was not phase shifted, but represents the average for the six subjects for each time of day. The mean percent decrease in t-PA activity from peak to nadir was 56+5%. PAI-
Activitv PAI- activity followed an inverse pattern compared to t-PA activity with a peak in the morning and a trough in the evening. Again the data were skewed (peak range 7.7 to 12.4 AU/mL, trough range 0.0 to 7.6 AU/mL) and the peak and trough times inconsistent among subjects (peak time range 0200 h to 0600 h, trough time range 1400 h to 2200 h). There were significant differences among average absolute PAI- activities (AU/mL) vs time of day (one-way ANOVA, p = 0.03), but as before the comparison reached greater significance when normalized percent PAIactivities were comwere pared (Fiq. 1, p = 0.002), and when the peak activities phase shifted-to the same time of day (p < 0.001). The mean percent decrease in PAI- activity from peak to nadir was 63+28%. Total t-PA Antiaen Total t-PA antigen followed a pattern similar to that for PAIactivity, with a peak in the morning and a trough in the t-PA antigen results ANOVA, .p _a = 0.001). evening (Fig. . 1, one-way _. . _ showed less inter-subject variation (peak range 6.2 to 10.7 ng/mL, trough range 3.4 to 5.1 ng/mL). Peak levels for t-PA antigen occurred between 0400 h and 0800 h. Between 1400 h and 2400 h there was little variation in results, no consistent trough time was noted. The mean percentage decrease in total t-PA antigen from peak to nadir was 45+13%.
Vol. 58, No. 1
CORTISOL & CATECHOLAMINES...
INSULIN.
FIG. 1
O’;;oo
“1
2400
Average circadian variations in insulin, cortisol, plasminogen activator inhibitor activity, type 1 (PAI-1) tissue plasminogen activator (TPA) activity, and TPA antigen in 6 healthy subjects for each time of day. Error bars indicate mean+SEM.
1
2400
o&o
12bo
l&O
24bO
100
50
1%
loo-
IJ6o---%-@++@ 2400 owa
1200
TIME
E 2lo 2400 I
I
I
I I
I
I
0600
1200
moo
I
MOO
2400
(h)
FIG. 2
2400
Circadian variations in tissue plasminogen activator (TPA) activity, plasminogen activator inhibitor type 1 (PAI-1) activity,and TPA antigen in subject A (upper panel) and subject B (lower panel). 2400
0600
TIME
1200
(h)
1200
2400
INSULIN, CORTISOL & CATECHOLAMINES...
Vol. 58, No. 1
7
Individual Patterns In general, two types of individual circadian patterns were seen in the fibrinolytic factors. The first pattern consisted of an almost continuous sinusoidal variation in PAIand t-PA activity (Subject A, Fig. 2). With this pattern there was essentially no time during which the levels were stable. Subject A demonstrated an almost 5 fold change in PAI- activity during the day and a 4 fold change in t-PA activity with a inverse correlation between PAIand t-PA activity (1:= -0.83, p < 0.001). As with the averaged data, total t-PA antigen followed PAI- activity but demonstrated a peak and trough that occurred approxmately 4 hours after the equivalent times for PAI-1. The second type of circadian pattern showed an essentially stable level of PAIactivity, t-PA activity and t-PA antigen from about 1200 h to 2400 h (Subject B, Fig. 2). Subject B also showed a inverse correlation between t-PA and PAI- activity (r = activity levels began to -0.77, p = < 0.01). At 2400 h, PAIincrease and t-PA activity to decrease reaching their respective peak and trough at 0600 h, then returning to baseline by 1200 h. Baseline (i.e. trough) levels of PAI- activity in Subject B were approximately 4 times higher than those seen in Subject A (sinusoidal type of pattern). Insulin The six subjects demonstrated significant circadian variations in plasma insulin (one-way ANOVA, p = 0.006). Insulin levels were highest and most variable during the waking hours (1000 h to 2400 h), dropping rapidly during the night (0200 h to 0800 h) to a lower, more stable level (Fig. 1). This pattern also corresponded to the subjects eating patterns. Insulin levels were 60
FIG. 3 Circadian variations in epinephrine and norepinephrine in 6 healthy subjects. Error bars indicate mean+SEM.
40 20
I
2400
0600
2400
0600
TIME
1200
1600
(h)
2400
8
INSULIN, CORTISOL & CATECHOLAMINES...
Vol. 58, No. 1
falling when PAI- activity was rising during the night. All six individuals demonstrated weak inverse correlations between insulin and PAI- activity (overall, r = -0.28, p < 0.02). Cortisol All six subjects demonstrated the expected diurnal change in cortisol: low levels at night rising rapidly after 0400 h to a peak at 0600 h, then slowly falling through the day back to baseline levels (Fig. 1, one-way ANOVA, p < 0.001). In contrast to the skewed values for PAI- and t-PA activity, cortisol levels demonstrated relatively narrow ranges at all times during the day with no evidence of phase shifts among subjects. At the time of the cortisol increase (0400 h to 0600 h), PAIactivity had already been rising for 4 to 6 hours (Fig. 1). Catecholamines Both plasma epinephrine and norepinephrine were uniformly low throughout the study, consistent with resting supine levels of a circadian variation in (Fig. 3). There was no evidence epinephrine or norepinephrine (one-way ANOVA, p > 0.7). DISCUSSION Circadian Variabilitv of Fibrinolvtic Factors This study, in which samples were drawn at two hour intervals, provides the most detailed description currently available of circadian variations in t-PA activity, PAIactivity and total t-PA antigen. The mean percentage fall from peak to nadir was 56+5% for t-PA activity, 63+28% for PAI- activity and 45+13% for total t-PA antigen. Thus, all three variables showed significant diurnal variability, confirming prior studies that sampled at less frequent intervals (1,18-21). To understand what controls the circadian rhythm of fibrinolytic activity, we must determine which fibrinolytic factors are responsible for the diurnal change and what the mechanism is that drives it. Overall, the lowest t-PA activity, highest PAIactivity, and highest total t-PA antigen were seen in the morning at approximately 0600 h. Assuming the clearance of these factors does not change during the day, then the trough in t-PA activity during the morning is most likely due to either decreased secretion of t-PA or increased inhibition of t-PA by PAI-1. If t-PA secretion were decreased we would expect the total level of t-PA antigen in the blood to fall in the morning, which did not occur. In fact, total t-PA antigen was highest in the morning when t-PA activity was at its nadir. It should be noted that in non-venous occlusion plasma, total t-PA antigen assays measure almost exclusively t-PA/PAI-1 complex (13). We hypothesize that the increased level of PAI- in the morning consumes t-PA, leading to decreased t-PA activity and increased t-PA/PAI-1 complex. Evaluation of the circadian rhythm of fibrinolysis is complicated by differences in the amplitude, phase and pattern of the rhythm among subjects (22). Healthy individuals in our study showed a wide range in both t-PA and PAIactivities. Trough levels in one individual were higher than the peak in another subject. Second, the timing of peak and trough levels were not
Vol. 58, No. 1
INSULIN, CORTISOL & CATECHOLAMINES...
9
consistent among individuals, but demonstrated phase shifts of 2 to 4 hours. Third, there appeared to be at least two types of circadian patterns. Three of the subjects demonstrated a sinusoidal variation throughout the day with essentially no stable periods, while the other three showed a single peak or trough in the morning that returned to a stable baseline for the rest of the day. The amount of variability noted in the fibrinolytic rhythms was greater than that seen for other factors that show a circadian rhythm. For example, cortisol levels in our subjects fell within narrow ranges at all times during the day, were all in phase showing a dramatic rise between 0400 h and 0600 h, and demonstrated only one type of circadian pattern. Hormonal Recrulation of Circadian Chancres in Fibrinolvsis We measured four hormones previously reported to show circadian rhythms, insulin, cortisol, epinephrine and norepinephrine (3) - Insulin followed a circadian rhythm with peak levels during the waking hours and lower levels at night. We found a weak inverse correlation between PAIactivity and insulin; this is in contrast to previous studies that have associated obesity and elevated levels of insulin with increased plasma PAIactivity and increased production of PAI- in hepatoma cells (4,23-25). In obese subjects that were fasted for 24 hours, insulin levels have been shown to fall on average from 22.3k2.2 to 16.3kl.l pU/mL, while PAIactivity decreased from 4.52kO.76 to 3.4420.63 AU/mL (26). The authors concluded that high plasma insulin levels may lead to increased levels of PAI- activity in obese subjects. In our study, elevated insulin levels during the waking hours were not associated with an increase in PAIactivity. Peak insulin levels in our study reached 103 pU/mL with sustained levels of 20 to 40 pU/mL during the waking hours with no associated increase activity levels actually fell in PAIactivity. In fact, PAIwhile insulin levels were rising. We conclude that increased insulin levels during the day in healthy non-obese subjects are not associated with an increase in PAI- activity. Regulation of PAIby insulin in obese subjects or over longer time periods may be different. The weak inverse correlation between insulin and PAIactivity was not sufficient, either in individuals or on average, to suggest that insulin is a major factor in controling the circadian rhythm for PAI- activity. Previous studies have shown that corticosteroids can suppress t-PA production while stimulating increased release of PAI(5,6). Thus, elevated levels of cortisol in the morning might in theory lead to increased PAIand decreased t-PA. Our activity begins to rise data did not support this theory. PAIand t-PA activity begins to fall between 2200 h and 2400 h, 4 to 6 hours before cortisol begins to rise. Furthermore, we found an increase in total t-PA antigen in the morning not a decrease. We conclude that cortisol is not associated with the diurnal changes in fibrinolysis. This is supported by a prior study showing that intravenous administration of ACTH in healthy individuals increased cortisol levels but had no effect on PAI- activity (27). Elevated catecholamine levels have been associated with increased fibrinolytic activity in vivo (7). We measured catecholamines in our study but found no evidence of circadian variation
10
INSULIN, CORTISOL & CATECHOLAMINES...
Vol.58, No. 1
in supine epinephrine or norepinephrine in this small sample. Catecholamine levels have been reported to be higher in the morning, but this was associated with samples being drawn in a sitting position: subjects that remained supine showed no change in epinephrine, as we found in our study (28). We conclude that catecholamines are not associated with the circadian variations in fibrinolysis. While our findings are useful, the present research has several limitations. First, we studied only 6 subjects. At least two types of individual circadian patterns were seen, an almost continuous sinusoidal variation and a single peak or trough during the day followed by stable values. Second, we sampled at only 2 hour intervals. We feel that more frequent sampling would not have substantially altered our findings. Third, we did not determine the cause of the increased PAI- activity found in the morning. In summary, we confirmed and better defined the diurnal variations in the principal fibrinolytic factors, t-PA and PAI-1. We hypothesize that the morning decrease in t-PA activity is primarily due to increased PAIactivity at that time. Lastly, we documented that insulin, cortisol, and catecholamines do not explain the circadian changes in fibrinolysis. ACKNOWLEDGEMENTS The authors would like to thank Swee-Chin Loo and Son Nguyen for their help during the project. This research was supported in part by the American Heart Association, Washington Affiliate and by the Medical Research Service of the Veterans Administration. REFERENCES 1. ANGLETON, P., CHANDLER, W.L., and SCHMER, G. Diurnal variation of tissue-type plasminogen activator and its rapid inhibitor (PAI-1). Circulation 79, 101-106, 1989. 2. MULLER, J.E., STONE, P.H., TURI, Z.G., et al. Circadian variation in the frequency of onset of acute myocardial infarction. N. Ensl. J. Med. 313, 1315-1322, 1986. 3. TURTON, M.B. and DEEGAN, T. Circadian variations of plasma catecholamine, cortisol, and immunoreactive insulin concentrations in supine subjects. Clin. Chim. Acta 55, 389-397, 1974. I., VAGUE, P., POISSON, C., AILLAUD, M.F., 4. JUHAN-VAGUE, MENDEZ, C., and COLLEN, D. Effect of 24 hours of normoglycaemia on tissue-type plasminogen activator plasma levels in insulin-dependent diabetes. Throm. Haemost. 51, 97-98, 1984. 5. KRUITHOF, E.K.O. Inhibitors of plasminogen activators. In: Tissue-tvoe nlasminosen activator (t-PA): ohvsioloqical and clinical aspects. Vol I. C. Kluft (Ed.) Boca Raton:CRC Press, 1988, p. 194.
Vol. 58, No. 1
6.
INSULIN, CORTISOL & CATECHOLAMINES...
11
LAUG, W.E. Glucocorticoids inhibit plasminogen activator production by endothelial cells. Throm. Haemost. 50, 888-892, 1983.
7. GADER, A.M.A., CLARKSON, A.R., and CASH, J.D. The plasminogen activator and coagulation factor VIII responses to adrenaline, noradrenaline, isoprenaline and salbutamol in man. Thrombosis Res. 2, 9-16, 1973. 8. FRIBERGER, P. Chromogenic peptide substrates - their use for the assay of factors in the fibrinolytic and plasma kallikrien-kinin systems. Methods for the measurement of plasminogen in plasma. Stand. J. Clin. Lab. Invest. 42 .(Sunnl 162), 49-54, 1982. 9. VERHEIJEN, J-H., MULLAART, E., CHANG, G.T.G., KLUFT, C., and WIJNGAARDS, G. A simple, sensitive spectrophotometric assay for extrinsic (tissue-type) plasminogen activator applicable to measurements in plasma. Throm. Haemost. 48, 266-269, 1982. study to 10. GAFFNEY, P.J. and CURTIS, A.D. A collaborative establish the 2nd international standard for tissue plasminogen activator (t-PA). Throm. Haemost. 58, 1085-1087, 1987. 11. CHANDLER, W.L., SCHMER, G., and STRATTON, J.R. Optimum conditions for the stabilization and measurement of tissue plasminogen activator in human plasma. J. Lab. Clin. Med. 113, 362-371, 1989. W.L., LOO, S.C., NGUYEN, S.V., SCHMER, G., and 12. CHANDLER, of methods for measuring STRATTON, J.R. Standardization plasminogen activator inhibitor (PAI-1) activity in human plasma. Clin. Chem. 35, 787-793, 1989. 13. RANBY, M., BERGSDORF, N., NILSSON, T., MELLBRING, G., WINBLAD, B., and BUCHT, G. Age dependence of tissue plasminogen activator concentrations in plasma as studied by an improved enzyme-linked immunosorbent assay. Clin. Chem. 32, 2160-2165, 1986. 14. WILSON, M.A. and MILES, L.E. Radioimmunoassay of insulin. In: Handbook of Radioimmunoassay. G.E. Abraham (Ed.) New York:Marcel Dekker 1977, p. 275. 15. BREUER, H., HAMEL, D., and KRUSEKEMPER, H. Methods of Hormone Analvsis. New York:John Wiley & Sons, 1976. 16. EVANS, M.I., HALTER, J.B., and PORTER, D. JR. Comparison of double- and single-isotope enzymatic derivative methods for measuring catecholamines in human plasma. Clin. Chem. 24, 567-570, 1978. 17. RUNYON, R.P. Fundamentals of Statistics in the biolosical, medical, and health sciences. Boston:Duxbury Press, 1985.
12
INSULIN, CORTISOL & CATECHOLAMINES...
Vol. 58, No. 1
18. ANDREOTTI, F., DAVIES, G.J., HACKETT, D., KHAN, M.I., DE BART, A., DOOIJEWAARD, G., MASERI, A., and KLUFT, C. Circadian variation of fibrinolytic factors in normal human plasma. Fibrinolvsis 2(Sunol 2), 90-92, 1988. 19. KLUFT, C., JIE, A.F.H., RIJKEN, D.C., and VERHEIJEN, J.H. Daytime fluctations in blood of tissue-type plasminogen activator (t-PA) and its fast-acting inhibitor (PAI-1). Throm. Haemost. 59, 329-332, 1988. 20. KOHLER, M. and MIYASHITA, C. Probleme bei der Messung von Parametern des Fibrinolytischen Systems: Circadianne Rhythmik von Gewebe-Plasminogen-Aktivator und Plasminogen-AktivatorInhibitor. Klin. Wochenschr. 66(Sunnl 12), 62-67, 1988. 21. SPEISER, W., BOWRY, S., ANDERS, E., BINDER, B.R., and MULLER-BERGHAUS, G. Method for the determination of fast acting plasminogen activator inhibitor capacity (PAI-CAP) in plasma, platelets and endothelial cells. Thrombosis Res. 44, 503-515, 1986. 22. KLUFT, C., and ANDREOTTI, F. Consequences of the circadian fluctuation in plasminogen activator inhibitor 1 (PAI-1) for studies on blood fibrinolysis. Fibrinolvsis 2(Sunnl 2), 93-95, 1988. between blood 23. SHAW, D.A. and MCNAUGHTON, D. Relationship fibrinolytic activity and body fatness. Lancet I, 352-354, 1963. 24. GRACE, C.S. and GOLDRICK, R.B. Fibrinolysis and body build. Interrelationships between blood fibrinolysis, body composition, and parameters of lipid and carbohydrate metabolism. J. Atherosclerosis Res. 8, 705-719, 1968. 25. ALESSI, M.C., JUHAN-VAGUE, I., KOOISTRA, T., DECLERCK, P.J., and COLLEN, D. Insulin stimulates the synthesis of plasminogen activator inhibitor 1 by the human hepatocellular cell line Hep G2. Throm. Haemost. 60, 491-494, 1988. 26. VAGUE, P., JUHAN-VAGUE, I., AILLAUE, M.F., BADIER, C., VIARD, R ., ALESSI, M.C., and COLLEN, D. Correlation between blood fibrinolytic activity, plasminogen activator inhibitor level, plasma insulin level, and relative body weight in normal and obese subjects. Metabolism 35, 250-253, 1986. 27. AILLUAD, M.F., JUHAN-VAGUE, I., ALESSI, M.C., MARECAL, M., VINSON, M.F., ARNAUD, C., VAGUE, P., and COLLEN, D. Increased PA-inhibitor levels in the postoperative period - no causeeffect relation with increased cortisol. Throm. Haemost. 54, 466-468, 1985. 28. TOFFLER, G.H., BREZINSKI, D., SCHAFER, A.I., CZEISLER, C.A., RUTHERFORD, M.B., WILLICH, S.N, GLEASON, R.E., WILLIAMS, G.H., and MULLER, J.E. Concurrent morning increase in platelet aggregability and the risk of myocardial infarction and sudden cardiac death. N. Ensl. J. Med. 316, 1514-1518, 1987.
INSULIN,
CORTISOL AND CATECHOLAMINES DO NOT REGULATE VARIATIONS IN FIBRINOLYTIC ACTIVITY
CIRCADIAN
Wayne L. Chandlerl, Dan Mornin' Richard 0. Whittenl Peter Angletonl, Federico M. Farid Thomas R. Fritsche?, Richard C. Veith3, John RI Strattona. Departments of Laboratory Medicine', Medicinea, and the Geriatric Research, Education and Clinical Center, Department of Psychiatry and Behavioral Sciences3, University of Washington, Seattle, WA, 98195, and the Seattle Veterans Administration Medical Center213, USA. (Received 18.8.1989; accepted in revised form 11.1.1990 by Editor N.U. Bang)
ABSTRACT To evaluate possible hormonal regulators of the diurnal rhythm in fibrinolytic activity, we measured tissue plasminogen activator (t-PA) activity, plasminogen activator inhibitor activity (PAI-l), t-PA antigen, insulin, cortisol, and catecholamines in 6 healthy males (age 34+5) every 2 hours for 24 hours. Fibrinolysis was characterized by a peak in PAI- activity and a trough in t-PA activity at 0600 h. PAI- activity increased 92% and t-PA activity decreased 56% between 2400 h and 0600 h. t-PA antigen (principally a measure of t-PA/PAI-1 complex), peaked at 0800 h. In comparison, insulin levels were lowest at night when PAI- activity was rising. The weak inverse correlation between insulin and PAI- activity (r = -0.28, p < 0.02), was not sufficient to explain the diurnal change in fibrinolysis. While cortisol demonstrated the expected circadian change, the increase in cortisol did not occur until 0400 h, 4-6 hours after the rise in PAIand decrease in t-PA activity started. Supine resting plasma epinephrine and norepinephrine showed no circadian rhythm. From this data, we hypothesize that the increased level of PAIin the morning consumes t-PA, leading to decreased t-PA activity and increased t-PA/PAI-1 complex. Further, we conclude that insulin, cortisol, and catecholamines are not responsible for the circadian rhythm of fibrinolysis.
KEY WORDS: fibrinolysis, catecholamine
circadian
1
rhythm,
insulin,
cortisol,
2
INSULIN, CORTISOL & CATECHOLAMINES...
Vol. 58, No. 1
INTRODUCTION The fibrinolytic system plays an important role in regulating the formation and removal of vascular thrombi. In a study in which samples were obtained at 0800 h and 2000 h, we demonstrated that fibrinolytic activity follows a diurnal rhythm with low activity in the morning and higher activity in the evening in both healthy individuals and patients with unstable angina and prior myocardial infarction (1). This circadian variation in fibrinolytic activity may be clinically significant in that arterial thrombotic disease also follows a circadian rhythm (2). To understand the mechanism underlying the circadian rhythm we need a detailed description of circadian of fibrinolysis, variations in the proteins that control fibrinolysis, tissue plasminogen activator (t-PA) and plasminogen activator inhibitor type 1 (PAI-1). Further, we need to understand the regulatory model would be factors that drive this rhythm. One possible circadian hormonal regulation, similar to a number of other better known diurnal rhythms such as the adrenocorticotropic hormone (ACTH) - cortisol system. To determine possible regulators of the diurnal variation in fibrinolytic activity, we measured t-PA activity, PAI- activity, insulin, t-PA antigen, and four possible hormonal regulators, cortisol and the catecholamines epinephrine and norepinephrine in 6 healthy males every 2 hours for 24 hours. These hormones were selected as they had all previously been reported to show a circadian rhythm and were reported to be regulators of fibrinolysis (3). Elevated insulin levels have been associated with increases in plasma PAI- activity (4). Cortisol has been shown to stimulate PAIsecretion and to inhibit t-PA secretion (5,6), while epinephrine infusions have been reported to increase fibrinolytic activity (7). The results from our study indicate, that while these four hormones may be important in regulating fibrinolysis in other situations, they are not responsible for the circadian rhythm of the fibrinolytic system. MATERIALS AND METHODS Human subiects Studies on human subjects principles of the Declaration obtained from all participants University of Washington Human
were carried out according to the of Helsinki. Informed consent was and the study was approved by the Subjects Review Committee.
Materials Cyanogen bromide (CNBr), bovine albumin (98% pure, salt free) and Triton-X-100 were obtained from Sigma Chemical Co. (St. Louis, MO). Chromogenic substrate D-valyl-phenylalanyllysyl-pnitro-anilide (S-2390) was obtained from Helena Diagnostics (Beaumont, TX). All other materials not described below were reagent or analytical grade. Human glu-plasminogen (product no. 400, lot no. GO187 and G0287) was obtained from American Diagnostica Inc. (Greenwich, CO). The specific activity of the plasminogen preparation was approximately 4.7 U/mg where 1 U = the plasminogen activity in 1 mL of pooled normal plasma (8). Human fi-
Vol. 58, No. 1
INSULIN, CORTISOL & CATECHOIAMINES...
3
brinogen (96% clottable, product no. F4883, lot no. 67F-9461) was obtained from Sigma Chemical Co. (St. Louis, MO). Human fibrinogen was cleaved with CNBr as described by Verheijen et al. (9). Gnechain melanoma derived t-PA (product no. 115, lot no. 7375), was obtained from American Diagnostica Inc. The specific activity of the t-PA preparation was 500,000 IU/mg as compared to the second international t-PA activity standard (86/670) from the National Institute for Biologic Standards and Control, London, UK (lo). Polyclonal goat-anti-human melanoma t-PA and its peroxidase conjugate were obtained from American Diagnostica Inc. Polyclonal anti-cortisol and 1251-cortisol tracer were obtained from Baxter Travenol Diagnostics (Cambridge, MA). Polyclonal anti-human insulin and 1251-insulin tracer were obtained from Radioassay Systems Laboratories, Inc. (Carson, CA). Blood samnlins and samnle orenaration Each subject had an intravenous catheter placed in a forearm vein prior to the start of the study. The catheters were closed with a heparin lock filled with 1 mL of 100 U/mL heparin between sampling periods. Samples were taken every two hours for 24 hours. The first samples were taken at 1000 h. The subjects followed their normal daily activity with the exception that prior to each sample the subjects rested in the supine position for 15 min. After the samples were taken at midnight, the subjects slept until approximately 0600 h except for brief disturbances during sample acquisition at 0200 h and 0400 h. At each sampling period, the following protocol was followed: 1) the heparin lock was flushed with 5 mL of sterile saline, 2) 3 mL of blood were drawn and discarded, 3) 2.5 mL of whole blood were drawn and immediately added to a prechilled tube containing 50 UL of 0.15 mol/L NaCl, 90 gm/L EGTA, and 60 gm/L reduced glutathione (catecholamine sample), 4) 3 mL of blood were drawn and allowed to clot (cortisol and insulin sample), 5) 9 mL of blood were drawn into 1 mL of 130 mmol/L sodium citrate (PAIactivity and total t-PA antigen samples), 6) 1 mL of the citrate anticoagulated whole blood was added to 0.5 mL of 0.5 mol/L sodium acetate buffer, pH 4.2 (t-PA activity sample), and 7) the heparin lock was refilled with 1 mL of 100 U/mL heparin. The catecholamine samples were centrifuged at 4'C for 15 min at 9OOg, the plasma removed and recentrifuged as before, and the supernatant removed and stored in tightly sealed plastic tubes at -8OOC. After clotting for 15 min at room temperature, the unanticoagulated blood samples were centrifuged at 4OC for 10 min at 25OOg, the serum removed and stored at -8OOC. The citrate anticoagulated and acidified blood samples were centrifuged at room temperature for 5 min at 25OOg, and the plasma removed and stored at -8OOC. Activated partial thromboplastin times run on titrated plasma samples indicated they contained less than 0.1 U/mL heparin activity. Measurement of t-PA activitv and PAI- activity t-PA activity and PAI- activity were measured as PreViOUSly described using amidolytic assays (11,12). The final results for both the t-PA activity and PAIactivity assays were corrected for dilution by the anticoagulant and acidification buffer. The
4
INSULIN, CORTISOL & CATECHOLAMINES...
Vol. 58, No. 1
imprecision for replicate analysis of a single sample of acidified plasma within a single run of the t-PA activity assay was 0.96+0.03 [SD] IU/mL (CV = 3%, n = 16). The imprecision of replicate analysis of a single sample run on different days for the t-PA activity assay was 0.92+0.06 [SD] IU/mL (CV = 7%, n = 5). In the PAI-I activity assay, one arbitrary unit (1 AU) of PAIactivity is defined as the amount of PAI- that inhibits 1 IU of assay is one-chain t-PA activity. In our laboratory the PAIdesigned to measure up to four plasma samples per run. The imprecision for replicate analysis of a single sample of plasma within a single run of the PAI- activity assay was 19.5kO.3 [SD] AU/mL of replicate analysis of a (CV = 2%, n = 4). The imprecision single sample run on different days for the PAI- activity assay was 10.220.7 [SD] AU/mL (CV = 7%, n = 20). Total
t-PA antisen assay Total t-PA antigen was determined as described previously using an enzyme-linked immunosorbent assay (13). The final results were corrected for dilution of the plasma by citrate antifor replicate analysis of a single coagulant. The imprecision sample of titrated plasma within a single run of the t-PA antigen assay was 7.9LO.2 [SD] ng/mL (CV = 3%, n = 8). The imprecision of replicate analysis of a single sample run on different days for the t-PA antigen assay was 5.8kO.5 [SD] ng/mL (CV = 9%, n = 9). Other
Assavs Cortisol and immunoreactive insulin were measured using radioimmunoassays (14,15). Epinephrine and norepinephrine were measured using a single isotope radioenzymatic assay (16). Statistics The mean and standard error of the mean (SEM) were determined for each sampling time for each assay type. For a given sampling time, results for t-PA antigen, insulin, cortisol, epinephrine and norepinephrine were normally distributed among subjects. Results for t-PA activity and PAIactivity demonstrated wide variations in values among different subjects (skewed distributions). Simple means for these two assays primarily reflected the results of the subjects with the highest values, not the average circadian rhythm of the group. Therefore, for these two assays, we normalized the results by dividing a given subject's results by the peak value for that subject. These results were expressed in percent, thus the peak value for each subject was 100% with all other values being expressed as a percent of the peak. These percent values were then averaged for each time of day for the six subjects. For each assay type, we first performed a one-way analysis of variance (ANOVA) vs time of day. Those results showing significant differences among different times of day (P c 0.01) were further analyzed using the Tukey Honestly-Significant-Difference IIIUltiCOIIIpariSOn test at P = 0.05 and P = 0.01 levels of significance (17). Original data and copies of the statistical evaluation tables are available from the authors.
Vol. 58, No. 1
INSULIN, CORTISOL & CATECHOLAMINES...
5
RESULTS t-PA Activitv All six subjects demonstrated a trough in t-PA activity in the morning and a peak in the afternoon or evening. There was a wide range in t-PA activity levels among different subjects. Peak activities ranged from 1.0 to 2.6 IU/mL, and trough activities from 0.4 to 1.1 IU/mL. The time of trough t-PA activity ranged from 0600 h to 1200 h, but occurred most commonly (4 of 6 subjects) at 0600 h. The time of peak t-PA activity ranged from 1200 h to 2200 h. While each of the individual subjects showed circadian variations in t-PA activity, there were no significant differences in average absolute t-PA activity (IU/mL) among the times tested (one-way ANOVA, p = 0.9). This was due to the wide range of values (skewed distribution) and variations in peak and trough times (phase shifts) among different subjects. To eliminate the effect of the wide range in values among subjects, we compared the average normalized percent t-PA activity (see methods) vs time of day (Fig. 1). Percent activity data demonstrated significant differences among different times of day (one-way ANOVA, p = 0.008), with lower t-PA activities between 0400 h and 1000 h compared to all other times of day (p = 0.01). If the trough values for each subject were phase shifted to the same time of day (0600 h), the apparent average rhythm became even stronger (one-way ANOVA, p CO.001). It should be noted that the data in Figure 1 was not phase shifted, but represents the average for the six subjects for each time of day. The mean percent decrease in t-PA activity from peak to nadir was 56+5%. PAI-
Activitv PAI- activity followed an inverse pattern compared to t-PA activity with a peak in the morning and a trough in the evening. Again the data were skewed (peak range 7.7 to 12.4 AU/mL, trough range 0.0 to 7.6 AU/mL) and the peak and trough times inconsistent among subjects (peak time range 0200 h to 0600 h, trough time range 1400 h to 2200 h). There were significant differences among average absolute PAI- activities (AU/mL) vs time of day (one-way ANOVA, p = 0.03), but as before the comparison reached greater significance when normalized percent PAIactivities were comwere pared (Fiq. 1, p = 0.002), and when the peak activities phase shifted-to the same time of day (p < 0.001). The mean percent decrease in PAI- activity from peak to nadir was 63+28%. Total t-PA Antiaen Total t-PA antigen followed a pattern similar to that for PAIactivity, with a peak in the morning and a trough in the t-PA antigen results ANOVA, .p _a = 0.001). evening (Fig. . 1, one-way _. . _ showed less inter-subject variation (peak range 6.2 to 10.7 ng/mL, trough range 3.4 to 5.1 ng/mL). Peak levels for t-PA antigen occurred between 0400 h and 0800 h. Between 1400 h and 2400 h there was little variation in results, no consistent trough time was noted. The mean percentage decrease in total t-PA antigen from peak to nadir was 45+13%.
Vol. 58, No. 1
CORTISOL & CATECHOLAMINES...
INSULIN.
FIG. 1
O’;;oo
“1
2400
Average circadian variations in insulin, cortisol, plasminogen activator inhibitor activity, type 1 (PAI-1) tissue plasminogen activator (TPA) activity, and TPA antigen in 6 healthy subjects for each time of day. Error bars indicate mean+SEM.
1
2400
o&o
12bo
l&O
24bO
100
50
1%
loo-
IJ6o---%-@++@ 2400 owa
1200
TIME
E 2lo 2400 I
I
I
I I
I
I
0600
1200
moo
I
MOO
2400
(h)
FIG. 2
2400
Circadian variations in tissue plasminogen activator (TPA) activity, plasminogen activator inhibitor type 1 (PAI-1) activity,and TPA antigen in subject A (upper panel) and subject B (lower panel). 2400
0600
TIME
1200
(h)
1200
2400
INSULIN, CORTISOL & CATECHOLAMINES...
Vol. 58, No. 1
7
Individual Patterns In general, two types of individual circadian patterns were seen in the fibrinolytic factors. The first pattern consisted of an almost continuous sinusoidal variation in PAIand t-PA activity (Subject A, Fig. 2). With this pattern there was essentially no time during which the levels were stable. Subject A demonstrated an almost 5 fold change in PAI- activity during the day and a 4 fold change in t-PA activity with a inverse correlation between PAIand t-PA activity (1:= -0.83, p < 0.001). As with the averaged data, total t-PA antigen followed PAI- activity but demonstrated a peak and trough that occurred approxmately 4 hours after the equivalent times for PAI-1. The second type of circadian pattern showed an essentially stable level of PAIactivity, t-PA activity and t-PA antigen from about 1200 h to 2400 h (Subject B, Fig. 2). Subject B also showed a inverse correlation between t-PA and PAI- activity (r = activity levels began to -0.77, p = < 0.01). At 2400 h, PAIincrease and t-PA activity to decrease reaching their respective peak and trough at 0600 h, then returning to baseline by 1200 h. Baseline (i.e. trough) levels of PAI- activity in Subject B were approximately 4 times higher than those seen in Subject A (sinusoidal type of pattern). Insulin The six subjects demonstrated significant circadian variations in plasma insulin (one-way ANOVA, p = 0.006). Insulin levels were highest and most variable during the waking hours (1000 h to 2400 h), dropping rapidly during the night (0200 h to 0800 h) to a lower, more stable level (Fig. 1). This pattern also corresponded to the subjects eating patterns. Insulin levels were 60
FIG. 3 Circadian variations in epinephrine and norepinephrine in 6 healthy subjects. Error bars indicate mean+SEM.
40 20
I
2400
0600
2400
0600
TIME
1200
1600
(h)
2400
8
INSULIN, CORTISOL & CATECHOLAMINES...
Vol. 58, No. 1
falling when PAI- activity was rising during the night. All six individuals demonstrated weak inverse correlations between insulin and PAI- activity (overall, r = -0.28, p < 0.02). Cortisol All six subjects demonstrated the expected diurnal change in cortisol: low levels at night rising rapidly after 0400 h to a peak at 0600 h, then slowly falling through the day back to baseline levels (Fig. 1, one-way ANOVA, p < 0.001). In contrast to the skewed values for PAI- and t-PA activity, cortisol levels demonstrated relatively narrow ranges at all times during the day with no evidence of phase shifts among subjects. At the time of the cortisol increase (0400 h to 0600 h), PAIactivity had already been rising for 4 to 6 hours (Fig. 1). Catecholamines Both plasma epinephrine and norepinephrine were uniformly low throughout the study, consistent with resting supine levels of a circadian variation in (Fig. 3). There was no evidence epinephrine or norepinephrine (one-way ANOVA, p > 0.7). DISCUSSION Circadian Variabilitv of Fibrinolvtic Factors This study, in which samples were drawn at two hour intervals, provides the most detailed description currently available of circadian variations in t-PA activity, PAIactivity and total t-PA antigen. The mean percentage fall from peak to nadir was 56+5% for t-PA activity, 63+28% for PAI- activity and 45+13% for total t-PA antigen. Thus, all three variables showed significant diurnal variability, confirming prior studies that sampled at less frequent intervals (1,18-21). To understand what controls the circadian rhythm of fibrinolytic activity, we must determine which fibrinolytic factors are responsible for the diurnal change and what the mechanism is that drives it. Overall, the lowest t-PA activity, highest PAIactivity, and highest total t-PA antigen were seen in the morning at approximately 0600 h. Assuming the clearance of these factors does not change during the day, then the trough in t-PA activity during the morning is most likely due to either decreased secretion of t-PA or increased inhibition of t-PA by PAI-1. If t-PA secretion were decreased we would expect the total level of t-PA antigen in the blood to fall in the morning, which did not occur. In fact, total t-PA antigen was highest in the morning when t-PA activity was at its nadir. It should be noted that in non-venous occlusion plasma, total t-PA antigen assays measure almost exclusively t-PA/PAI-1 complex (13). We hypothesize that the increased level of PAI- in the morning consumes t-PA, leading to decreased t-PA activity and increased t-PA/PAI-1 complex. Evaluation of the circadian rhythm of fibrinolysis is complicated by differences in the amplitude, phase and pattern of the rhythm among subjects (22). Healthy individuals in our study showed a wide range in both t-PA and PAIactivities. Trough levels in one individual were higher than the peak in another subject. Second, the timing of peak and trough levels were not
Vol. 58, No. 1
INSULIN, CORTISOL & CATECHOLAMINES...
9
consistent among individuals, but demonstrated phase shifts of 2 to 4 hours. Third, there appeared to be at least two types of circadian patterns. Three of the subjects demonstrated a sinusoidal variation throughout the day with essentially no stable periods, while the other three showed a single peak or trough in the morning that returned to a stable baseline for the rest of the day. The amount of variability noted in the fibrinolytic rhythms was greater than that seen for other factors that show a circadian rhythm. For example, cortisol levels in our subjects fell within narrow ranges at all times during the day, were all in phase showing a dramatic rise between 0400 h and 0600 h, and demonstrated only one type of circadian pattern. Hormonal Recrulation of Circadian Chancres in Fibrinolvsis We measured four hormones previously reported to show circadian rhythms, insulin, cortisol, epinephrine and norepinephrine (3) - Insulin followed a circadian rhythm with peak levels during the waking hours and lower levels at night. We found a weak inverse correlation between PAIactivity and insulin; this is in contrast to previous studies that have associated obesity and elevated levels of insulin with increased plasma PAIactivity and increased production of PAI- in hepatoma cells (4,23-25). In obese subjects that were fasted for 24 hours, insulin levels have been shown to fall on average from 22.3k2.2 to 16.3kl.l pU/mL, while PAIactivity decreased from 4.52kO.76 to 3.4420.63 AU/mL (26). The authors concluded that high plasma insulin levels may lead to increased levels of PAI- activity in obese subjects. In our study, elevated insulin levels during the waking hours were not associated with an increase in PAIactivity. Peak insulin levels in our study reached 103 pU/mL with sustained levels of 20 to 40 pU/mL during the waking hours with no associated increase activity levels actually fell in PAIactivity. In fact, PAIwhile insulin levels were rising. We conclude that increased insulin levels during the day in healthy non-obese subjects are not associated with an increase in PAI- activity. Regulation of PAIby insulin in obese subjects or over longer time periods may be different. The weak inverse correlation between insulin and PAIactivity was not sufficient, either in individuals or on average, to suggest that insulin is a major factor in controling the circadian rhythm for PAI- activity. Previous studies have shown that corticosteroids can suppress t-PA production while stimulating increased release of PAI(5,6). Thus, elevated levels of cortisol in the morning might in theory lead to increased PAIand decreased t-PA. Our activity begins to rise data did not support this theory. PAIand t-PA activity begins to fall between 2200 h and 2400 h, 4 to 6 hours before cortisol begins to rise. Furthermore, we found an increase in total t-PA antigen in the morning not a decrease. We conclude that cortisol is not associated with the diurnal changes in fibrinolysis. This is supported by a prior study showing that intravenous administration of ACTH in healthy individuals increased cortisol levels but had no effect on PAI- activity (27). Elevated catecholamine levels have been associated with increased fibrinolytic activity in vivo (7). We measured catecholamines in our study but found no evidence of circadian variation
10
INSULIN, CORTISOL & CATECHOLAMINES...
Vol.58, No. 1
in supine epinephrine or norepinephrine in this small sample. Catecholamine levels have been reported to be higher in the morning, but this was associated with samples being drawn in a sitting position: subjects that remained supine showed no change in epinephrine, as we found in our study (28). We conclude that catecholamines are not associated with the circadian variations in fibrinolysis. While our findings are useful, the present research has several limitations. First, we studied only 6 subjects. At least two types of individual circadian patterns were seen, an almost continuous sinusoidal variation and a single peak or trough during the day followed by stable values. Second, we sampled at only 2 hour intervals. We feel that more frequent sampling would not have substantially altered our findings. Third, we did not determine the cause of the increased PAI- activity found in the morning. In summary, we confirmed and better defined the diurnal variations in the principal fibrinolytic factors, t-PA and PAI-1. We hypothesize that the morning decrease in t-PA activity is primarily due to increased PAIactivity at that time. Lastly, we documented that insulin, cortisol, and catecholamines do not explain the circadian changes in fibrinolysis. ACKNOWLEDGEMENTS The authors would like to thank Swee-Chin Loo and Son Nguyen for their help during the project. This research was supported in part by the American Heart Association, Washington Affiliate and by the Medical Research Service of the Veterans Administration. REFERENCES 1. ANGLETON, P., CHANDLER, W.L., and SCHMER, G. Diurnal variation of tissue-type plasminogen activator and its rapid inhibitor (PAI-1). Circulation 79, 101-106, 1989. 2. MULLER, J.E., STONE, P.H., TURI, Z.G., et al. Circadian variation in the frequency of onset of acute myocardial infarction. N. Ensl. J. Med. 313, 1315-1322, 1986. 3. TURTON, M.B. and DEEGAN, T. Circadian variations of plasma catecholamine, cortisol, and immunoreactive insulin concentrations in supine subjects. Clin. Chim. Acta 55, 389-397, 1974. I., VAGUE, P., POISSON, C., AILLAUD, M.F., 4. JUHAN-VAGUE, MENDEZ, C., and COLLEN, D. Effect of 24 hours of normoglycaemia on tissue-type plasminogen activator plasma levels in insulin-dependent diabetes. Throm. Haemost. 51, 97-98, 1984. 5. KRUITHOF, E.K.O. Inhibitors of plasminogen activators. In: Tissue-tvoe nlasminosen activator (t-PA): ohvsioloqical and clinical aspects. Vol I. C. Kluft (Ed.) Boca Raton:CRC Press, 1988, p. 194.
Vol. 58, No. 1
6.
INSULIN, CORTISOL & CATECHOLAMINES...
11
LAUG, W.E. Glucocorticoids inhibit plasminogen activator production by endothelial cells. Throm. Haemost. 50, 888-892, 1983.
7. GADER, A.M.A., CLARKSON, A.R., and CASH, J.D. The plasminogen activator and coagulation factor VIII responses to adrenaline, noradrenaline, isoprenaline and salbutamol in man. Thrombosis Res. 2, 9-16, 1973. 8. FRIBERGER, P. Chromogenic peptide substrates - their use for the assay of factors in the fibrinolytic and plasma kallikrien-kinin systems. Methods for the measurement of plasminogen in plasma. Stand. J. Clin. Lab. Invest. 42 .(Sunnl 162), 49-54, 1982. 9. VERHEIJEN, J-H., MULLAART, E., CHANG, G.T.G., KLUFT, C., and WIJNGAARDS, G. A simple, sensitive spectrophotometric assay for extrinsic (tissue-type) plasminogen activator applicable to measurements in plasma. Throm. Haemost. 48, 266-269, 1982. study to 10. GAFFNEY, P.J. and CURTIS, A.D. A collaborative establish the 2nd international standard for tissue plasminogen activator (t-PA). Throm. Haemost. 58, 1085-1087, 1987. 11. CHANDLER, W.L., SCHMER, G., and STRATTON, J.R. Optimum conditions for the stabilization and measurement of tissue plasminogen activator in human plasma. J. Lab. Clin. Med. 113, 362-371, 1989. W.L., LOO, S.C., NGUYEN, S.V., SCHMER, G., and 12. CHANDLER, of methods for measuring STRATTON, J.R. Standardization plasminogen activator inhibitor (PAI-1) activity in human plasma. Clin. Chem. 35, 787-793, 1989. 13. RANBY, M., BERGSDORF, N., NILSSON, T., MELLBRING, G., WINBLAD, B., and BUCHT, G. Age dependence of tissue plasminogen activator concentrations in plasma as studied by an improved enzyme-linked immunosorbent assay. Clin. Chem. 32, 2160-2165, 1986. 14. WILSON, M.A. and MILES, L.E. Radioimmunoassay of insulin. In: Handbook of Radioimmunoassay. G.E. Abraham (Ed.) New York:Marcel Dekker 1977, p. 275. 15. BREUER, H., HAMEL, D., and KRUSEKEMPER, H. Methods of Hormone Analvsis. New York:John Wiley & Sons, 1976. 16. EVANS, M.I., HALTER, J.B., and PORTER, D. JR. Comparison of double- and single-isotope enzymatic derivative methods for measuring catecholamines in human plasma. Clin. Chem. 24, 567-570, 1978. 17. RUNYON, R.P. Fundamentals of Statistics in the biolosical, medical, and health sciences. Boston:Duxbury Press, 1985.
12
INSULIN, CORTISOL & CATECHOLAMINES...
Vol. 58, No. 1
18. ANDREOTTI, F., DAVIES, G.J., HACKETT, D., KHAN, M.I., DE BART, A., DOOIJEWAARD, G., MASERI, A., and KLUFT, C. Circadian variation of fibrinolytic factors in normal human plasma. Fibrinolvsis 2(Sunol 2), 90-92, 1988. 19. KLUFT, C., JIE, A.F.H., RIJKEN, D.C., and VERHEIJEN, J.H. Daytime fluctations in blood of tissue-type plasminogen activator (t-PA) and its fast-acting inhibitor (PAI-1). Throm. Haemost. 59, 329-332, 1988. 20. KOHLER, M. and MIYASHITA, C. Probleme bei der Messung von Parametern des Fibrinolytischen Systems: Circadianne Rhythmik von Gewebe-Plasminogen-Aktivator und Plasminogen-AktivatorInhibitor. Klin. Wochenschr. 66(Sunnl 12), 62-67, 1988. 21. SPEISER, W., BOWRY, S., ANDERS, E., BINDER, B.R., and MULLER-BERGHAUS, G. Method for the determination of fast acting plasminogen activator inhibitor capacity (PAI-CAP) in plasma, platelets and endothelial cells. Thrombosis Res. 44, 503-515, 1986. 22. KLUFT, C., and ANDREOTTI, F. Consequences of the circadian fluctuation in plasminogen activator inhibitor 1 (PAI-1) for studies on blood fibrinolysis. Fibrinolvsis 2(Sunnl 2), 93-95, 1988. between blood 23. SHAW, D.A. and MCNAUGHTON, D. Relationship fibrinolytic activity and body fatness. Lancet I, 352-354, 1963. 24. GRACE, C.S. and GOLDRICK, R.B. Fibrinolysis and body build. Interrelationships between blood fibrinolysis, body composition, and parameters of lipid and carbohydrate metabolism. J. Atherosclerosis Res. 8, 705-719, 1968. 25. ALESSI, M.C., JUHAN-VAGUE, I., KOOISTRA, T., DECLERCK, P.J., and COLLEN, D. Insulin stimulates the synthesis of plasminogen activator inhibitor 1 by the human hepatocellular cell line Hep G2. Throm. Haemost. 60, 491-494, 1988. 26. VAGUE, P., JUHAN-VAGUE, I., AILLAUE, M.F., BADIER, C., VIARD, R ., ALESSI, M.C., and COLLEN, D. Correlation between blood fibrinolytic activity, plasminogen activator inhibitor level, plasma insulin level, and relative body weight in normal and obese subjects. Metabolism 35, 250-253, 1986. 27. AILLUAD, M.F., JUHAN-VAGUE, I., ALESSI, M.C., MARECAL, M., VINSON, M.F., ARNAUD, C., VAGUE, P., and COLLEN, D. Increased PA-inhibitor levels in the postoperative period - no causeeffect relation with increased cortisol. Throm. Haemost. 54, 466-468, 1985. 28. TOFFLER, G.H., BREZINSKI, D., SCHAFER, A.I., CZEISLER, C.A., RUTHERFORD, M.B., WILLICH, S.N, GLEASON, R.E., WILLIAMS, G.H., and MULLER, J.E. Concurrent morning increase in platelet aggregability and the risk of myocardial infarction and sudden cardiac death. N. Ensl. J. Med. 316, 1514-1518, 1987.
