Placenta in SARS-CoV-2 infection: a new target for inflammatory and thrombotic events

Infection by SARS-CoV-2 has been shown to involve a wide range of organs and tissues, leading to a kaleidoscope of clinical conditions. Within this spectrum, an increased rate of preterm deliveries has been reported in women with COVID-19. High expression of proteins (ACE2/TMPRSS2) required for SARS- CoV-2 cell entry has been observed in the maternal-fetal tissues, supporting the possibility of placental viral involvement. A consecutive series of 6 placentas, delivered by 5 women with COVID-19 have been investigated. Three out of six placentas showed changes consistent with chronic villitis, while in one case chronic histiocytic intervillositis was diagnosed. Vascular abnormalities consisting of thrombo- hemorrhagic alterations were identified in the three cases with chronic villitis. These pathological findings: i) provide a basis to explain the higher rate of obstetric complications in these patients;2) suggest the need to investigate heparin use in COVID-19 affected pregnant patients to prevent adverse outcomes.


Background
On March 12, 2020, due do the rapid escalation of the coronavirus disease 2019 (COVID-19) outbreak, the WHO declares a pandemic.1 COVID-19 is caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), a betacoronavirus similar to SARS-CoV and MERS-CoV, with multiple possible transmission routes and characterized by a high infectivity.2, 3 Emerging data on maternal outcomes of COVID-19 suggest that clinical course is similar to non-pregnant patients4, 5, 6 and outcomes of perinatal COVID-19 infection are reassuring.7 However, adverse effects on newborns, such as fetal and respiratory distress, thrombocytopenia accompanied by abnormal liver function and even death have been reported.8 In addition, evidence of an higher incidence of preterm birth is being increasingly reported.9, 10 Concern is also rising on the possibility that SARS-CoV-2 may be vertically transmitted,11, 12, 13 but most patient series do not report cases of proved vertical transmission.9, 10, 14 To date, reported data regarding placental histopathology in COVID-19 are extremely limited (four cases in total),15, 16 but high expression of proteins (ACE2/TMPRSS2) required for SARS-CoV-2 cell entry has been observed in maternal-fetal interface tissues enabling the possibility of placental viral involvement.17 Moreover, the increased thrombotic risk and the in ammatory activation described in patients with COVID-1918, 19, 20 make the placenta a potential target of pathophysiological phenomena which could affect pregnancy outcomes and support the preliminary evidence of increased preterm birth risk.
For all these reasons, histological analysis of placentas delivered by women with COVID-19 could add insightful information. The aim of this work was to evaluate the consecutive series of placentas delivered at our Institution from women with COVID-19 infection, investigating the occurrence of pathological changes and the potential correlations with obstetric outcomes.

Case Series
Case 1 -Pathological ndings Placenta weight was 590 grams before formalin xation, lateral dimensions were 16 cm x 15 cm with thickness ranging from 2 cm to 3 cm. Umbilical cord length was 45 cm, its structure normal and a peripheral insertion was observed. Placental macroscopic features were normal with translucent membranes. Histological examination showed only mild abnormalities including a scant lymphocytic in ammation which involved both the decidua and basal placental villi with initial villous agglutination (Figure 2A-2B). Lymphocytes showed a prevalent T cell cytotoxic (CD8-positive) phenotype ( Figure 2C-2D). These ndings were consistent with chronic villitis (low grade, multifocal since in ammatory foci were present on more than one slide).21 Focal intervillous hemorrhage was also identi ed, while maturation was normal with features suggesting initial hypermaturation.

Case 2 -Pathological ndings
Placenta weighted 429 grams before xation, dimensions were 15.5 cm x 15 cm x 2 to 3.1 cm thickness. Umbilical cord (length: 16 cm) and membranes showed a normal morphology. After sectioning, focal and small whitish areas were observed close to the chorial surface and multiple, small (largest dimension <1 cm) hemorrhagic areas were also identi ed. Microscopically, this case showed areas of intervillous thrombosis, hemorrhage and brin laminar deposition ( Figure 3A) as well as microvascular thrombosis with intimal proliferations (Figure 3B). Signi cant foci of lymphocytic in ammation (CD8-positive) were present, involving the decidua and multiple villi ( Figure 3C-3D), consistent with chronic villitis (low grade, multifocal).21 Villous maturation was consistent with the gestational age.

Case 3 -Pathological ndings
Placenta weight was 460 grams prior to formalin xation, dimensions: 14 cm x 14 cm x 2 to 3 cm thickness. Umbilical cord morphology was normal and insertion central, length: 28 cm. Membranes were translucent. Sectioning showed focal whitish areas (largest dimension: 8 mm). Histological features included foci of villous/intervillous lymphocytic in ammation involving both the maternal and the fetal side as well the decidua in which this in ltrate was associated with signs of thrombo-hemorrhagic vasculopathy and partial trophoblast loss ( Figure 4A-4B). Prevalence of CD8-positive cells was observed ( Figure 4C). Similarly to the previous cases, these ndings were consistent with a chronic villitis (low grade, multifocal).21 Villous infarction areas were also present ( Figure 4D) as well as avascular villi consistent with fetal thrombotic vasculopathy. Villous maturation was within normal range with just focal chorangiosis.

Case 4 -Pathological ndings
The rst placenta weighted 410 grams after formalin xation, dimensions were 16 cm x 14 cm with a thickness ranging 1 to 3 cm. Umbilical cord length was 16 cm with a normal morphology and central insertion. Membranes presented a greenish discoloration and had a normal insertion. Sectioning showed extensive hemorrhagic subchorial areas consistent with partial placental abruption ( Figure 5A). Histological examination con rmed the presence of hemorrhagic areas located below the fetal surface ( Figure 5B), associated with chronic hypoxia signs (intervillous necrotic areas with focal calci cations) ( Figure 5C). Accelerated villous maturation was also focally present. Decidual morphology was normal.
The second placenta weight was 235 grams after xation and dimensions were 14 cm x 15 cm with thickness ranging between 0.5 and 2.5 cm. Umbilical cord insertion was paracentral, and length was 20 cm. Membranes showed a greenish discoloration similarly to the rst placenta. Microscopically, features consistent with accelerated villous maturation were noted. No other abnormalities were present.
Case 5 -Pathological ndings Placenta weight was 370 grams after xation, dimensions: 15 cm x 12 cm with a thickness ranging between 1 to 3.2 cm. Umbilical cord length was 18 cm with a paracentral insertion. Membranes appearance was normal. After sectioning, focal abruption was observed ( Figure 6A). Histological examination showed intervillous histiocytic in ltration (CD68-positive) with perivillous brin deposition ( Figure 6B-6C), a nding consistent with chronic histiocytic intervillositis. Areas showing villous conglutination, loss of trophoblasts and focal microcalci cations were also observed ( Figure 6D). Focal features of accelerated villous maturation were also present.

Discussion
In this manuscript, we report the pathological ndings observed in six consecutive placentas delivered from ve women affected by COVID-19. Differently from the limited data reported so far,15 we describe for the rst time pathological abnormalities potentially related to SARS-CoV-2 infection which could help understand the incoming reports suggesting higher rates of obstetric complications in these patients.
In 4 cases, we found microscopic signs of in ammation involving the decidua and the chorial villi: three out of four cases showed a CD8-positive T cell lymphocytic in ltrate, while in the remaining case an intervillous histiocytic in ltrate was identi ed.
The presence of an in ammatory lymphocytic in ltrate (cytotoxic CD8+ T cell lymphocytes in particular, as observed in our series) and of villous damage are consistent with the diagnosis of chronic villitis.22, 23 Of note, none of these three patients had comorbidities that could justify the presence of placental in ammation. Chronic villitis etiology is often related to an undergoing infection, although in most cases it is not possible to identify the speci c etiopathogenetic agent: these cases are de ned as "villitis of unknown etiology" (VUE) and can be found in 2-33.8% of placentas.24 Within the group of chronic villitis secondary to maternal infections, viruses are the most frequently involved pathogens and many speci c agents, such as varicella zoster, herpes simplex virus, cytomegalovirus, and in uenza A/H1N1, have been found to be associated with this condition. 25,26,27,28 Little is known about the association between placental pathology and coronavirus infections. Seven placentas from coronavirus-infected women during the 2002-2004 SARS outbreak have been previously described:29 histological examination found no abnormalities in 2/7, an increase in subchorionic and intervillous brin in 3/7, and massive thrombotic vasculopathy associated with intra-uterine growth restriction (IUGR) in the remaining 2/7 cases. No signs of chronic villitis were detected in any case and the same is true for MERS infections.30 Based on present knowledge, infection by SARS-CoV-2 is characterized by an extensive immune response with activation of both CD4+ and CD8+ lymphocytes18, 31, 32, thus ndings consistent with chronic villitis could be expected in the presence of placental involvement, also considering the high expression of ACE2 in decidual cells, villous cytotrophoblast and syncytiotrophoblast cells.17 Moreover, experiences derived from in uenza A/H1N1 infections in pregnancy suggest that activation of cellular immune response with release of in ammatory cytokines can lead to indirect placental damage. 28 Considering this evidence, the placenta could be potentially harmed because of intrinsic active virus replication and/or through indirect activity of in ammatory cytokines.33 Viral in situ assessment (by immunohistochemistry for viral proteins and/or in situ hybridization of viral RNA) will be important to de ne which speci c mechanisms are involved in this setting.
In the fourth case with in ammatory abnormalities, features consistent with chronic histiocytic intervillositis were identi ed. This rare entity is usually associated with maternal immunologic conditions like systemic lupus erythematosus, lupus anticoagulant and antiphospholipid syndrome.34, 35 Despite the different type of in ammatory in ltrate (histiocytic versus lymphocytic) and of etiologies (chronic histiocytic intervillositis is usually not related to viral infections), chronic histiocytic villositis is frequently associated with chronic villitis (30% to 50% of cases), supporting a potential overlap in terms of involved immunologic pathways.36, 37 Moreover, alveolar histiocytic in ltration seems to be a frequent autoptic hallmark of SARS-CoV-2 infection. 38 Lastly, it has to be noted that this patient's signi cant comorbidities (chronic hypertension, severe obesity, and gestational diabetes) are not usually associated with chronic histiocytic intervillositis.
Our data open up an interesting perspective: the three patients with chronic villitis had a shorter/milder history of COVID-19 (<14 days, estimated time for IgG appearance in patients serum)39, while chronic histiocytic intervillositis was found in a patient with long-standing and more severe symptoms (Figure 1). Could placental alterations mirror the stage of immune response to SARS-CoV-2 and/or infection severity? Should this prove to be true, placenta could be used as a representative tissue to investigate SARS-CoV-2 pathogenetic mechanisms helping avoid the biases of autoptic samples. Perspective and integrated studies (i.e. including seriated viral and serological testing) on larger series are needed to characterize these clinico-pathological correlations and address this hypothesis.
Vascular alterations represent the second type of abnormalities observed in our series. Thrombohemorrhagic alterations were mainly found in association with chronic villitis (Cases 1, 2 and 3): in these samples, thrombi involving the arterial microcirculation (<1 mm in diameter) were present in association with an in ammatory in ltrate, similarly to the alterations observed in the lungs of patients who died of COVID-19. 38 The relationship between COVID-19 and systemic coagulopathy with a thromboin ammatory state has been well documented40, 41 and placental thromboses were also described in two cases of SARS-CoV during the previous outbreak as previously listed. In Case 4 of our series, an extensive hemorrhagic placental abruption was observed, but this nding is probably unrelated to SARS-CoV-2 infection.
Regarding our knowledge about the clinical outcomes of pregnancy in COVID-19 patients, it is based on limited series and systematic reviews. Chen et al. 6 reported the outcomes of a relatively large series of pregnant women affected by COVID-19 observing a high rate (14/68, 21%) of preterm delivery. A systematic review evaluating 41 pregnancies affected by COVID-19 con rmed this observation, reporting preterm birth (< 37 weeks) in above 40% of pregnancies.9 Could the poorer obstetrics outcomes reported in women with COVID-19 be related to SARS-CoV-2induced in ammatory and/or vascular abnormalities? The ndings of our study provide an initial pathological basis for these clinical ndings: chronic villitis/VUE as well as vascular abnormalities are associated with several obstetric and neonatal adverse outcomes including preterm birth. 22,23,27,42,43,44 Moreover, it is well known that an in ammatory/infectious reaction can lead to preterm delivery by increasing prostaglandins secretion or by impairing the fetal immune tolerance.
Negative results of SARS-CoV-2 placental swabs observed in our series should not be taken as a conclusive evidence of SARS-CoV-2 absence in these tissues. This nding could rather depend on compartmentalization of virus replication within speci c cell types. This hypothesis is supported by the differential expression of ACE2 and TMPRSS2, proteins required for SARS-CoV-2 cell entry, observed in the maternal-fetal interface tissues. 17,45 This open question could also be e caciously addressed by in situ viral assays which will hopefully become more widely available in the coming weeks.
The main limitation of our observations is that our pathological ndings are shared by multiple conditions and etiologies, which could have been pre-existent at time of SARS-CoV-2 infection. Nevertheless, some considerations support this relationship: i) the observed rate of abnormalities is higher compared to what could have been expected in a consecutive unselected series of placentas; ii) ndings are consistent with other viral infections and with our current knowledge about SARS-CoV-2 pathogenetic mechanisms; iii) patients' histories did not include comorbidities justifying the identi ed pathological hallmarks. Analysis of larger series of cases and in situ viral assays will be needed to de nitively prove and characterize the causative role of SARS-CoV-2. Nevertheless, our ndings are important to shed a light on the reported associations between COVID-19 infection and poorer obstetrics outcomes.
Finally, can therapeutic implications be derived from our data? As prophylactic low-molecular-weight heparin is recommended in pregnant women with high thrombotic risk,46, 47 our ndings provide a pathological basis to consider heparin use in COVID-19 affected pregnant patients to prevent in ammatory and vascular thrombotic complications. This suggestion is in line with the recent proposal of The Society of Thrombosis and Haemostasis which recommends (if not otherwise contraindicated) the use of prophylactic low molecular weight heparin in all patients who require hospitalization for COVID-19, regardless of illness severity.48

Materials And Methods
Our study is based on the pathological analysis of 6 consecutive placentas delivered from 5 women affected by COVID-19 admitted to the Obstetrics and Gynecology clinic of S. Anna Hospital-"Città della Salute e delle Scienza di Torino", University of Turin from March 22, 2020 to April 23, 2020. Sample collection was performed within the framework of the "SARS-CoV-2 infection during pregnancy and puerperium: an Italian Obstetric Surveillance System (ItOSS)" study by the Italian National Institute of Health. All patients showed symptomatic COVID-19 disease con rmed by a nasopharyngeal swab according to World Health Organization (WHO) guidelines. SARS-CoV-2 RNA was detected using an automated real-time RT-PCR assay [DiaSorin Molecular Simplexa™ COVID-19 Direct, target genes S and ORF1ab]. Pathological examination was performed according to routine procedures and immunohistochemistry was set up on a BenchMark ULTRA platform (Ventana Medical Systems Inc., Tucson, AZ, USA). The study was conducted in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki and following amendments) for studies involving humans and within the guidelines and regulations de ned by the Research Ethics Committee of the University of Turin. Patients signed a written informed consent. Data supporting study results are included within the manuscript.

Con icts of Interest
The authors declare they have no con ict of interest.     Villous infarction areas were also present (D: HE, 40X).

Figure 5
Pathological ndings of Case 4. Extensive subchorial placental abruption was noted during macroscopic examination (stars) (A). Histological examination con rmed the presence of hemorrhagic areas located below the fetal surface (B: HE, 3X), associated with chronic hypoxia signs (C: HE, 60X).