OUP user menu

Microvascular permeability is related to circulating levels of tumour necrosis factor-α in pre-eclampsia

Nick Anim-Nyame , John Gamble , Suren R. Sooranna , Mark R. Johnson , Philip J. Steer
DOI: http://dx.doi.org/10.1016/S0008-6363(02)00844-1 162-169 First published online: 1 April 2003

Abstract

Introduction: The mechanism for the increased microvascular permeability which, underline many of the complications of pre-eclampsia, remain unexplained. It has been suggested that a factor present in the maternal circulation in pregnancies complicated by the disease may be responsible for increased microvascular permeability. In this study, we have investigated the relationship between filtration capacity (Kf), an index of microvascular permeability, and maternal levels of VEGF, leptin and TNF-α, all of which are known permeability factors whose plasma levels are increased in pre-eclampsia. Methods: We used a small cumulative pressure step venous congestion plethysmography protocol to compare Kf, an index of microvascular permeability, during the third trimester of 20 women with pre-eclampsia, 18 normal pregnant women and 18 non-pregnant female matched controls. Blood samples were obtained to measure plasma levels of VEGF, leptin, TNF-α plasma protein concentrations and full blood count. Results: Microvascular filtration capacity (Kf) was significantly increased in pre-eclampsia compared to the other groups (P≪0.0001, ANOVA). Kf was also increased in the normal pregnant group when compared to the non-pregnant controls (P = 0.02). Plasma levels of VEGF, leptin and TNF-α were significantly greater in pre-eclampsia compared to normal pregnancy and non-pregnant controls (P<0.0001, ANOVA, for all three analyses). Total plasma protein and albumin concentrations were significantly lower in the normal pregnant and pre-eclamptic groups, compared to the non-pregnant controls (P<0.0001, ANOVA). Kf was significantly related to TNF-α in pre-eclampsia (r = 0.53, P = 0.018), and with VEGF in the non-pregnant controls (r = 0.6, P = 0.02). No significant relationship was observed between Kf and VEGF, leptin and TNF-α during normal pregnancy. There was a significant inverse correlation between plasma albumin concentration and filtration capacity in the normal pregnant (r = −0.94, P<0.0001) and non-pregnant (r = −0.87, P<0.0001) groups but not in the women with pre-eclampsia (r = −0.18, P = 0.8). Conclusions: These data show that that microvascular filtration capacity is significantly increased in pre-eclampsia, and correlates with circulating levels of TNF-α but not leptin or VEGF.

Keywords
  • Growth factors
  • Microcirculation

Time for primary review 26 days.

1 Introduction

Pre-eclampsia is a multisystem disorder of the second half of pregnancy, characterized by generalized endothelial cell dysfunction [1]. It appears the release of pro-inflammatory cytokines such as TNF-α [2] and reactive oxygen species [3] from the ischaemic placenta, which results from abnormal placentation [4] in pre-eclampsia, contributes to the endothelial dysfunction. Cytokines may also contribute directly to oxidative stress induced endothelial dysfunction [5] or indirectly through neutrophil activation [6]. Reduced maternal plasma volume is a feature of pre-eclampsia [7] and appears to be secondary to increased microvascular permeability [8]. The mechanism(s) for the increased microvascular permeability, however, remains unexplained. Of the Starling forces, which influence microvascular fluid exchange, changes in oncotic [9] and hydrostatic pressure gradients [10] do not appear to adequately to explain the increased fluid flux observed in pregnancies complicated by the disease.

There is evidence that sera from women with pre-eclampsia rather that normal pregnancy increases the permeability of human umbilical vein endothelial cell (HUVEC) monolayers [11]. This suggests that there may be a circulating factor(s) present in pre-eclampsia that may be responsible for the increased microvascular permeability in pregnancies complicated by the disease. Plasma levels of vascular endothelial growth factor (VEGF) [12], leptin [13] and tumour necrosis factor-α (TNF-α) [14] are elevated in pre-eclampsia. These three substances have been shown to cause a marked increase in microvascular permeability in a variety of animal model studies. The role of VEGF in microvascular permeability in the non-pregnant state is well established [15]. Recent evidence suggests that leptin receptors are expressed by vascular endothelial cells [16]. Leptin induces angiogenesis and may increase microvascular permeability [17]. Structurally, leptin resembles the class I cytokines [18] and may regulate placental cytokine production during pregnancy [19]. The increased circulating levels of leptin during pregnancy appear to be of placental origin [20], although its functions in the placenta are not known, in addition to its angiogenic effect, it may increase the exchange of small molecules between maternal circulation and the fetus by the induction and maintenance of vascular permeability [21]. TNF-α is associated with increased capillary leak (for example in sepsis) and has been shown to increase venular fluid filtration coefficient (Lp) [22] in the non-pregnant state. In this study, we have investigated whether Kf correlates with plasma levels of VEGF, leptin and TNF-α, in pre-eclampsia.

2 Methods

2.1 Subjects

We used a small cumulative pressure step venous congestion plethysmography protocol to compare filtration capacity (Kf) during the third trimester of normal and pre-eclamptic pregnancies. Filtration capacity was measured in the calves of 20 women with pre-eclampsia, 18 normal pregnant women, and 18 non-pregnant controls. All the women were of similar in age and body mass index (BMI), and the pregnant women matched for gestational age. The pregnant women were all nulliparous white Europeans with spontaneous pregnancies and were recruited from the antenatal clinics or antenatal ward at the Chelsea and Westminster Hospital, London. The non-pregnant controls were health workers. The normal pregnant controls were women with no history of medical illnesses, attending the routine antenatal clinics and who were invited to take part in the study. They were chosen to be similar to the pre-eclamptic group with regard to the latter's booking body mass index (BMI) and, in the pregnant controls, gestational age. All the women were non-smokers and were not on any medication. None of the non-pregnant controls were taking oral contraceptives. None of the subjects received any intravenous infusion before or during the study. Women with previous or present history of peripheral vascular disease, peripheral neuropathy, chronic hypertension, infective or inflammatory disorders or any other underlying medical disorders were excluded from the study.

Pre-eclampsia was defined according to the criteria of hypertension and proteinuria occurring for the first time after 20 weeks gestation, and reversal of both after delivery. Hypertension was defined as an absolute blood pressure greater than 140 mmHg systolic or 90 mmHg diastolic taken twice, 6 h apart. The first and fifth Korotkoff sounds were used to determine the systolic and diastolic components, respectively. Proteinuria was defined as more than 0.5 g urinary protein excretion in a 24-h urine sample [23]. The urine specimens were collected into plastic jugs containing phenyl mercuric acetate as a preservative. Protein was measured by colorimetric reactions using an autoanalyser, as described by Watanabe et al. [24]. The obstetric records of all pregnant groups were reviewed after delivery to confirm reversal of hypertension and proteinuria. The investigation conformed to the principles outlined in the Declaration of Helsinki. The Local Ethics Committee approved the study and informed consent was obtained from each patient.

2.2 Study protocol

Studies were performed in a quiet room at a steady temperature (23–24°C). Subjects rested for at least 15 min before the study. Observations were made in the left lateral position, to prevent aorto-caval compression and with the right mid-calf supported at the level of the heart. In addition to the blood pressure assessment at diagnosis, arterial blood pressure was measured, non-invasively, on the ipsi-lateral calf and arm, using a Dinamap Vital Sign Monitor (Type 1800, Critikon, Tampa, FL, USA). Averaged values of systolic, diastolic and mean arterial blood pressures were calculated from triplicate measurements. Filtration capacity (Kf), an index of vascular permeability, was estimated using the Filtrass strain gauge plethysmograph (Filtrass, DOMED, Munich, Germany) [25]. The device, a modification of a standard strain gauge plethysmograph, has been fully described previously [26]. Briefly, the congestion pressure cuff, which is attached to a compressor pump built into the apparatus, was placed around the right thigh and enclosed in a rigid corset to reduce filling volume and thus filling time. Changes in calf circumference in response to a rapid increase in cuff pressure (pcuff), were measured using a passive inductive transducer with an accuracy of ±5 μm. The files were recorded and saved for subsequent ‘off-line’ analysis.

2.3 Assessment of filtration capacity

Microvascular filtration capacity (Kf) was measured using an established small cumulative pressure step strain gauge plethysmography protocol [25,26]. Briefly, a series of five to seven small (8 mmHg) cumulative pressure steps were applied to the venous congestion cuff and the resulting changes in tissue volume, derived from alterations in calf circumference (Fig. 1a), were recorded. Using this protocol, no change in the recorded signal is observed until the ambient venous pressure in the limb is exceeded. At congestion cuff pressures greater than this value, each additional pressure increment causes a change in limb volume that is attributed to vascular filling (Va) (Fig. 1a). The volume change representing Va can be fitted by exponential analysis and it takes 30–45 s to reach a steady-state [26]. When congestion cuff pressure exceeds the isovolumetric venous pressure (pvi); the equilibrium pressure at the microvascular interface reflecting the transmicrovascular balance of Starling forces, a steady-state change in volume is observed, reflecting fluid filtration (Jv) (Fig. 1a).

Fig. 1

(a) Calf volume response (top panel) to a step increase in congestion cuff pressure (pcuff). The upper trace of the top panel reflects the whole volume response and the superimposed dotted line, the regression slope (Jv). The lower trace of the top panel shows the volume response after the regression slope (Jv) has been subtracted. The volume (Va) reflects the compliance volume change in response to this pressure step. (b). The relationship between Jv and pcuff (top panel) and the vascular compliance (lower panel) of the subject in (a).

Interpretation of these data is, perhaps, made easier by referring to the Starling equation itself. Movement of fluid and plasma proteins between the vascular and interstitial compartments is governed by the Starling forces, which are described in Eq. (1): Embedded Image(1) where Kf is the fluid filtration capacity; pc and pt are the hydrostatic pressures in the capillaries and the tissue, respectively; α is the osmotic reflection coefficient, an index of the vascular permeability to plasma proteins; and τc and τt are plasma and tissue oncotic pressures, respectively. The filtration capacity Kf reflects the product of the area available for fluid filtration and the permeability per unit surface area. The isovolumetric venous pressure (pvi) is an index of the equilibrium pressure at the microvascular interface, reflecting the local plasma oncotic pressure απ (Fig. 1b).

Computer-based analysis enables differentiation between volume and filtration responses [27]. The value of Kf is determined by linear regression, based on the pcuff and Jv, co-ordinates obtained at pressures above those causing measurable increase in Jv (see Fig. 1b). The slope of this relationship is Kf (Fig. 1b) and the units are expressed as KfU (ml min−1 100 ml−1 mmHg−1×10−3) [26]. The intercept of the line on the abscissa, that is where Jv, represents pvi (Fig. 1a), the pressure that has to be exceeded to induce net filtration at the level of the strain gauge.

2.4 Sample collection and analysis

The venous blood samples used to assay for circulating levels of total VEGF, leptin, TNF-α, plasma albumin, total protein, uric acid and creatinine concentrations and also full blood count, were obtained from the antecubital vein. Care was taken in sample handling to avoid platelet activation, since this is known to increase VEGF levels in vitro. The heparinized samples were centrifuged for 10 min at 3000 rpm and 4°C, the plasma separated within an hour of venepuncture and stored at −70°C until assayed. Total immunoreactive VEGF was assayed by quantitative sandwich enzyme immunoassay technique, using the Quantikine human VEGF commercial kit (R&D Systems Europe limited, Abingdon, Oxon, UK). The assays were done in duplicate, using an enzyme-linked polyclonal antibody specific to VEGF. Total VEGF (pg/ml) was measured because of the effects of pregnancy on its free form. Factors produced by the placenta results in underestimation of the free forms of VEGF [28]. Total circulating leptin concentrations (ng/ml) were measured in duplicate by radioimmunoassay as described previously [29]. The assay employed a polyclonal (rabbit) antibody raised against recombinant human leptin. Standards and 125I-tracer were also made from recombinant human leptin [29]. The average intra- and inter-assay coefficients of variation were 3 and 8%, respectively. TNF-α was assayed using a solid phase sandwich commercial ELISA kit (Diaclone Research, France) with inter- and intra-assay coefficients of variation of 1.7 and 6.0%, respectively. Full blood count, plasma albumin, total protein, uric acid and creatinine concentrations were also measured, using an autoanalyser.

2.5 Statistical analysis

All the data were normally distributed data and are presented as mean±S.D. Statistical differences between the groups were compared using analysis of variance with Bonferroni correction for multiple comparisons. The relationships between Kf and the clinical parameters, VEGF, leptin and TNF-α concentrations were determined using Pearson correlation coefficients. Multiple regression analysis was performed to determine which of the parameters were independently related to Kf and to assess associations between Kf and measured clinical and biochemical variables. Statistical significance was assumed at a P value less than 0.05. The Statistical Package for Social Sciences (SPSS, version 10) was used for these analyses.

3 Results

The clinical and demographic characteristics for the three groups are shown in Table 1. There were no significant differences in age or BMI among the three groups, or in gestational age between the pregnant groups. Babies born to women with pre-eclampsia were smaller compared to the normal pregnant controls (P = 0.02). As expected from the recruitment criteria, the women with pre-eclampsia had higher systolic and diastolic blood pressures (P = 0.001), higher serum uric acid levels (P = 0.001) and lower platelet counts (P = 0.002) than the pregnant control group.

View this table:
Table 1

Clinical and demographic characteristics of the three groups

VariablesNon-pregnantNormal pregnantPre-eclampsiaP value
(n = 18)(n = 18)(n = 20)
Age (years)29.9±4.229.0±5.129.6±4.4NS
Gestation (weeks)N/A35.6±1.635.8±2.0NS
BMI (kg/m2)22.7±2.623.4±3.324.6±4.0NS
SBP (mmHg)113±6.0119.4±10.9141.3±13.30.001
DBP (mmHg)67±6.272.3±6.993.0±4.70.001
Birth weight (kg)N/A3.44±0.152.95±0.120.02
Plasma uric acid
(mmol/l)0.24±0.040.22±0.050.39±0.030.001
Plasma albumin (g/l)38.8±1.126.2±2.624.0±1.6NS
Total protein (g/l)74.3±0.969.8±4.164.19±3.4NS
Platelets (×106)257.9±21.5230.9±42.5138.8±32.70.002

Microvascular filtration capacity was significantly increased in pre-eclampsia (6.32±0.3 KfU), compared to the normal pregnant and non-pregnant controls (4.35±0.18 and 3.13±0.16 KfU, respectively, P≪0.0001, ANOVA with Bonferroni correction). Kf was also increased in the normal pregnant group when compared to the non-pregnant controls (P = 0.02) (Fig. 2). Plasma levels of VEGF, leptin and TNF-α were significantly greater in pre-eclampsia (149.79±7.9 pg/ml, 29.59±3.04 μg/l and 191.02±12.37±3.89 μg/l, respectively) compared to normal pregnancy (53.03±4.9 pg/ml, 12.62±0.97 μg/l and 59.75±1.59 μg/l, respectively) and non-pregnant controls (38.5±5.65 pg/ml, 10.05 μg/l and 34.89±4.0 μg/l, respectively) (P<0.0001, ANOVA, for all three factors) (Fig. 3a–c). Total plasma protein and albumin concentrations were significantly lower in the normal pregnant and pre-eclamptic groups, compared to the non-pregnant controls (P<0.0001, ANOVA) (Table 1). Although plasma protein levels were lower in pre-eclampsia than normal pregnancy, the differences were not statistically significant (P = 0.05) (Fig. 3d).

Fig. 3

Graphs demonstrating the relationship between microvascular filtration capacity and circulating levels of (a) TNF-α (μg/l), (b) VEGF (pg/l), (c) leptin (pg/l), and (d) plasma albumin (g/l) in pre-eclampsia (●), normal pregnancy (○), and non-pregnant controls (▴).

Fig. 2

Dot plots comparing microvascular filtration capacity in women with pre-eclampsia (n = 20), normal pregnant controls (n = 18) and non-pregnant control women (n = 18). The dots represent Kf values for women in each group and the bars mean values.

Kf was significantly related to TNF-α in pre-eclampsia (Fig. 4b, r = 0.53, P = 0.018), and with VEGF in the non-pregnant controls (Fig. 3a, r = 0.6, P = 0.02). No significant relationship was observed between Kf and VEGF, leptin and TNF-α during normal pregnancy. There were highly significant inverse correlations between plasma albumin concentration and filtration capacity in the normal pregnant (r = −0.94, P<0.0001) and non-pregnant (r = −0.87, P<0.0001) groups, but not in the women with pre-eclampsia (r = −0.18, P = 0.8) (Fig. 3d). A significant inverse relationship was also observed between centile birth weight and filtration capacity in pre-eclampsia (r = −0.67, P = 0.001) but not in normal pregnancy (r = 0.07, P = 0.77) (Fig. 4). Plasma uric acid concentrations correlated significantly with Kf, and plasma levels of VEGF, leptin and TNF-α in pre-eclampsia in pre-eclampsia (r = 0.64, 0.68 and 0.71; respectively, P = 0.001).

Fig. 4

This graph demonstrates the relationship between birth weight (kg) and microvascular filtration capacity (KfU) in pre-eclampsia (●), normal pregnancy (○).

4 Discussion

In this study we compared microvascular permeability in normal pregnancy, pregnancies complicated by pre-eclampsia, and in the non-pregnant state. Whereas microvascular permeability increased during normal pregnancy and pre-eclampsia, the increase was significantly greater in pre-eclampsia. Furthermore, we observed that changes in microvascular permeability in pre-eclampsia correlated significantly with circulating levels of TNF-α but not with either VEGF or leptin. No significant relationship was observed between permeability and these permeability-inducing factors in the normal pregnant controls. This is the first study to compare changes in microvascular permeability with those of specific circulating permeability factors in pregnancy. However, while suggestive, these data do not prove a causal relationship between microvascular permeability and TNF-α in pre-eclampsia.

Although Haller et al. [11], reported that a circulating factor(s) present in maternal circulation in pre-eclampsia increased endothelial cell permeability, no specific permeability factors were identified. The relationship between microvascular permeability and circulating TNF-α observed in the present study is consistent with the hypothesis that pre-eclampsia may be an exaggeration of inflammatory responses common to all pregnancies [30]. It is also consistent with the role of TNF-α in inducing capillary leak by increased fluid filtration coefficient (Lp) [22] in the non-pregnant state, particularly in sepsis. The placenta expresses TNF-α receptors, the expression increases with gestational age [31], and is known to be up regulated in pre-eclampsia [32]. It appears that the increased circulating levels, during pregnancy, may be in response to trophoblast–decidual cell interactions, and may thus play a role in trophoblast differentiation and invasion of decidua and spiral arteries [33,34].

Although, this study did not investigate the mechanisms by which TNF-α may increase microvascular permeability in pre-eclampsia several are possible. TNF-α has been shown to up-regulate the expression of the cell adhesion molecules VCAM-1 and ICAM-1 on the endothelial cells [35]. Furthermore, circulating levels of TNF-α correlate with VCAM-1 expression [36]. Thus, the increase in VCAM-1 and ICAM-1 expression may increase leucocyte–endothelial cell interactions and, in addition, increase microvascular permeability. TNF-α may also increase endothelial permeability by inducing oxidative stress in pre-eclampsia [37]. This is because TNF-α has been shown to interfere with mitochondria electron flow with resultant release of oxidizing free radicals, leading to lipid peroxidation [38].

VEGF is known to be one of the most potent microvascular permeability [15] inducing agents and circulating levels are elevated in pre-eclampsia [12]. Thus, the lack of significant correlation between the level of VEGF and microvascular permeability is an interesting observation. Hypoxia is known to be a very potent stimulus for upregulation of VEGF production [39] and we have previously shown that tissue blood flow is reduced in pre-eclampsia [40] and precedes the clinical onset of the disease [41]. Although, the increase in VEGF might have been related to the reduced tissue blood flow, we did not observe any correlation between circulating levels of VEGF and Kf. There are other possible explanations for the apparent lack of correlation between VEGF and Kf in the pre-eclamptic group. In animal studies, Bates and co-workers [15,42], investigated differences in vascular permeability following acute and chronic exposure to VEGF. They observed that, whereas the greatest permeability increases followed acute exposure to VEGF, the effects were transient. By contrast, chronic exposure resulted in more sustained increases in permeability. Pre-eclampsia presents a chronic exposure state, and may therefore explain the findings in this study. It is also possible that the mechanism(s) by which VEGF exerts its action had been maximally up regulated. If this is a correct interpretation of these data, it must be assumed that the further increases in permeability in response to TNF-α, were operating via another pathway. Beynon et al. [43] have demonstrated that cytokines have synergistic effects on permeability of endothelial cell monolayers. Therefore the diverse array of agonists or growth factors released in a complex pathophysiological state such as pre-eclampsia, may give rise to responses via synergistic mechanisms. A significant correlation was observed between VEGF and Kf in the non-pregnant controls. Whilst the plasma VEGF levels were much lower, it is possible that we were observing the effect of VEGF on permeability, without the synergistic action of the other factors normally raised in pre-eclampsia. Haller et al. [11] demonstrated that the circulating factor(s) present in maternal circulation in pre-eclampsia which induced increased microvascular permeability was mediated by the protein kinase C (PKC) pathway. In their study, incubation of pre-eclamptic sera induced translocation of α and ε isoforms of PKC within endothelial cells. This is consistent with recent evidence that VEGF increased microvascular permeability in vivo by phospholipase C (PLC) stimulation but not by the PKC pathway [44], observations that are in line with our own results.

Thus it seems that VEGF is unlikely to be the only circulating factor responsible for the increased microvascular permeability in pre-eclampsia. It is also possible that other factors present in maternal circulation in pre-eclamptics inhibit the vascular permeability effects of VEGF. Angiopoietins, for instance, are vascular endothelial cell-specific growth factors that play an important role principally during the last stages of angiogenesis after the induction of new capillaries by VEGF. There is evidence that over-expression of angiopoietin-1 (Ang1) results in non-leaky vessels by inhibiting the effects of VEGF [45]. Whereas combined overexpression of VEGF and Ang1 had an additive effect on new vessel formation, this combination results in leak resistant-vessels typical of Ang-1 [46,47]. Furthermore, Ang1 suppresses VEGF-induced expression of the cell adhesion molecules VCAM-1, ICAM-1 and E-selectin, thus inhibiting endothelial adhesiveness [48]. It is therefore possible that the effect of increased VEGF on vascular permeability is suppressed by the presence of other circulating factor(s), such as Ang1, in pregnancies complicated by pre-eclampsia. Further work is required to investigate circulating levels of angiopoietins in normal pregnancy and pre-eclampsia and their relationships with VEGF.

Although leptin has both angiogenic and vascular permeability effects [21], circulating levels did not correlate with Kf in all three groups of women in this study. Since the increased level of leptin during pregnancy is predominantly placental in origin, it is possible it acts locally to increase placental exchange between the mother and fetus. Such effects would be consistent with previous evidence that placental expression of leptin mRNA is upregulated during hypoxic conditions including pre-eclampsia [49] and diabetes [50]. The lack of correlation between Kf and leptin could also be explained by synergism between the different circulating permeability factors in all three groups of women. Another observation worth further attention is the striking correlation between plasma albumin concentration and permeability in the pregnant and non-pregnant controls. Since albumin is one of the major contributors to plasma oncotic pressure, the major expected effect of changes in concentration would be in terms of the value of σπ, not Kf (Fig. 1b). That the pre-eclamptics do not also show this, despite their high permeability values, probably reflects the intervention of other causative agents, as discussed above. Increased microvascular permeability is associated with reduced plasma volume, which is a characteristic feature of the disease [7]. Thus, the inverse correlation between filtration capacity and birth weight is consistent with previous observations of inverse relationship between plasma volume and birth weight [51].

In summary, we have demonstrated that the increased microvascular permeability in pre-eclampsia significantly correlates with increased circulating levels of TNF-α but not VEGF and leptin. However, the data do not prove a causal relationship between Kf and TNF-α.

Acknowledgements

This work was supported by a grant from The Henry Smith Charity.

References

  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
View Abstract