Molecular Insights into the Thrombotic and Microvascular Injury in Placental Endothelium of Women with Mild or Severe COVID-19

Clinical manifestations of coronavirus disease 2019 (COVID-19) in pregnant women are diverse, and little is known of the impact of the disease on placental physiology. Severe acute respiratory syndrome coronavirus (SARS-CoV-2) has been detected in the human placenta, and its binding receptor ACE2 is present in a variety of placental cells, including endothelium. Here, we analyze the impact of COVID-19 in placental endothelium, studying by immunofluorescence the expression of von Willebrand factor (vWf), claudin-5, and vascular endothelial (VE) cadherin in the decidua and chorionic villi of placentas from women with mild and severe COVID-19 in comparison to healthy controls. Our results indicate that: (1) vWf expression increases in the endothelium of decidua and chorionic villi of placentas derived from women with COVID-19, being higher in severe cases; (2) Claudin-5 and VE-cadherin expression decrease in the decidua and chorionic villus of placentas from women with severe COVID-19 but not in those with mild disease. Placental histological analysis reveals thrombosis, infarcts, and vascular wall remodeling, confirming the deleterious effect of COVID-19 on placental vessels. Together, these results suggest that placentas from women with COVID-19 have a condition of leaky endothelium and thrombosis, which is sensitive to disease severity.


Introduction
Coronavirus infections have ranged from asymptomatic to mild, moderate or severe symptoms in pregnant women. In the Middle East respiratory syndrome coronavirus (MERS-CoV) pandemic, which originated on the Arabian Peninsula in 2012, the pregnancy outcomes remained apparently unaffected [1]. However, during the severe acute respiratory syndrome coronavirus (SARS-CoV) epidemic of [2002][2003] in China, pregnant women had spontaneous miscarriages in the first trimester, delivered preterm, or had pregnancies complicated by intrauterine growth restriction [2]. With the emergence of the novel SARS-CoV-2 and the coronavirus disease 2019 (COVID- 19), clinical manifestations of pregnant women with COVID-19 have varied widely from asymptomatic to very severe, and pregnancy complications have included miscarriages [3], fetal distress, premature rupture of membranes, preterm labor, stillbirth [4] and uncontrolled hemorrhage during cesarean delivery [5]. According to the Centers for Disease Control and Prevention in the USA, a higher proportion (12.9%) of preterm live births among women with SARS-CoV-2 infection during pregnancy is present in comparison to the general population (10.2%) [6].
Vertical transmission of SARS-CoV-2 has been suggested, due to positive RT-PCR nasopharyngeal swabs in newborns delivered by cesarean section with neonatal isolation implemented immediately after birth [7][8][9]. Some infants born to mothers with confirmed COVID-19 have displayed problems, including fetal distress, lethargy, vomiting, fever, respiratory distress, thrombocytopenia accompanied by abnormal liver function and even death [9][10][11]. Such symptoms have been displayed in newborns with a SARS-CoV-2 positive and negative RT-PCR test (for review, see [12]), and even with mothers having COVID-19, but with negative RT-PCR tests in amniotic fluid, vaginal secretions, placenta, and breast milk [13].
With the COVID-19 pandemic, interest has arisen concerning the impact of the disease on placental physiology. In pregnant women with COVID-19, placental swabs have been positive for SARS-CoV-2 RNA by RT-PCR in cases where the newborns tested either positive [14] or negative [15,16] for SARS-CoV-2, as in a second-trimester miscarriage with a SARS-CoV-2-negative fetus [3]. The presence of this coronavirus in the placental syncytiotrophoblast (STB) cell layer has been further demonstrated by immunostaining with antibodies against SARS-CoV-2 proteins [17,18] or by RNA in situ hybridization of SARS-CoV-2 spike antigen [19]. Moreover, SARS-CoV-2 virus has been observed invading the human placenta using electron microscopy [20]. However, this observation has been questioned, suggesting that the structures identified are clathrin-coated vesicles and not SARS-CoV-2 virus particles [21]. SARS-CoV-2 cell entry depends on binding of the spike protein organized in trimers to receptor angiotensin-converting enzyme 2 (ACE2) [22]. This receptor is critical, since sequestering ACE2 inside cells due to the loss of Rab7A, a key regulator of endo-lysosomal trafficking, reduces viral entry [23]. SARS-CoV-2 spike protein harbors a multibasic site S1/S2 that undergoes proteolytic cleavage by host proteases like furin, TMPRSS2 and Cathepsin L, which allow the posterior fusion of the viral membrane with a cellular membrane in the endocytic pathway and the release of viral RNA in the cytoplasm of the host cell [22,24].
ACE2 is abundantly present in the lung and small intestine epithelia, as in arterial and venous endothelial cells in all organs [25]. In the human placenta, ACE2 is present in the stromal and perivascular cells of decidua [26], fetal placental vessels [27], and in cytotrophoblast [26]. In the STB cell layer, ACE2 is also present and has been proposed to promote maternal vasodilation through Ang 1-7 release into the maternal circulation [27].
Von Willebrand Factor (vWf) plays a critical role in hemostasis as it binds and stabilizes factor VIII in the circulation and mediates platelet-endothelial and platelet-platelet interaction at high shear (for review, see [37]). vWf is assembled as a multimeric protein in endothelial Weibel Palade bodies and is exocyted in response to several stimuli, including inflammatory cytokines. The ultra-large multimers of vWf tether circulating platelets to damaged endothelial sites under high shear stress conditions. In normal circumstances, the metalloproteinase ADAMTS13 cleaves vWf into smaller and less thrombogenic units (for review, see [37]). In thrombotic microangiopathy, like that observed in purpura and other thrombocytopenic conditions like severe sepsis, disseminated intravascular coagulation, and complicated malarial infections, an excess of vWf with a deficiency of ADAMTS13 is observed [38]. Likewise, in patients with severe COVID-19, plasma levels of vWf antigen are increased [39][40][41][42][43][44], while ADAMTS13 activity is normal [42] or diminished [39,40].
Vascular endothelial (VE) cadherin and claudin-5 are adherens and tight junctions (AJ, TJ) proteins, respectively, involved in endothelial cell-cell adhesion and barrier function. VE-cadherin is a classical cadherin present at the adherens junction of endothelial cells, required to prevent the disassembly of blood vessels [45]. Deletion of the VE-cadherin gene leads to early embryonic death associated with severe vascular anomalies [46] and endothelial apoptosis [47]. Likewise, antibodies against VE-cadherin ectodomain block endothelial cell-cell adhesion and increase vascular paracellular permeability [48].
Claudins are major constituents of TJ and are responsible for the ionic selectivity of the paracellular pathway (for review, see [49]). Claudin-5 is specifically present in endothelial cells, and when transfected in L fibroblasts that lack TJs, forms strands that resemble those of endothelial cells where the extracellular, and not the protoplasmic, face of the membrane associates to TJ filaments [50]. In the brain, claudin-5 is required for the establishment of the blood-brain barrier against small molecules (<800 D) [51].
Considering this, we aimed to analyze the impact of COVID-19 in placental endothelium. For this purpose, we have studied the expression of vWf, claudin-5, and vascular endothelial (VE) cadherin in the decidua and chorionic villi of placentas derived from women with mild and severe COVID-19. Here, we found that in the endothelium of decidua and chorionic villi of placentas derived from women with COVID-19, the expression of vWf is increased, being higher in severe cases, suggesting the existence of a thrombotic condition. The altered state of the endothelium in the decidua and chorionic villi of placentas from women with COVID-19 is further confirmed by a decreased expression of both claudin-5 and VE-cadherin in the placentas of women with severe COVID-19, suggesting enhanced vessel permeability. The histological analysis of these placentas revealed thrombosis, infarcts, and remodeling of vascular walls in chorionic villi and decidua, indicating fetal and maternal malperfusion.

Ethics Statement
The study protocol followed the Declaration of Helsinki Ethical Principles for Medical Research Involving Human Subjects. The participants signed the informed consent before their inclusion in this work. The study was approved by the Ethics and Research Internal Review Board of the Instituto Nacional de Perinatología in Mexico City (2020-1-32).

Patient Selection and Specimens
Universal testing with nasopharyngeal swabs and RT-PCR test (La Charité, Berlin protocol) to detect SARS-CoV-2 infection was implemented at the Instituto Nacional de Perinatología for all women who were admitted for delivery, even if asymptomatic; for all positive cases, newborns were tested for SARS-CoV-2 infection using saliva samples.
Placental tissues were obtained immediately after delivery from 11 women with COVID-19 (five mild and six severe) and four control women who delivered by cesarean section, with no evidence of labor. The RT-PCR diagnosis was done during the acute phase of SARS-CoV-2 infection upon admission to delivery. Healthy controls were paired by gestational age. Clinical data and outcomes of enrolled women were obtained from the electronic medical records.

RT-PCR for Placental SARS-CoV-2 Infection
For detection of SARS-CoV-2 RNA in the placenta, the tissue was disrupted through mechanical lysis using a FastPrep instrument (MP Biomedicals, Solon, OH, USA), within 1 h after obtention. Then, RNA was purified using the Direct-zol RNA Miniprep Kit (Cat. R2050; Zymo Research, Irvine, CA, USA). SARS-CoV-2 RNA was detected following La Charité, Berlin protocol [52], using TaqPath 1 step RT-PCR master Mix, CG commercial kit (Cat. A15299 Thermo Fisher Scientific, Waltham, MA, USA), and probes and primers designed for RdRP and E viral genes. RNase P human gene was used as RNA isolation control. RT-qPCR was performed on a StepOnePlus instrument (Applied Biosystems/Thermo Fisher Scientific, Waltham, MA, USA). Each RT-PCR reaction contained an enzyme mix, primers, probes, and RNA (5 µL each one) [52]. Conditions at the thermocycler were set as previously reported [52]. Ct values were collected using threshold at 0.035 fluorescence level.

Histochemical Staining and Hofbauer Cell Assessment
Placental tissues were fixed with 10% para-formaldehyde and then embedded in paraffin. Sections of 3 µm were cut and stained with hematoxylin and eosin according to standard protocols [54]. Hofbauer cells were identified in these tissues by immunohistochemical analysis, using a rat monoclonal antibody against CD68 (Cat. ab53444, Abcam, San Diego, CA, USA; dilution 1:1000), counterstained with hematoxylin. The number of Hofbauer cells present in the tissue was determined using the analysis of software Zen (version ZEN 2.3 lite, Carl Zeiss Microscopy, Jena, Germany). Fields were selected at 20×, and CD68+ macrophage count was performed at 40× in five different high-power optical fields per placenta.

Statistical Analysis
The three tests of normality of D'Agostino (Skewness, Kurtosis and Omnibus), and the equal variances test of Levene and Bartlett were employed to confirm the assumptions about residuals in One-Way ANOVA. The F test of ANOVA for equal variances and the F test with Welch correction for unequal variances were used, followed by the multiple comparison tests of Bonferroni, Dunnett, or Duncan. The Kruskal-Wallis test and its multiple comparison test were used for not normally distributed data. The legend of each figure indicates the detailed statistical analysis employed. Data are expressed as mean ± SD, and statistical significance was considered for p < 0.05.

Results
Here, we studied the expression of vWf, claudin-5, and VE-cadherin in the decidua and chorionic villi of placentas derived from control and SARS-CoV-2 infected women. The clinical data and outcomes of control women and those with mild and severe COVID-19 are summarized in Tables 1-3, respectively.

The Expression of Claudin-5 Diminishes in the Endothelium of Decidua and Chorionic Villi of Placentas from Women with Severe COVID-19
Next, we analyzed the expression of claudin-5 in the endothelium of the decidua and chorionic villi of control women and those with mild or severe COVID- 19   Next, we analyzed the expression of claudin-5 in the endothelium of the decidua and chorionic villi of control women and those with mild or severe COVID-19. By immunofluorescence, we did not observe significant changes in the expression of claudin-5 in decidua (Figure 3a,b) and chorionic villi (Figure 4a,b) of women with mild COVID-19 in comparison to control. However, in placentas of women with the severe form of the disease, we found a significant decrease in claudin-5 expression in decidua (Figure 3a,b) and chorionic villi (Figure 4a,b). Since claudin-5 is the main claudin of endothelial cells [50], these results strongly suggest that TJs in the decidual endothelium and chorionic villi of women with COVID-19 become leaky as the severity of the disease augments.
fluorescence, we did not observe significant changes in the expression of claudin-5 in decidua (Figure 3a,b) and chorionic villi (Figure 4a,b) of women with mild COVID-19 in comparison to control. However, in placentas of women with the severe form of the disease, we found a significant decrease in claudin-5 expression in decidua (Figure 3a,b) and chorionic villi (Figure 4a,b). Since claudin-5 is the main claudin of endothelial cells [50], these results strongly suggest that TJs in the decidual endothelium and chorionic villi of women with COVID-19 become leaky as the severity of the disease augments.

VE-Cadherin Expression Diminishes in Decidua and Chorionic Villi Endothelium of Placentas from Women with Severe COVID-19
Since VE-cadherin is a crucial protein of the AJs of endothelial cells [45] whose alteration leads to an increase in vascular permeability [48], we next analyzed the expression of this protein in the endothelium of the decidua and chorionic villi. By immunofluorescence, we detected no significant change in VE-cadherin expression in the decidua of

VE-Cadherin Expression Diminishes in Decidua and Chorionic Villi Endothelium of Placentas from Women with Severe COVID-19
Since VE-cadherin is a crucial protein of the AJs of endothelial cells [45] whose alteration leads to an increase in vascular permeability [48], we next analyzed the expression of this protein in the endothelium of the decidua and chorionic villi. By immunofluorescence, we detected no significant change in VE-cadherin expression in the decidua of women with mild COVID-19 compared to control women. In contrast, VE-cadherin expression in the decidua is significantly reduced in women with severe COVID-19 (Figure 5a,b). A similar pattern is observed in the chorionic microvilli, where the decreased expression of VE-cadherin is detected in women with severe COVID-19 but not in those with mild disease (Figure 6a,b). These results reinforce the observations done with claudin-5, suggesting that COVID-19 augments the paracellular permeability of the endothelium in the chorionic villi.
women with mild COVID-19 compared to control women. In contrast, VE-cadherin expression in the decidua is significantly reduced in women with severe COVID-19 ( Figure  5a,b). A similar pattern is observed in the chorionic microvilli, where the decreased expression of VE-cadherin is detected in women with severe COVID-19 but not in those with mild disease (Figure 6a,b). These results reinforce the observations done with claudin-5, suggesting that COVID-19 augments the paracellular permeability of the endothelium in the chorionic villi.

Placentas of Women with COVID-19 Display Histological Alterations Indicative of Vasculopathy and a Higher Number of Hofbauer Cells is Observed in Placentas from Women with Severe COVID-19
Since thrombotic and microvascular injury syndrome has been observed in patients with COVID-19 (for review, see [35]), we next analyzed if histological alterations indicative of fetal vascular malperfusion (FVM)

Placentas of Women with COVID-19 Display Histological Alterations Indicative of Vasculopathy and a Higher Number of Hofbauer Cells Is Observed in Placentas from Women with Severe COVID-19
Since thrombotic and microvascular injury syndrome has been observed in patients with COVID-19 (for review, see [35]), we next analyzed if histological alterations indicative of fetal vascular malperfusion (FVM) or maternal vascular malperfusion (MVM) were present in the placentas of women with COVID-19. The histological analysis revealed the presence in chorionic villi of subacute thrombosis with remodeling of vascular wall, and extensive parenchymal infarcts with intervillositis. In the decidua of women with COVID-19, we found vasculopathy with remodeling of the vascular wall. Instead, placentas from control women had no histological alterations (Figure 7). In addition, to confirm the inflammatory state of placentas of women with COVID-19, we assessed the number of fetal macrophages, known as Hofbauer cells, in the parenchyma of chorionic villi. Figure 8 shows an increase in Hofbauer cells in chorionic villi of women with severe COVID-19, compared to control placentas and tissues from women with the mild form of the disease (** p < 0.01).
Cells 2021, 10, x FOR PEER REVIEW 14 of 21 present in the placentas of women with COVID-19. The histological analysis revealed the presence in chorionic villi of subacute thrombosis with remodeling of vascular wall, and extensive parenchymal infarcts with intervillositis. In the decidua of women with COVID-19, we found vasculopathy with remodeling of the vascular wall. Instead, placentas from control women had no histological alterations (Figure 7). In addition, to confirm the inflammatory state of placentas of women with COVID-19, we assessed the number of fetal macrophages, known as Hofbauer cells, in the parenchyma of chorionic villi. Figure 8 shows an increase in Hofbauer cells in chorionic villi of women with severe COVID-19, compared to control placentas and tissues from women with the mild form of the disease (** p < 0.01).

Discussion
COVID-19 induces, in some patients, a thrombotic and microvascular injury syndrome triggered by several mechanisms, including a cytokine storm, hypoxic vaso-occlusion, direct activation of immune and vascular cells by virus infection, and the development of pathogenic autoantibodies targeting phospholipids and phospholipid-binding proteins [55] (for review, see [35]).
In patients with severe COVID-19, plasma levels of vWf antigen are increased [39][40][41][42][43][44], and in the placenta of healthy women, vWf has been found in endothelium, STB and chorionic villous stroma [56]. In placentas from women with COVID-19 we found that this factor, whose increase may predict an augmented risk of thrombosis [57], is elevated compared to controls in the endothelium of both decidua and chorionic villi, especially in severe cases. This observation suggests that COVID-19 represents a thrombotic risk in human placenta.
Changes in vWf expression have been explored in other pathological conditions of pregnancy. Thus, in pregnancies with intrauterine growth restriction that occurs when a fetus does not reach its growth potential, the expression of vWf is higher, although not at a statistically significant level [58]. In preeclampsia, a higher amount of vWf is found in maternal plasma [56,59], with no differences in the expression in the chorionic villous endothelium and stroma in comparison to normal pregnancies [56]. Others instead reported a decrease in placental vWf in the STB accompanied by an increase in the intervillous space in preeclampsia, thus suggesting injury to the STB cell layer that favors the release of vWf from Weibel-Palade bodies into the maternal space [60].
The cell-cell adhesion complex of endothelial cells constituted by AJ and TJs is critical to prevent vascular leakage and for proper placental perfusion. To test if COVID-19 in-

Discussion
COVID-19 induces, in some patients, a thrombotic and microvascular injury syndrome triggered by several mechanisms, including a cytokine storm, hypoxic vaso-occlusion, direct activation of immune and vascular cells by virus infection, and the development of pathogenic autoantibodies targeting phospholipids and phospholipid-binding proteins [55] (for review, see [35]).
In patients with severe COVID-19, plasma levels of vWf antigen are increased [39][40][41][42][43][44], and in the placenta of healthy women, vWf has been found in endothelium, STB and chorionic villous stroma [56]. In placentas from women with COVID-19 we found that this factor, whose increase may predict an augmented risk of thrombosis [57], is elevated compared to controls in the endothelium of both decidua and chorionic villi, especially in severe cases. This observation suggests that COVID-19 represents a thrombotic risk in human placenta.
Changes in vWf expression have been explored in other pathological conditions of pregnancy. Thus, in pregnancies with intrauterine growth restriction that occurs when a fetus does not reach its growth potential, the expression of vWf is higher, although not at a statistically significant level [58]. In preeclampsia, a higher amount of vWf is found in maternal plasma [56,59], with no differences in the expression in the chorionic villous endothelium and stroma in comparison to normal pregnancies [56]. Others instead reported a decrease in placental vWf in the STB accompanied by an increase in the intervillous space in preeclampsia, thus suggesting injury to the STB cell layer that favors the release of vWf from Weibel-Palade bodies into the maternal space [60].
The cell-cell adhesion complex of endothelial cells constituted by AJ and TJs is critical to prevent vascular leakage and for proper placental perfusion. To test if COVID-19 induced damage to placental vessels, we analyzed the expression of VE-cadherin and claudin-5, the main molecular components of endothelial AJ [45][46][47][48] and TJs, respectively. We found that while there is no significant change in expression of VE-cadherin and claudin-5 in the decidua and chorionic villi of placentas from women with mild COVID-19, the amount of both adhesion molecules decreases significantly in the tissues derived from women with severe COVID-19. Hence, these results indicate that AJ and TJ proteins are sensitive to damage induced by COVID-19, and further reveal that the deleterious effect detected on the apical junctional complex of placental endothelium correlates with COVID-19 severity and the development of inflammation determined by the increased number of Hofbauer cells present in the parenchyma of chorionic villi.
In human placenta, VE-cadherin was previously observed in endothelial cells of the decidua [64] and the chorionic villus [65,66], as well as in the STB, where the expression decreases towards term in normal pregnancies but not in those complicated by preeclampsia [67]. However, in preeclampsia, VE-cadherin expression in chorionic villi endothelial cells was not altered [66]. Diabetes mellitus is another disease linked to endothelial dysfunction (for reviews, see [68,69]), and in insulin-treated gestational diabetes (GDM), a reduction in VE-cadherin expression in chorionic vessels was found [70]. In contrast, placentas of women with diabetes type 1, type 2, and GDM display normal levels of VE-cadherin in the chorionic vessels if not treated with insulin [71].
Previous studies have shown that the amount of claudins -1 and -5 present in Triton X-100 insoluble fractions, which corresponds to claudin associated with the actomyosin cytoskeleton, diminishes in preeclamptic compared to healthy placentas [61], whereas the expression of claudin-4 at the basolateral membrane of the STB diminishes in placentas derived from ZIKV-infected women [62].
Our histological analysis of COVID-19 placentas revealed the presence of thrombosis, infarcts, and vascular wall remodeling in chorionic villi and decidua. These findings are in agreement with previous observations done in placentas from women with COVID-19, showing that the chorionic terminal villi have features of fetal vascular malperfusion including infarctions [14,72], avascular villi [72][73][74], fibrin deposition in fetal vessels [72], intervillous hemorrhages [16] and thrombi [72,73], capillary congestion and focal microchorangiosis [16], a vascular hyperplastic process observed in placental tissue after periods of low-grade hypoxia [75]. Maternal vascular malperfusion has also been reported with decidual arteriophathy, including maternal vessels with artherosis and fibrinoid necrosis as well as hypertrophy of membrane arterioles [72].
We also observed a higher number of Hofbauer cells in placentas from women with severe COVID-19. Similar observations have been done in placental pathologies involving infection and inflammation [76], and, in the case of placentas of women with COVID-19, other studies have reported intervillositis with inflammatory infiltrate of macrophages, T lymphocytes, neutrophils, and monocytes [3,[17][18][19].
A preeclampsia-like syndrome has been described in pregnant women with COVID-19, with indistinguishable clinical symptoms and similar histopathological findings [77]. This could be due to the pyroptosis induced by SARS-CoV-2 replication and release [78]. In early preeclampsia, pyroptosis is known to induce the release of alarmins and placental debris into maternal circulation [79], which triggers thrombosis, intramural fibrin deposition, villous stromal-vascular karyorrhexis, and villous infarction [72][73][74]. Altogether, these changes lead to placental dysfunction and fetal growth restriction [80]. In this context, our findings pose a challenge to the differential diagnosis of the hypertensive disease of pregnancy, highlighting the importance of considering the molecular profile of the disease and not only the associated signs and symptoms.
Although placental vasculopathy associated with COVID-19 has been reported previously, it was described in placentas from a heterogeneous population, including women with associated comorbidities (e.g., severe preeclampsia, infection, placenta previa, preterm labor, or diabetes), and vaginal delivery or cesarean section [81][82][83][84]. Instead, our study is more homogeneous, including only women in the acute phase of infection, without associated comorbidities, and having babies delivered by cesarean section. This approach allows us to assume that the placental histological findings and the molecular changes that correlate with these histological observations and with the severity of the disease are due to COVID-19 infection. On the other hand, the limitations of our study are that we were not able to measure circulating vWf in the studied women and that severity markers, such as fibrinogen, D-dimer, and procalcitonin as well as X-rays and CAT, were determined only in women with severe COVID-19.
In conclusion, our study reveals that placentas from women with COVID-19 display a thrombotic and microvascular injury syndrome, including the overexpression of vWf in endothelium coupled with the decreased expression of VE-cadherin and claudin-5 in chorionic villus and decidua that correlate with disease severity. The presence of thrombosis, infarcts, and vascular wall remodeling in chorionic villi and decidua further confirm the deleterious effect of COVID-19 on placental vessels.