Elevated tissue plasminogen activator as a potential marker of endothelial dysfunction in pre-eclampsia: correlation with proteinuria

Authors

  • Luís Belo,

    Corresponding author
    1. Department of Biochemistry, Faculty of Pharmacy, University of Porto, Portugal
    2. Institute for Molecular and Cell Biology, University of Porto, Portugal
      * L. Belo, Department of Biochemistry, Faculty of Pharmacy, University of Porto, 4050-047 Porto, Portugal.
    Search for more papers by this author
  • Alice Santos-Silva,

    1. Department of Biochemistry, Faculty of Pharmacy, University of Porto, Portugal
    2. Institute for Molecular and Cell Biology, University of Porto, Portugal
    Search for more papers by this author
  • Ann Rumley,

    1. University Department of Medicine, Glasgow Royal Infirmary, University of Glasgow, UK
    Search for more papers by this author
  • Gordon Lowe,

    1. University Department of Medicine, Glasgow Royal Infirmary, University of Glasgow, UK
    Search for more papers by this author
  • Luís Pereira-Leite,

    1. Department of Obstetrics and Gynaecology, Hospital of S. João, Porto Medical School, Portugal
    Search for more papers by this author
  • Alexandre Quintanilha,

    1. Institute for Molecular and Cell Biology, University of Porto, Portugal
    2. Institute of Biomedical Sciences of Abel Salazar, University of Porto, Portugal
    Search for more papers by this author
  • Irene Rebelo

    1. Department of Biochemistry, Faculty of Pharmacy, University of Porto, Portugal
    2. Institute for Molecular and Cell Biology, University of Porto, Portugal
    Search for more papers by this author

* L. Belo, Department of Biochemistry, Faculty of Pharmacy, University of Porto, 4050-047 Porto, Portugal.

Abstract

Objective To clarify the role played by endothelial cell production of fibrinolytic factors in normal pregnancy and pre-eclampsia.

Design A longitudinal study performed during normal pregnancy and a cross sectional study performed in healthy and pre-eclamptic pregnant women in the third trimester of pregnancy.

Setting Department of Obstetrics and Gynaecology, University Hospital of S. João, Porto, Portugal.

Population Fourteen normal pregnant women followed through the three trimesters of gestation. Two groups of women (normal, n= 56; pre-eclamptic, n= 37) evaluated at the third trimester of gestation.

Methods Measurement of platelet number, plasma levels of fibrinogen, tissue plasminogen activator (t-PA) antigen, plasminogen activator inhibitor-1 (PAI-1) activity, and fibrin fragment D-dimer.

Main outcome measures Pre-eclampsia, proteinuria.

Results All the substances, except platelet count, increased significantly throughout normal pregnancy. Comparison of the values in the third trimesters of normal and pre-eclamptic pregnancies showed similar values for the fibrinogen and platelet counts, and higher values of t-PA (almost twice normal median value; P < 0.0001), PAI-1 and D-dimer in the pre-eclamptic women. t-PA correlated positively and significantly with the degree of proteinuria in pre-eclamptic women (r= 0.575, P= 0.0002).

Conclusion These findings suggest that elevated t-PA antigen may reflect endothelial disturbance in pre-eclampsia, and may be a potential biomarker of risk.

Introduction

Profound changes in coagulation and fibrinolytic mechanisms occur during normal human pregnancy1–3. A hypercoagulable state develops in normal pregnancy and is a physiologic adaptation to prevent major haemorrhage during and after placental separation. This hypercoagulable state may also contribute to the increased susceptibility of pregnant women to thrombotic disorders4. Plasma levels of several coagulation factors are increased1,5, and higher concentrations of plasminogen activator inhibitors types 16–9 and 27,9 (PAI-1 and -2, respectively), known to suppress fibrinolysis, are also reported. Thus, there is a predisposition to thromboembolism in late pregnancy.

Pre-eclampsia, a major cause of maternal and fetal morbidity and mortality, is a multisystem disorder of unknown cause. The imbalance of haemostasis observed in normal pregnancy seems to be increased in pre-eclampsia10–12. Endothelial cell damage or activation is believed to play a central role in pre-eclampsia and may underlie the haemostatic changes observed in this syndrome13,14. While normal endothelial cells participate in the regulation of haemostasis, perturbed vascular cells may express prothrombotic properties promoting pathologic events15.

Tissue-type plasminogen activator (t-PA) produced by endothelial cells plays an important role in fibrinolysis by promoting the conversion of plasminogen to plasmin. The protease plasmin cleaves fibrin to produce fibrin degradation products, usually assayed as cross-linked fibrin fragments (D-dimer). The evaluation of D-dimer can be used as a measure of the formation (via thrombin) and lysis (via plasmin) of fibrin. The activity of plasminogen activators is controlled by PAI. The endothelial cell-type PAI (PAI-1) is believed to be the main inhibitor of t-PA, both being produced by the endothelial cell. PAI-2 also plays an important role in pregnancy, being produced by the placenta.

The aim of our work was to clarify the role played by endothelial cells in both normal and pre-eclamptic pregnancies. We measured the changes in the platelet count, plasma fibrinogen concentration and in fibrinolytic products, which may reflect endothelial cell function, namely t-PA antigen, PAI-1 activity and fibrin fragment D-dimer. A longitudinal study was performed in the three trimesters of pregnancy in healthy women. A cross sectional study was also performed to compare healthy and pre-eclamptic pregnant women in the third trimester.

Methods

The protocol used for all women during pregnancy was approved by the Committee on Ethics of the University Hospital S. João, Porto and women gave their consent to participate in the study. Clinical data regarding the women were collected at the obstetric service of this institution. Normal pregnancy was diagnosed on the basis of clinical and ultrasound findings. Healthy pregnant women had a normal course and outcome of pregnancy and did not receive any medication known to interfere with haemostasis. Pre-eclampsia was diagnosed by gestational hypertension accompanied by proteinuria, oedema or both. In agreement with the Committee on Terminology of the American College of Obstetricians and Gynaecologists, gestational hypertension was defined as an increase by at least 30 mmHg in systolic or 15 mmHg in diastolic blood pressure, when compared with the values obtained before 20 weeks of gestation; or a sustained blood pressure of at least 140/90 mmHg after 20 weeks of gestation, if prior blood pressure was not known. Proteinuria was defined as the excretion of 300 mg or more of protein in a 24-h urine collection specimen. This usually agrees with 30 mg/dl (1+ on dipstick testing) or higher in a random specimen of urine. Since proteinuria may be confounded by infection, routine bacteriologic tests were performed. Oedema was diagnosed clinically, or as a rapid increase of weight. Women with pre-eclampsia received antihypertensive therapy and a low salt diet. Pregnant women with significant medical disorders other than pre-eclampsia were excluded from the study.

A longitudinal study was performed during the three trimesters of normal pregnancy in 14 healthy pregnant women. A cross sectional study was also performed in women with pre-eclampsia and in healthy women, matched for gestational age, in the third trimester of pregnancy.

Blood samples obtained during pregnancy were non-fasting and processed within 2 h of collection. Blood was obtained by venipuncture in sterile tubes containing EDTA as anticoagulant. After centrifugation, plasma was separated and aliquots were taken for immediate storage at −70°C until assayed. Plasma fibrinogen was measured using an immunologic technique on a Behring nephelometer, using reagents and standards from the manufacturer. Commercially available kits from Biopool (Umeå, Sweden) were used to measure plasma levels of t-PA antigen, fibrin D-dimer (enzyme linked immunosorbent assays) and PAI-1 activity (a two-step, indirect enzymatic assay). Platelet count was performed by using an automatic blood cell counter (ABX Micros 60-OT).

Kolmogorov–Smirnov analyses were used to test if the results were normally distributed. In the longitudinal study, total comparison between trimesters was performed using repeated measures analysis of variance (Huynh–Feldt adjustment). When differences reached statistical significance, multiple comparisons between trimesters were made using Bonferroni adjustment. In the cross sectional study, the groups were compared using Student's unpaired t test. Data not normally distributed were compared by the Mann–Whitney U test. The strength of the association between the substances was estimated by Pearson's correlation coefficient (r) or Spearman's rank correlation coefficient (rs, for data that are not normally distributed). Moderate and extreme outliers were defined as cases with values between 1.5 and 3 interquartile ranges or more than 3 interquartile ranges (respectively) from the third and first quartiles.

Results

Table 1 summarises the clinical characteristics of the healthy pregnant women in the longitudinal study. As normal pregnancy progresses, there was an increase in weight and in uric acid level. No significant changes were seen in the blood pressure of healthy pregnant women.

Table 1.  Longitudinal study in normal pregnancy. Values are mean (SD).
 TrimesterTotal comparison (first/second/third)Multiple comparisons
First (n= 14)Second (n= 14)Third (n= 14)FPFirst/second (P)Second/third (P)First/third (P)
  1. Repeated measures analysis of variance (Huynh–Feldt adjustment).

  2. Repeated measures analysis of variance (Bonferroni adjustment).

  3. aOne missing case.

  4. bTwo missing cases.

Clinical data
Age (years)24.2 (3.9)       
Gestational age (weeks)9.9 (2.5)23.8 (1.5)34.3 (1.9)     
Weight (kg)63.7 (9.7)69.5 (9.9)74.8 (9.3)104.5<0.0001<0.0001<0.0001<0.0001
Uric acid (mg/dL)3.2 (0.6)3.5 (0.8)4.3 (0.9)21.4a<0.00010.10750.00280.0006
Blood pressure (mmHg)
  Systolic114 (13)115 (8)116 (11)0.1670.8472   
  Diastolic59 (9)58 (8)60 (6)0.2760.7612   
Haemostatic data
Fibrinogen (g/L)3.70 (1.00)4.32 (0.77)5.12 (0.89)19.9<0.00010.08630.00290.0001
t-PA (ng/mL)5.7 (1.4)7.6 (1.6)8.5 (2.3)13.70.00020.00170.38980.0032
PAI-1 (U/mL)19.5 (3.9)38.1 (4.4)68.8 (9.6)203.5<0.0001<0.0001<0.0001<0.0001
D-dimer (ng/mL)122 (72)272 (201)377 (279)10.5a0.00630.02310.02610.0202
Platelets (103/μL)220 (53)215 (49)180 (44)7.3b0.00730.84720.00560.0245

In the third trimester, women with pre-eclampsia had proteinuria, greater weight, greater systolic and diastolic blood pressures and greater uric acid concentrations (Table 2). No significant difference was found between maternal age and gestational age at sampling in both groups.

Table 2.  Cross sectional study: comparison of normal and pre-eclamptic women in the third trimester. Values are mean (SD) or median [25th to 75th centiles].
 Third trimesterP
Normal (n= 56)Pre-eclamptic (n= 37)
  1. n.d. = non-detectable.

  2. *Student's unpaired t test.

  3. Mann–Whitney U test.

Clinical data
Age (years)25.3 (3.6)27.0 (5.8)0.1216*
Gestational age (weeks)33.6 (2.9)34.4 (3.5)0.2918*
Weight (kg)74.2 (10.3)80.3 (15.1)0.0362*
Uric acid (mg/dL)4.0 (1.1)6.3 (1.9)<0.0001*
Blood pressure (mmHg)
  Systolic118 (11)152 (12)<0.0001*
  Diastolic60 (8)91 (9)<0.0001*
Proteinuria (+)n.d.2 (1)<0.0001
Haemostatic data
Fibrinogen (g/L)5.10 (1.04)5.06 (0.87)0.8344*
t-PA (ng/mL)8.0 (6.3–10.9)14.4 (10.4–18.6)<0.0001
PAI-1 (U/mL)56.1 (16.9)67.2 (29.1)0.0410*
D-dimer (ng/mL)348 (245–502)484 (290–741)0.0204
Platelets (103/μL)177 (58)187 (62)0.4222*

Haemostatic data obtained from the longitudinal study of normal pregnant women are shown in Table 1. All the haemostatic substances increased throughout pregnancy, except platelet count, which decreased. In the third trimester, PAI-1 activity is three times that of the first trimester and the same finding occurred for D-dimer.

In pre-eclampsia, t-PA and D-dimer were not normally distributed and therefore the concentrations of those substances are presented as medians (25th to 75th centiles). The concentration of fibrinogen and the platelet count were the same in women with pre-eclampsia compared with the controls, but t-PA, PAI-1 and D-dimer were increased. The concentration of t-PA almost doubled in pre-eclampsia (P < 0.0001) (Table 2). In normal pregnancy we detected four moderate outliers in D-dimer (944, 992, 1113 and 1184 ng/mL). In pre-eclampsia three outliers were detected in D-dimer (>2000 ng/mL) and one moderate (36.3 ng/mL) and two extreme (45.6 and 75.0 ng/mL) outliers in t-PA. When outliers were removed from analysis, the pre-eclamptic group still had, compared with the normal one, significantly higher medians for D-dimer (465 vs 337 ng/mL; P= 0.0146) and t-PA (13.5 vs 8.0 ng/mL; P < 0.0001).

A significant positive correlation was found between fibrinogen and t-PA in normal pregnancy, but this was reversed in pre-eclampsia, and the same was true for D-dimer and t-PA (Table 3). Differences in those correlations were more pronounced when outliers were removed from the analysis (Figs 1 and 2, respectively). In pre-eclampsia, t-PA was correlated with the degree of proteinuria measured by dipstick (Fig. 3). The platelet count was inversely correlated with systolic blood pressure (r=−0.364, P= 0.027) (Fig. 4).

Table 3.  Correlation of t-PA with fibrinogen and D-dimer in the third trimester of gestation in normal pregnancy and pre-eclampsia.
t-PA FibrinogenD-dimer
  1. rs= Spearman's rank correlation coefficient.

Normal pregnancyrs0.3130.268
n5656
P0.0190.046
Pre-eclampsiars−0.202−0.139
n3737
P0.2320.413
Figure 1.

Correlation of fibrinogen and t-PA in the third trimester of gestation (outliers removed) in normal pregnancy (r= 0.387, n= 56, P= 0.003) and pre-eclampsia (r=−0.248, n= 34, P= 0.157). r= Pearson's correlation coefficient.

Figure 2.

Correlation of D-dimer and t-PA in the third trimester of gestation (outliers removed) in normal pregnancy (r= 0.281, n= 52, P= 0.044) and pre-eclampsia (r=−0.246, n= 32, P= 0.175). r= Pearson's correlation coefficient.

Figure 3.

Correlation of proteinuria (dipstick test) and t-PA in the third trimester of pre-eclampsia with (rs= 0.575, n= 37, P= 0.0002) and without (rs= 0.454, n= 34, P= 0.007) the three outliers of t-PA (white dots). rs= Spearman's rank correlation coefficient.

Figure 4.

Correlation of platelet count and systolic blood pressure (SBP) in the third trimester of pre-eclampsia (r=−0.364, n= 37, P= 0.027). r= Pearson's correlation coefficient.

Discussion

The longitudinal study showed large changes in the coagulation and fibrinolytic systems, which agree with other reports indicating that fibrinogen1, t-PA antigen9,16 and D-dimer levels8,16,17 as well as PAI-1 activity7 increase throughout normal pregnancy. Some authors claim that this physiologic adaptation may in part be explained by a hormonal influence16.

The rise in PAI-1 activity during pregnancy was higher than that observed for t-PA antigen. t-PA showed only a slight rise (although significant) when compared with the rise in PAI-1, which trebles its value throughout pregnancy. These relative changes in t-PA and PAI-1 may indicate inactivation of t-PA by PAI (formation of t-PA–PAI complexes) and may explain the reduced fibrinolytic potential during normal pregnancy reported elsewhere1. The significant rise in the levels of D-dimer during normal pregnancy indicates a compensated state of low grade intravascular coagulation. Moreover, it shows that, notwithstanding the marked impairment in fibrinolytic potential, the fibrinolytic system remains functionally active.

In the cross sectional part of the study, we compared the haemostatic parameters between women with pre-eclampsia and healthy women, matched for gestational age. When the data were analysed, we detected some outliers regarding D-dimer (both groups) and t-PA (pre-eclampsia). The presence of outliers in both groups may represent high variability in those substances between subjects. On the other hand, they may reflect an undiagnosed pathology in women presenting such high values. They may even represent a technical error during the analytic determination. As we cannot reject any of these hypothesis, we performed statistical analysis with and without outliers and realised that such values had no substantial effect on the statistical results obtained and on the subsequent conclusions.

Women with pre-eclampsia had significantly higher levels of t-PA, D-dimer and PAI-1. Some studies have already reported similar results in PAI-1 activity7,18–21, t-PA antigen18,22 and D-dimer23 levels. As we measured these substances together (unlike previous reports), we were also able to compare their relative changes.

t-PA and PAI-1 are both produced by endothelial cells and their greater levels in the third trimester in pre-eclampsia may suggest the development of endothelial dysfunction. The increased concentrations of D-dimer indicate increased intravascular coagulation and activation of fibrinolysis.

In pre-eclampsia, t-PA showed an inverse correlation with fibrinogen (Table 3; Fig. 1), suggesting that intravascular coagulation is occurring and that the fibrinolytic system is activated, consuming fibrinogen. We were, therefore, expecting to find a positive correlation of t-PA with D-dimer. However, we observed an inverse correlation (Table 3; Fig. 2), although without statistical significance. It seems there is increased production of t-PA by endothelial cells, the activator of plasminogen to degrade fibrinogen and fibrin. The overproduction of t-PA is, therefore, likely to be related to endothelial disturbance and could explain the positive correlation of t-PA with the degree of proteinuria (Fig. 3), a known marker of the severity of pre-eclampsia. We could hypothesise an overproduction of t-PA in pre-eclampsia by endothelial cells in order to accomplish some other functions. For instance, it has been reported that t-PA has a mitogenic role in smooth muscle cells, suggesting that its overproduction may contribute to vascular repair, although it may also favour atherogenesis24. Hypertriglyceridemia occurs in pre-eclampsia25, which may compromise endothelial cell function and thus may favour thrombosis and atherosclerosis.

No differences were found in platelet count between the pre-eclamptic and the control groups. Platelet count may be useful in predicting coagulation abnormalities in hypertensive disorders of pregnancy26. Women with pre-eclampsia have a greater decrease in platelet count in pregnancy than normal women, although no difference was found in the platelet count at the end of gestation27, in agreement with our results. In our study, there was an inverse correlation between platelet count and systolic blood pressure (Fig. 4), suggesting that in severe pre-eclampsia thrombocytopenia is likely.

In conclusion, we describe modifications in the haemostatic system during pregnancy. Our data suggest that platelet count (which correlated inversely with systolic blood pressure) and t-PA levels (which positively correlated with proteinuria) may be useful in defining the severity of pre-eclampsia. We also suggest that t-PA antigen may be a valuable biomarker for the assessment of endothelial dysfunction in pre-eclampsia. Further, larger studies are required to test this hypothesis.

Acknowledgements

The authors would like to thank the University of Porto and FCT (Portugal) for financial support (PRAXIS XXI/BD/16061/98).

Ancillary