Hofbauer Cells and COVID-19 in Pregnancy: Molecular Pathology Analysis of Villous Macrophages, Endothelial Cells, and Placental Findings From 22 Placentas Infected by SARS-CoV-2 With and Without Fetal Transmission

This case-based retrospective study enrolled maternal–fetal pairs consisting of women having had a positive test result for SARS-CoV-2 during pregnancy using reverse-transcriptase polymerase chain reaction (RT-PCR) occurring before delivery in which placenta was submitted for pathology examination and found to be infected with SARS-CoV-2 by direct visualization of fetal-derived placental cells, either using immunohistochemistry for SARS-CoV-2 antigens, RNA in situ hybridization for SARS-CoV-2 nucleic acid, or both. These cases were contributed to the investigation by a collaborative group of 29 perinatal physicians and molecular pathologists and biologists working with the disease.

Hofbauer cells were identified using immunohistochemistry to a variety of macrophage antigens including CD14, CD68, and CD163. In some cases, double staining was performed using immunohistochemical staining for Hofbauer cells and either immunohistochemistry for SARS-CoV-2 antigens or RNA in situ hybridization for SARS-CoV-2 nucleic acid. The placentas from Cases 10 and 15 were immunohistochemically evaluated for Hofbauer cell proliferation by staining with antibody to the Ki-67 nuclear proliferation marker. In each case, pertinent maternal and neonatal clinical data and the results of testing for SARS-CoV-2, including maternal–fetal transmission status (when available), were obtained, and the presence of significant placental pathology abnormalities was recorded.

In those cases that previously had some aspect of the case published, the references are provided. For all 22 cases occurring in 7 countries that comprised this study group, the 29 pathologists, clinicians, or others involved with these patients were personally contacted by 1 of the authors (DAS) requesting reexamination of the placenta to evaluate Hofbauer cell involvement and confirmation of the clinical, laboratory, and pathology findings. In some cases, this led to additional molecular testing of placental tissues. For all cases there was either approval received from the local institutional review boards or institutional waiver and parental permission obtained, and there was compliance with the Declaration of Helsinki for Human Research.

A total of 22 placentas were examined for the presence of SARS-CoV-2 within Hofbauer cells using immunohistochemistry, RNA in situ hybridization, or both (Tables 1 through 3). With 1 exception (Case 20, discussed below), all placentas had strong positive staining of the syncytiotrophoblast for SARS-CoV-2 using immunohistochemistry, RNA in situ hybridization, or both methods.

Nineteen of 22 placentas (86.4%) did not reveal staining of any villous stromal cells, including Hofbauer cells, for SARS-CoV-2 (Figure 1; Figure 2, A and B; Figure 3, A and B; and Figure 4). In some of these placentas, double-staining immunohistochemistry was performed using antibodies to macrophages and to SARS-CoV-2 nucleocapsid protein, which were also negative for Hofbauer cell involvement. In some cases, Hofbauer cells were identified that closely approached, and in some cases abutted upon, the trophoblast basement membrane zone (Figure 4), but these macrophages were also negative for viral staining.

Among 22 placentas there were 4 placentas (18.2%) that demonstrated positive staining of Hofbauer cells for SARS-CoV-2. These included Cases 15 (Figure 5 and Figure 6, A through C), Case 19 (Figure 7, A and B), Case 20 (Figure 8 and Figure 9, B through D) and Case 21 (Figure 10, A and B). In 7 cases with documented transplacental infection of the fetus (Cases 10–12, 14, 15, 19, 22), 2 occurred in placentas with Hofbauer cell staining positive for SARS-CoV-2.

In Case 15, immunohistochemistry revealed cells with strong intracytoplasmic staining for SARS-CoV-2 nucleocapsid protein present within the stroma of multiple chorionic villi. These cells had the morphologic features of macrophages, consistent with Hofbauer cells (Figure 5). Double staining using antibodies to macrophage markers CD68, CD163, and CD14 together with antibody to SARS-CoV-2 nucleocapsid protein demonstrated that Hofbauer cells were positive for intracytoplasmic SARS-CoV-2 staining (Figure 6, A through C). In addition, there was intense staining for SARS-CoV-2 of the syncytiotrophoblast. No viral staining of the villous capillary endothelium was present. Hofbauer cell hyperplasia was present, and double staining using antibody to Ki-67 nuclear proliferation antigen and SARS-CoV-2 nucleocapsid protein revealed that occasional Hofbauer cells were in the proliferation phase of the cell cycle.

Case 19 was a preterm placenta in which there was positive intracytoplasmic staining of villous stromal cells morphologically consistent with Hofbauer cells as well as intense positive staining of the syncytiotrophoblast using antibody to SARS-CoV-2 nucleocapsid protein (Figure 7, A and B).

Case 20 had both a unique clinical history as well as an array of placenta pathology findings. The mother had initially been hospitalized for symptomatic SARS-CoV-2 infection at 26 weeks gestation. She delivered her infant 11 weeks later, at 37 weeks gestation, at which time both mother and neonate had negative nasopharyngeal swabs for SARS-CoV-2 using polymerase chain reaction (PCR) testing. The placenta had extensive fibrin deposition (Figure 8), syncytiotrophoblast necrosis, and mild chronic histiocytic intervillositis. Immunohistochemistry using antibody to SARS-CoV-2 nucleocapsid protein demonstrated numerous villous stromal cells having the morphologic features of Hofbauer cells with positive intracytoplasmic staining (Figure 9). Capillary villous endothelial cells also stained positively for the virus, and in some vessels, it appeared that more than 1 endothelial cell stained positively for SARS-CoV-2 (Figure 9, C). The syncytiotrophoblast did not stain for SARS-CoV-2—the only placenta in this cohort to show negative staining in these cells.

Case 21 was a placenta from a stillborn infant at 35 4/7 weeks gestation. The placenta showed extensive intervillous fibrin deposition suggesting massive perivillous fibrin deposition, syncytiotrophoblast necrosis, and mild chronic histiocytic intervillositis. Immunohistochemical staining with antibody to SARS-CoV-2 nucleocapsid protein showed intense positive staining of syncytiotrophoblast. There were numerous villous stromal cells having the morphologic features of Hofbauer cells that stained positively for SARS-CoV-2 (Figure 10, A and B). Positive staining for SARS-CoV-2 was also present in occasional villous capillary endothelial cells (Figure 10, C). The stillborn infant was not tested for SARS-CoV-2.

Three of 22 placentas (13.6%; Cases 10, 15 and 20) demonstrated Hofbauer cell hyperplasia (Figure 3, A and B, and Figure 9, A). Double immunohistochemical staining of a placenta (Case 15) from a fetus with transplacental SARS-CoV-2 infection using the proliferation marker Ki-67 and antibody to SARS-CoV-2 nucleocapsid protein showed numerous cytotrophoblast cells in the proliferative phase of the cell cycle, but relatively few Hofbauer cells (Figure 11, A). In the placenta of another fetus having acquired infection through transplacental transmission (Case 10), immunohistochemistry with antibody to Ki-67 showed that there were many cytotrophoblast but very few Hofbauer cells in the proliferative phase of the cell cycle (Figure 11, B). The lack of prominent Hofbauer cell proliferation is in marked contrast to that seen in placentas of fetuses having congenital Zika syndrome (Figure 11, C).

In 2 placentas (Cases 20 and 21) there was positivity of villous capillary endothelial cells identified for SARS-CoV-2 (Figure 9 and Figure 10, C). These 2 placentas also had staining of Hofbauer cells present. Both Cases 20 and 21 appeared to show that some villous vessels had greater than 1 endothelial cell staining positive per lumen for SARS-CoV-2. No villous capillary endothelial staining was present in the 19 placentas in which Hofbauer cell staining was absent.

The syncytiotrophoblast stained positive for SARS-CoV-2 in 21 of 22 (95.5%) placentas. The only placenta in which the syncytiotrophoblast did not stain was Case 20, in which there was an 11-week interval between maternal illness and test positivity for SARS-CoV-2 and the delivery.

In one case (Case 15), staining of both syncytiotrophoblast and cytotrophoblast for SARS-CoV-2 was present.

Chronic histiocytic intervillositis was present in all placentas (100%). In all placentas the characteristic accumulation of intervillous histiocytes was identified by both routine histologic staining as well as immunohistochemical staining for macrophage antigens. The degree of intervillous inflammation varied between the different cases; in some cases where there was extensive fibrin deposition or massive perivillous fibrin deposition, the extent of chronic histiocytic intervillositis appeared to be less severe than in cases where the fibrin deposition was not as prominent.

Although it was not a focus of our study, it was observed in 5 placentas having chronic histiocytic intervillositis that histiocytes within the inflammatory infiltrate stained positive for SARS-CoV-2 nucleocapsid antibody.

The COVID-19 pandemic has been remarkable for the wide-range spectrum and duration of disease that it produces, affecting persons of all ages and ethnicities. In particular, infection of pregnant women and vertical transmission of SARS-CoV-2 has remained a special public health problem due to the susceptibility of mothers, fetuses, and neonates to viral infections.^41–43  Following the demonstration that SARS-CoV-2 can be transmitted from a pregnant woman to her fetus through the placenta,^19,31,32  this coronavirus can now be considered a TORCH (toxoplasma, other, rubella, cytomegalovirus, herpes) infection.

Examination of placentas in cases of maternal–fetal infection due to TORCH agents has been greatly beneficial in understanding not only the specific placental cell types that are susceptible to infection, but also in evaluating potential mechanisms of transplacental transmission. In studying transplacental viral infections, among the most important cells to be examined are the resident population of immunologically active placental macrophages, the Hofbauer cells. Hofbauer cells are large, pleomorphic fetal-derived macrophages that constitute an important cellular component of the maternal–fetal interface and the placental immune system. They are located in an advantageous position to perform this function, within the chorionic villous stroma adjacent to both the fetal villous capillaries and the overlying trophoblast layer, where potential transfer of substances and microbial agents across the maternal–placental interface can occur.⁴⁴  Hofbauer cells are present throughout all trimesters of pregnancy and have features that most closely resemble alternatively activated macrophages that have been termed M2a, M2b, M2c, and M2d polarity subtypes that have an antiinflammatory or regulatory phenotype.⁴⁵  Hofbauer cells serve many functions—angiogenesis and vasculogenesis, host defense, villous development and stromal maturation, fluid homeostasis, and clearance of apoptotic cells.^45–47  Evidence suggests that Hofbauer cells may also be involved in the vasoregulation of placental blood vessels, as in vitro studies have found that they produce both prostaglandin E₂ and thromboxane.⁴⁶  It has been experimentally demonstrated that when exposed to an infectious agent or inflammatory stimulus, Hofbauer cells may express an M1 or proinflammatory phenotype that can release cytokines, damage the villus, and result in a fibrotic response.^45,48–50 

Similar to other types of macrophages, Hofbauer cells have plasticity in their functions, and in addition to their role in placental development and homeostasis they have been implicated in the pathogenesis of a number of TORCH agents.⁵¹  In particular, they can be the targets for several viruses including HIV and Zika virus.^39,51–53  Hofbauer cells are susceptible to HIV infection—they express not only the CD4 receptor, which is bound by the viral env receptor, but also HIV coreceptors including dendritic cell specific intercellular adhesion molecule-3-grabbing non integrin, CCR5, and C-X-C chemokine receptor type 4.^39,54  Before the onset of the COVID-19 pandemic, the most recent example of a newly emergent virus causing transplacental infection of the fetus was a flavivirus—Zika virus. Similar to HIV, placentas infected with Zika virus from neonates with congenital infection are not typically associated with an inflammatory process of either maternal or fetal origin. Although the mechanism(s) by which Zika virus penetrates the trophoblastic barrier remains unclear, the virus can undergo productive infection of Hofbauer cells in vitro.^55–57  In placentas from fetuses with congenital Zika syndrome after intrauterine infection, Hofbauer cells have been found to contain the Zika virus and to undergo proliferation in response to placental infection.^40,58 

In the early stages of the COVID-19 pandemic, pathology investigations of placentas from pregnant women having SARS-CoV-2 infection revealed a range of findings that included fetal vascular malperfusion,⁵⁹  maternal vascular malperfusion,⁶⁰  and even no specific abnormalities.⁶¹  In the majority of these cases, the placentas and neonates were not found to be infected with SARS-CoV-2. However, the placental pathology findings from fetuses infected with SARS-CoV-2 after transplacental transmission are very different from those of neonates with congenital Zika virus infection. Intrauterine transplacental transmission of SARS-CoV-2 has been found to be associated in most cases with 3 significant pathology findings—chronic histiocytic intervillositis, necrosis of the syncytiotrophoblast, and identification of the virus in syncytiotrophoblast using immunohistochemistry or RNA in situ hybridization.^31,32  In addition to these 3 findings, increased fibrin deposition within the intervillous space, which in some placentas may be so severe that it reaches the criteria for massive perivillous fibrin deposition, appears to accompany placental infection with SARS-CoV-2.^31,32  In contrast to placental Zika virus infection in which virus is most frequently identified in Hofbauer cells, in placentas infected with SARS-CoV-2 the syncytiotrophoblast appears to be the most frequently involved cell type.⁶²  However, with the demonstration that SARS-CoV-2 can be identified in the syncytiotrophoblast from infected placentas as well as from fetuses having intrauterine infection, the question arises as to whether the virus can pass beyond the trophoblast and into cells in the underlying villous stroma.

There have been previous individual case reports of both positivity and negativity of Hofbauer cells for SARS-CoV-2 staining,^26,38,63  but this present article describes the first systematic investigation of SARS-CoV-2 involving Hofbauer cells and other cell types in a large cohort of placentas infected with the virus. Evaluation of 22 placentas infected with SARS-CoV-2 revealed that Hofbauer cells were immunohistochemically positive for viral staining in 4 cases (18.2%; Cases 15, 19, 20, 21). In some cases, Hofbauer cells that were adjacent to and even in intimate contact with, the trophoblast basement membrane zone stained negative for SARS-CoV-2. While we were unable to determine if positive staining represented intact virions, it indicates that, at the least, SARS-CoV-2 material can penetrate through the trophoblast layer into the core of the chorionic villus. It should be emphasized that in this article we have referred to staining positivity of Hofbauer cells for SARS-CoV-2 as involvement, not infection, because positive staining of Hofbauer cells is only indicative of viral material within the cytoplasm of macrophages. Whether productive SARS-CoV-2 replication is occurring within these Hofbauer cells cannot be determined but would be important to know. Hofbauer cell involvement was also not a requirement for transplacental viral transmission in this series. Among the 22 infected placentas, there were 7 cases (31.8%) of transplacental SARS-CoV-2 transmission (Cases 10, 11, 12, 14, 15, 19, 22). In only 2 (9.1%) of these transmitting cases (Cases 15 and 19) were Hofbauer cells found to stain positive for SARS-CoV-2.

Because Hofbauer cell proliferation and hyperplasia were significant features in some placentas infected with Zika virus, we looked for these features in our placental cohort. Although Hofbauer cell hyperplasia was present in 3 placentas (13.6%), Case 10 (Figure 3, A and B), Case 15, and Case 20 (Figure 9, A), none demonstrated the extent of hyperplasia seen in Zika virus–infected placentas. Immunohistochemical staining of 2 placentas (9.1%; Cases 10 and 15) infected with SARS-CoV-2 using the nuclear proliferation marker Ki-67 showed some proliferative activity of Hofbauer cells, but the extent of Hofbauer cell proliferation was not as high as present in placentas infected with Zika virus (Figure 11, C).

Villous capillary endothelial staining for SARS-CoV-2 was identified in 2 of 22 placentas (9.1%; Cases 20 and 21). Although these endothelial cells lining the villous capillaries are in direct contact with fetal blood circulating in the chorionic circulation, there was no identifiable association with viral staining of these cells and maternal–fetal SARS-CoV-2 transmission. An interesting observation is that in both placentas that had positive staining of endothelial cells for SARS-CoV-2, Hofbauer cells also stained positively for the virus. Staining of endothelial cells for SARS-CoV-2 was not observed in placentas that did not also demonstrate Hofbauer cell staining for the virus. Although these numbers are too small for statistical analysis, it seems not only biologically feasible but almost beyond coincidence that these findings are unassociated and suggests that when SARS-CoV-2 penetrates through the protective trophoblastic layer and gains access to the chorionic villous stroma, viral material (or virions) can enter or be sequestered within such stromal cells as Hofbauer cells and endothelial cells. Endothelial cells that line the blood vessels have many roles that include limiting access of infectious agents, toxins, and other materials between the bloodstream and surrounding tissues. Although not phagocytic, endothelial cells have been reported to internalize, or sequester, microsize objects that include blood clots, senescent cells, and even bacteria, such as Listeria monocytogenes.^64–66  This phenomenon may explain the staining positivity of endothelial cells in these 2 placentas.

Chronic histiocytic intervillositis has been found to occur together with placental infections with SARS-CoV-2 as well and transplacental transmission and stillbirths.^31,32,62  All 22 placentas (100%) that were infected with SARS-CoV-2 in this cohort had the finding of chronic histiocytic intervillositis in varying degrees of intensity. Another published finding associated with placental infection with SARS-CoV-2, trophoblast necrosis, was also present in all 22 cases.

One case needs to be discussed in detail because of its atypical features. Case 20 was unusual because of the 11-week interval between the diagnosis of SARS-CoV-2 at 26 weeks gestation in a symptomatic pregnant mother and the delivery of her uninfected infant at 37 weeks.^38,67  The placenta demonstrated chronic histiocytic intervillositis, trophoblast necrosis, and increased fibrin, with immunohistochemical positivity for SARS-CoV-2 in Hofbauer cells and villous capillary endothelial cells but not syncytiotrophoblast. Although speculative, given the 11-week interval between symptomatic maternal infection and delivery of the placenta (and infant), we suggest the following explanation. An initial placental infection around the time of maternal SARS-CoV-2 infection at 26 weeks gestation was accompanied by infection and necrosis of the syncytiotrophoblast with chronic histiocytic intervillositis and fibrin deposition. During the subsequent weeks, extension of SARS-CoV-2 into cells in the villous stroma occurred together with residual inflammation and fibrin deposition, but with resolution and clearing of virus from necrotic trophoblast and regeneration of uninfected syncytiotrophoblast. It has been previously observed that neonatal test positivity for SARS-CoV-2 is frequently transient, becoming negative in many newborn infants a short time after a positive result.^15,68 

This investigation of 22 placentas infected with SARS-CoV-2 has reinforced previous research that showed chronic histiocytic intervillositis and trophoblast necrosis are associated with placental infection.^31,32,62  It has also demonstrated that varying degrees of increased fibrin deposition, up to the level of massive perivillous fibrin deposition, are also associated with placental infection with SARS-CoV-2. Significantly, we have demonstrated that in a small number of infected placentas, SARS-CoV-2 is not limited to the trophoblast layer but can extend beyond it and into the chorionic villous stroma to involve Hofbauer cells and villous capillary endothelial cells, and that positive viral staining in both villous cell types can occur together. Unlike Zika virus infection, in most placentas infected with SARS-CoV-2 the Hofbauer cells did not demonstrate any excessive hyperplastic or proliferative activity. Although our series of placentas is small, there was no evidence that viral staining of either Hofbauer cells or capillary endothelial cells was associated with a greater probability of maternal–fetal transmission or poor outcome. It seems probable that most cases of transplacental infection of the fetus with SARS-CoV-2 occur in the absence of Hofbauer cell and endothelial involvement.

The authors thank Ke Cheng, PhD, Justin Mann, BS, and the staff of HistoWiz, Brooklyn, New York, for their generous cooperation in performing routine and molecular pathology testing of selected specimens. We would also like express our gratitude to Rodrigo Munoz Mitev, MD, Department of Clinical Pathology, Lund University, Lund, Sweden for his assistance with the placenta microscopy pictures from the Swedish case.