Background
COVID-19 presents a spectrum of signs and symptoms in pregnant women that might resemble preeclampsia. Differentiation between severe COVID-19 and preeclampsia is difficult in some cases.
Objective
To study biomarkers of endothelial damage, coagulation, innate immune response, and angiogenesis in preeclampsia and COVID-19 in pregnancy in addition to in vitro alterations in endothelial cells exposed to sera from pregnant women with preeclampsia and COVID-19.
Study Design
Plasma and sera samples were obtained from pregnant women with COVID-19 infection classified into mild (n=10) or severe (n=9) and from women with normotensive pregnancies as controls (n=10) and patients with preeclampsia (n=13). A panel of plasmatic biomarkers was assessed, including vascular cell adhesion molecule-1, soluble tumor necrosis factor-receptor I, heparan sulfate, von Willebrand factor antigen (activity and multimeric pattern), α2-antiplasmin, C5b9, neutrophil extracellular traps, placental growth factor, soluble fms-like tyrosine kinase-1, and angiopoietin 2. In addition, microvascular endothelial cells were exposed to patients’ sera, and changes in the cell expression of intercellular adhesion molecule 1 on cell membranes and von Willebrand factor release to the extracellular matrix were evaluated through immunofluorescence. Changes in inflammation cell signaling pathways were also assessed by of p38 mitogen-activated protein kinase phosphorylation. Statistical analysis included univariate and multivariate methods.
Results
Biomarker profiles of patients with mild COVID-19 were similar to those of controls. Both preeclampsia and severe COVID-19 showed significant alterations in most circulating biomarkers with distinctive profiles. Whereas severe COVID-19 exhibited higher concentrations of vascular cell adhesion molecule-1, soluble tumor necrosis factor-α receptor I, heparan sulfate, von Willebrand factor antigen, and neutrophil extracellular traps, with a significant reduction of placental growth factor compared with controls, preeclampsia presented a marked increase in vascular cell adhesion molecule-1 and soluble tumor necrosis factor-α receptor I (significantly increased compared with controls and patients with severe COVID-19), with a striking reduction in von Willebrand factor antigen, von Willebrand factor activity, and α2-antiplasmin. As expected, reduced placental growth factor, increased soluble fms-like tyrosine kinase-1 and angiopoietin 2, and a very high soluble fms-like tyrosine kinase-1 to placental growth factor ratio were also observed in preeclampsia. In addition, a significant increase in C5b9 and neutrophil extracellular traps was also detected in preeclampsia compared with controls. Principal component analysis demonstrated a clear separation between patients with preeclampsia and the other groups (first and second components explained 42.2% and 13.5% of the variance), mainly differentiated by variables related to von Willebrand factor, soluble tumor necrosis factor-receptor I, heparan sulfate, and soluble fms-like tyrosine kinase-1. Von Willebrand factor multimeric analysis revealed the absence of von Willebrand factor high-molecular-weight multimers in preeclampsia (similar profile to von Willebrand disease type 2A), whereas in healthy pregnancies and COVID-19 patients, von Willebrand factor multimeric pattern was normal.
Sera from both preeclampsia and severe COVID-19 patients induced an overexpression of intercellular adhesion molecule 1 and von Willebrand factor in endothelial cells in culture compared with controls. However, the effect of preeclampsia was less pronounced than the that of severe COVID-19. Immunoblots of lysates from endothelial cells exposed to mild and severe COVID-19 and preeclampsia sera showed an increase in p38 mitogen-activated protein kinase phosphorylation. Patients with severe COVID-19 and preeclampsia were statistically different from controls, suggesting that both severe COVID-19 and preeclampsia sera can activate inflammatory signaling pathways.
Conclusion
Although similar in in vitro endothelial dysfunction, preeclampsia and severe COVID-19 exhibit distinctive profiles of circulating biomarkers related to endothelial damage, coagulopathy, and angiogenic imbalance that could aid in the differential diagnosis of these entities.
Introduction
Preeclampsia is a pregnancy complication and a leading cause of maternal and perinatal morbimortality and iatrogenic prematurity. Although its etiology is not completely understood, , it is accepted that this condition relies on placental insufficiency and maternal cardiovascular maladaptation underlined by angiogenic imbalance, endothelial dysfunction, coagulopathy, and complement dysregulation, which lead clinically to hypertension and proteinuria that can progress to multiorgan dysfunction during pregnancy. The multifactorial nature of preeclampsia explains a variable clinical/laboratory presentation, mainly determined by gestational age at onset (early vs late).
Why was this study conducted?
We conducted this study to characterize the profile of endothelial damage, coagulation, innate immune response, and angiogenesis in preeclampsia and COVID-19 in pregnancy, which are both considered disorders associated with endothelial dysfunction.
Key findings
Severe COVID-19 in pregnancy and preeclampsia share a similar end-stage in vitro-induced p38 mitogen-activated protein kinase phosphorylation in endothelial cells but a differential profile of circulating endothelial and angiogenic biomarkers. Severe COVID-19 is characterized by higher vascular cell adhesion molecule-1 (VCAM-1), soluble tumor necrosis factor-α receptor I (sTNFRI), heparan sulfate (HS), von Willebrand factor (VWF) antigen, and neutrophil extracellular traps (NETs) and reduced placental growth factor (PlGF), whereas preeclampsia is marked by increased VCAM-1, sTNFRI, soluble fms-like tyrosine kinase-1 (sFlt-1), angiopoietin-2, C5b9, and NETs and a reduction in VWF antigen, VWF activity, α2-antiplasmin, and PlGF.
What does this add to what is known?
Soluble biomarkers of coagulopathy (VWF), endothelial inflammation (sTNFRI), barrier damage (HS), and angiogenesis (sFlt-1) seem to be highly specific in differentiating preeclampsia from severe COVID-19 in pregnancy. These findings improve our understanding of the pathophysiological pathways in preeclampsia and COVID-19 and may help in the differential diagnosis of these disorders during pregnancy.
Clinical and analytical data from patients infected by SARS-CoV-2, which causes COVID-19, suggest that endothelial dysfunction plays an important role in the pathophysiology of this condition, involving extrapulmonary manifestations of COVID-19 like hypertension, kidney disease, thrombocytopenia, and liver injury. Some of these clinical features overlap with those observed in preeclampsia. In addition, an increased incidence of preeclampsia has been reported in association with COVID-19. Despite their clinical resemblance, the mechanisms underlying endothelial dysfunction might differ between COVID-19 and preeclampsia. Understanding endothelial and angiogenic profiles could enlighten the pathophysiological basis of these 2 entities.
The endothelium is a monolayer of cells that lines the interior of blood vessels, acting as a protective layer between circulating blood and other tissues. The endothelium is crucial for the regulation of vascular homeostasis, coagulation cascade, immune response, and angiogenesis. Circulating biomarkers related to endothelial activation and loss of barrier integrity seem to be associated with disease severity in COVID-19. , Inflammatory effects on these damaged endothelial cells activate the innate immune response and induce a hypercoagulable state with impaired fibrinolysis and angiogenic imbalance. , On the contrary, angiogenesis dysregulation has emerged as 1 of the main pathophysiological features in the development of preeclampsia. , Finally, in vitro studies enabled us to describe the endothelial cell proinflammatory and thrombogenic response in COVID-19.
The aim of the present study was to comprehensively investigate the endothelial and angiogenic profiles in preeclampsia and SARS-CoV-2 infection in pregnancy using circulating biomarkers and in vitro studies.
Materials and Methods
Study populations and design
Pregnant women with laboratory-confirmed SARS-CoV-2 infection were selected from a large multicenter prospective population-based cohort study conducted from March 15 to May 31, 2020, in Barcelona, Spain, including consecutive cases detected during the study period. SARS-CoV-2 infection was confirmed by a positive real-time polymerase chain reaction (RT-PCR) on nasopharyngeal swab or a positive serologic result. SARS-CoV-2–positive pregnancies were subdivided into mild (n=9) and severe disease (n=8) according to the presence of pneumonia or coexistence of fever, dry cough, and dyspnea. In addition, we also included SARS-CoV-2–negative pregnant women, including preeclampsia (n=13) and normotensive pregnancies as controls (n=10) who were matched to COVID-19 cases by gestational age at blood sampling. Preeclampsia was defined as high blood pressure (systolic blood pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg on 2 occasions, at least 4 hours apart) developed after 20 weeks of gestation with proteinuria (≥300 mg/24 h or protein/creatinine ratio ≥0.3), thrombocytopenia (platelet count <100 × 10 9 /L), renal insufficiency (serum creatinine concentrations >1.1 mg/dL), impaired liver function (elevated blood concentrations of liver transaminases to twice-normal concentration), pulmonary edema, or a new-onset headache unresponsive to medication and not accounted for by alternative diagnoses or visual symptoms. Early-onset preeclampsia was defined by gestational age at delivery <34 weeks. Baseline and perinatal data were obtained by interviews and from electronic medical records. Gestational age was calculated on the basis of the crown-rump length at first-trimester ultrasound. Birthweight centiles were assigned according to local standards. Pregnancies with chromosomal/structural anomalies or intrauterine infection were excluded. Endothelial and angiogenic profiles were studied in all participants by analyzing circulating molecules in maternal peripheral blood and by in vitro study of endothelial cells exposed to patients’ sera. Details of the laboratory methodology used are included in the Supplemental Material .
This study was approved by the ethics committee of the Hospital Clínic de Barcelona (HCB/2020/0401) and conformed to the ethical guidelines of the Declaration of Helsinki. All participants provided informed written consent before sample collection.
Maternal blood sample collection
Peripheral maternal blood was obtained by venipuncture within 24 to 48 hours after onset of symptoms and before starting any treatment. Plasma and sera samples were obtained by centrifugation of blood anticoagulated with ethylenediaminetetraacetic acid and by incubation for 30 minutes at room temperature to allow clotting, and subsequently centrifuged at 1500× g for 10 minutes at 4°C to separate the serum from clots, respectively. All samples were aliquoted and stored at −80°C until used.
Assessment of circulating biomarkers
Endothelial damage was assessed by measuring plasmatic concentrations of vascular cell adhesion molecule-1 (VCAM-1), soluble tumor necrosis factor-α receptor I (sTNFRI), and heparan sulfate (HS) with enzyme-linked immunosorbent assays (ELISA) (R&D Systems, Minneapolis, MN; Biomatik Corporation, Wilmington, DE; and AttendBio Research, Barcelona, Spain, respectively). The kit used for the detection of HS did not show any significant cross-reactivity or interference between HS and analogs according to the manufactureŕs instructions.
Biomarkers for coagulation/fibrinolysis included von Willebrand factor (VWF) antigen (VWF:Ag) and activity (VWF:GPIbM) and α2-antiplasmin (α2AP) evaluated by immunoturbidimetry (Atellica 180 360 COAG, Siemens Healthineers, Erlangen, Germany). Visualization of VWF multimers was achieved using a commercially available enhanced chemiluminescence kit for detecting horseradish peroxidase-labeled antibodies on Western blots. In addition, VWF-cleaving protease (ADAMTS13) activity was assessed by fluorescence resonance energy transfer (Fluoroskan Ascent FL, Thermo Fisher Scientific, Waltham, MA). Plasminogen activator inhibitor antigen (PAI) and thrombomodulin (TM) were measured by ELISA (Imubind, BioMedica Diagnostics, Windsor, Canada and Biomatik Corporation, respectively).
Activation of innate immune response was determined by circulating terminal complement complex (C5b9) and double-stranded DNA (dsDNA) for neutrophil extracellular traps (NETs) quantified by Quant-iT PicoGreen dsDNA Assay Kit (Invitrogen, Thermo Fisher Scientific) on a fluorescence reader.
Angiogenic profile was assessed by sera concentrations of free placental growth factor (PlGF) and soluble fms-like tyrosine kinase-1 (sFlt-1) with ELISA (R&D Systems Europe Ltd, Abingdon, United Kingdom) and angiopoietin-2 (Ang2) ELISAs (R&D Systems, Minneapolis, MN). The sFlt-1 to PlGF ratio was calculated as previously described.
In vitro studies
For the in vitro studies, human dermal microvascular endothelial cells (ATCC, CRL-3243, Lot: 62630587) in culture were exposed to patients’ sera to study cell response to: (1) the expression of adhesion receptors at the cell surface (intercellular adhesion molecule 1 [ICAM-1]) as an indicator of a proinflammatory cell response; (2) the presence of the adhesive protein VWF, involved in thrombogenicity, on the extracellular matrix generated by these cells; and (3) the activation of the endothelial intracellular signaling pathway related to inflammation—p38 mitogen-activated protein kinase (p38 MAPK). Details of the laboratory methodology used are included in the Supplemental Material .
Statistical analysis
Baseline and perinatal data were analyzed with the statistical software STATA 14.2 (StataCorp LLC, College Station, TX) and results are expressed as median and interquartile range or percentage as appropriate. Statistical analysis comprised the comparison of each group of complicated pregnancies with controls. Soluble markers are expressed as median (interquartile range). Further statistical analyses were performed in R version 4.0.0 (R Foundation for Statistical Computing, Vienna, Austria) using the Student t test with the Benjamini-Hochberg correction for multiple comparisons after checking data normality and homoscedasticity. Results were considered statistically significant when adjusted P value was <.05. Data were ordinated and plotted using principal component analysis. An additional unsupervised hierarchical clustering was performed on the basis of the univariate results comparing severe COVID-19 with preeclampsia. A subanalysis comparing early- with late-onset preeclampsia was performed using the Student t test and Benjamini-Hochberg procedure for multiple pairwise comparisons and included in the Supplemental Material .
Results
Baseline and perinatal characteristics of the study populations
Baseline characteristics of the study populations are summarized in the Table . Study groups were mainly similar in terms of maternal and perinatal characteristics. However, patients with preeclampsia had higher rates of Asian ethnicity and a tendency to younger age. Chronic hypertension was present in 2 patients with preeclampsia and systemic lupus erythematosus in 1 control. None of the patients included in this study had pregestational diabetes mellitus or previous respiratory disorders. All the pregnancies complicated by preeclampsia were proteinuric, 4 were early-onset cases that were treated with corticosteroids for fetal lung maturity, and 5 patients had preeclampsia with severe features that was treated with magnesium sulfate. Patients with preeclampsia showed an earlier gestational age at delivery, with a trend toward higher rates of small-for-gestational-age fetuses and admissions to the neonatal intensive care unit. Three cases of preeclampsia were complicated by peripartum hemorrhage. Severe COVID-19 cases were all detected by RT-PCR. Among the mild cases, 2 were detected by RT-PCR and the rest by positive serology. Given that this study was conducted at the beginning of the pandemic, convalescent subjects should have been infected during the 4 weeks preceding the blood analysis. Two cases of mild COVID-19 had hypertension, and 1 of them had associated proteinuria. None of the COVID-19 patients (mild or severe) had thrombocytopenia, elevated liver enzymes, or elevated creatinine. All COVID-19 cases were followed up to 40 days postpartum to exclude the diagnosis of evolving preeclampsia. The diagnosis of atypical preeclampsia in COVID-19 cases was excluded because none of them presented signs of placental insufficiency nor abnormal sFlt-1 to PlGF ratio (according to our institutional protocol for the differential diagnosis of hypertensive disorders in pregnancy). Severe COVID-19 cases were not critically ill (no mortality and only 1 case required invasive mechanical ventilation). Six patients with severe COVID-19 were treated with low-molecular-weight heparin, 3 of them were additionally treated with hydroxychloroquine and azithromycin, and 1 of these 3 was also given lopinavir/ritonavir and corticosteroids. As mentioned earlier, maternal blood samples were obtained before starting any treatment. Gestational age at sampling was similar between the study groups at a median (interquartile range) of 40.2 (38.9–41) weeks in controls, 39.1 (38.7–39.6) weeks in mild COVID-19 cases, 39.3 (34.9–41.1) weeks in severe COVID-19 cases, and 39.1 (35.1–39.6) weeks in preeclampsia cases. No cases of perinatal mortality were observed in the study population.
| Controls n=10 | Mild COVID-19 n=9 | Severe COVID-19 n=8 | Preeclampsia n=13 | |
|---|---|---|---|---|
| Maternal characteristics | ||||
| Age (y) | 36.9 (31.6–38.7) | 36 (30.6–37.7) | 35.2 (24.7–39.1) | 29 (26–35.9) |
| Ethnicity | ||||
| White | 8 (80) | 5 (55.6) | 4 (50) | 4 (30.8) |
| African | 0 (0) | 0 (0) | 1 (12.5) | 2 (15.4) |
| Latin | 2 (20) | 3 (33.3) | 2 (25) | 2 (15.4) |
| Asian | 0 (0) | 1 (11.1) | 1 (12.5) | 5 (38.5) a |
| Pregestational body mass index (kg/m 2 ) | 22.4 (21.1–25.6) | 22.7 (20.3–28.7) | 21.8 (21–23.9) | 25.9 (21.9–28.4) |
| Nulliparity | 7 (70) | 5 (55.6) | 2 (25) | 7 (53.8) |
| Use of assisted reproductive technologies | 2 (20) | 0 (0) | 0 (0) | 0 (0) |
| Smoking during pregnancy | 0 (0) | 0 (0) | 1 (12.5) | 0 (0) |
| Perinatal outcomes | ||||
| Gestational age at delivery (weeks) | 40.2 (38.9–41) | 39.1 (38.7–39.6) | 39.2 (38.3–41.1) | 39.1 (35.1–39.6) a |
| Preterm delivery b | 1 (10) | 1 (11.1) | 2 (25) | 4 (30.8) |
| Cesarean delivery | 1 (10) | 3 (33.3) | 2 (25) | 5 (38.5) |
| Female gender | 4 (40) | 4 (44.4) | 5 (62.5) | 6 (46.1) |
| Birthweight (g) | 2975 (2780–3220) | 3280 (2940–3335) | 3290 (2780–3670) | 2558 (2010–3268) |
| Small-for-gestational age c | 3 (30) | 0 (0) | 0 (0) | 7 (53.8) |
| APGAR score at 5 min <7 | 0 (0) | 0 (0) | 1 (12.5) | 1 (7.7) |
| Umbilical artery pH | 7.21 (7.15–7.23) | 7.18 (7.12–7.21) | 7.17 (7.12–7.2) | 7.22 (7.17–7.24) |
| Admission to neonatal intensive care unit | 1 (10) | 0 (0) | 1 (12.5) | 5 (38.5) |
a P <.05 by Mann–Whitney U, Pearson chi square, or Fisher exact tests as appropriate, compared with controls
b Preterm delivery defined as delivery occurring before 37 weeks of gestation
c Small-for-gestational-age defined as birthweight below the 10th centile according to local standards.
Endothelial and angiogenic circulating biomarkers are differentially altered in COVID-19 vs preeclampsia
Results on soluble biomarkers in the study populations are displayed in Figure 1 and Supplemental Table 1 . Most soluble biomarkers were similar in mild COVID-19 and controls with the exception of a significant increase in VWF:Ag. In contrast, profound alterations in endothelial, coagulation, immune, and angiogenic biomarkers were detected in severe COVID-19 including significantly higher concentrations of VCAM-1, sTNFRI, HS, VWF:Ag, and NETs, with a significant reduction of PlGF compared with controls. No differences were observed in Ang2, sFlt-1, C5b9, ADAMTS13, PAI, nor TM between patients with severe COVID-19 and controls. Pregnant women with preeclampsia exhibited remarkable alterations in soluble biomarkers in a distinct profile from the one observed in COVID-19. Cases of preeclampsia showed a marked increase in VCAM-1 and sTNFRI (significantly increased compared with controls and severe COVID-19 cases), with a striking reduction in VWF:Ag, VWF:GPIbM, VWF:Ag/VWF:GPIbM, and α2AP. As expected, reduced PlGF, increased sFlt-1 and Ang2, and a very high sFlt-1 to PlGF ratio were also observed in preeclampsia. In addition, a significant increase in C5b9 and NETs was also detected in preeclampsia compared with controls. HS, ADAMTS13, PAI, and TM remained unchanged in preeclampsia.
Principal component analysis demonstrated a clear separation between preeclampsia and the other study populations (controls and mild and severe COVID-19 cases) ( Figure 2 , A). The first and second components explained 42.2% and 13.5% of the variance between groups. Unsupervised hierarchical clustering also showed a complete separation between severe COVID-19 cases and preeclampsia ( Figure 2 , B), with the most remarkable differences observed in VWF:GPIbM, VWF:Ag, and VWF:Ag/VWF:GPIbM followed by HS (significantly lower in preeclampsia) and sTNFRI, sFlt-1, and sFlt-1 to PlGF ratio (significantly higher in preeclampsia).
VWF multimeric analysis revealed the absence of VWF high-molecular-weight multimers in preeclampsia, comparable to a diagnosis of von Willebrand disease type 2A, with an accumulation of low-molecular-weight multimers ( Figure 3 ). In healthy pregnancies and SARS-CoV-2–positive patients, VWF multimeric pattern was normal.
