Context.—The placenta is an important component in understanding the fetal response to intrauterine Zika virus infection, but the pathologic changes in this organ remain largely unknown. Hofbauer cells are fetal-derived macrophages normally present in the chorionic villous stroma. They have been implicated in a variety of physiological and pathologic processes, in particular involving infectious agents.
Objectives.—To characterize the fetal and maternal responses and viral localization in the placenta following Zika virus transmission to an 11 weeks' gestation fetus. The clinical course was notable for prolonged viremia in the mother and extensive neuronal necrosis in the fetus. The fetus was delivered at 21 weeks' gestation after pregnancy termination.
Design.—The placenta was evaluated by using immunohistochemistry for inflammatory cells (macrophages/monocytes [Hofbauer cells], B and T lymphocytes) and proliferating cells, and an RNA probe to Zika virus. The fetal brain and the placenta were previously found to be positive for Zika virus RNA by reverse transcription–polymerase chain reaction.
Results.—The placenta demonstrated prominently enlarged, hydropic chorionic villi with hyperplasia and focal proliferation of Hofbauer cells. The degree of Hofbauer cell hyperplasia gave an exaggerated immature appearance to the villi. No acute or chronic villitis, villous necrosis, remote necroinflammatory abnormalities, chorioamnionitis, funisitis, or hemorrhages were present. An RNA probe to Zika virus was positive in villous stromal cells, presumably Hofbauer cells.
Conclusions.—Zika virus placental infection induces proliferation and prominent hyperplasia of Hofbauer cells in the chorionic villi but does not elicit villous necrosis or a maternal or fetal lymphoplasmacellular or acute inflammatory cell reaction.
Zika virus infection has been confirmed to be a cause of fetal malformation1—an unprecedented event for an arthropod-borne infection. The spectrum of disease caused by Zika virus among adults and children is still being investigated and defined, but it is clear in the most recent epidemic in South and Central America, Mexico, and the Caribbean islands that when infection occurs in pregnant women, fetuses appear to be at increased risk for a poor clinical outcome.2–9 The most significant effect of the Zika virus on the fetus is the occurrence of neuronal necrosis and microcephaly—this has been demonstrated in neuroimaging studies performed both before delivery and following the birth of infected infants.2,3,7 Pathology studies have been shown to be an important component of the public health investigation of emerging infections, contributing to the understanding of the mechanisms of tissue injury, for example, human immunodeficiency virus, Ebola virus, hantavirus, West Nile virus, and many others.10–13 In those situations in which infections may be transmitted from pregnant women to their fetuses, placental pathology can provide important information on the mechanism(s) of maternal-fetal (vertical) transmission, the nature of the inflammatory response of the mother and fetus to the infection, and the effects of the organisms on the placenta, such as necrosis, hemorrhage, or vascular disease.14 This report details the microscopic changes in the placenta from a 21 weeks' gestation fetus whose mother acquired symptomatic Zika virus infection at 11 weeks' gestation.15
MATERIALS AND METHODS
The clinical history of the mother and infant have been previously described.15 In summary, a 33-year-old pregnant woman developed symptomatic Zika virus infection after her return from travel in Central America. She was bitten by mosquitoes at 11 weeks' gestational age and subsequently developed mild fever, myalgia, rash, and ocular pain beginning at 12 weeks' gestation. Four weeks later, her serum was positive for immunoglobulin (Ig) G and IgM antibodies to Zika virus. A nested reverse transcription–polymerase chain reaction (RT-PCR) performed at 16 weeks' gestation confirmed Zika virus in her serum. The United States Centers for Disease Control and Prevention in Atlanta confirmed maternal Zika virus infection with serum positivity for Zika IgM of greater than 1:2560, using a plaque-reduction neutralization test at 17 weeks' gestation. Between the 16th and 20th weeks of gestation, ultrasonography revealed a decrease in the fetal head circumference from the 47th to the 24th percentile. Fetal ultrasonography at 19 weeks' gestation revealed abnormal intracranial anatomy. Following this, fetal magnetic resonance imaging at 20 weeks' gestation demonstrated more severe abnormalities of the brain, including diffuse atrophy in the cerebral mantle, absence of the normal lamination patterns of the cerebral mantle, a short corpus callosum, and mild enlargement of the third and lateral ventricles. On the day before an elective termination of the pregnancy at 21 weeks' gestation, the mother was found to still have serum positive for Zika virus RNA with a low viral count (21.1 × 103 copies/mL). The amniotic fluid was also positive for Zika virus RNA at the time of pregnancy termination. Following delivery, an autopsy of the fetus was performed that revealed diffuse cerebral cortical thinning with extensive apoptosis of postmigratory neurons of the neocortex, micromineralization, prominent volume loss of the white matter and subventricular zone, axonal rarefaction, and macrophage infiltration. Viral particles were seen with electron microscopy, and high loads of Zika viral RNA were detected with RT-PCR.15 The placenta was submitted for pathologic examination. Substantial Zika viral loads were identified in the placenta, extraplacental membranes, and umbilical cord with quantitative RT-PCR.
All immunohistochemical staining procedures were performed on a Leica Bond-MAX (Leica Biosystems, Wetzler, Germany). Antigen retrieval and antibody conditions are presented in the Table.
RNAscope In Situ Hybridization
In situ expression analyses of mRNA were performed by using a Zika virus mRNA probe (catalog No. 463781) and developed with an RNAscope kit (Advanced Cell Diagnostics, Inc., Newark, California) by using 3,3′-diaminobenzidine-based detection and hematoxylin counterstain. Briefly, paraffin-embedded sections were pretreated according to the manufacturer's protocol. Probes for Zika virus mRNA were applied and positive and negative control probes (peptidylprolyl isomerase B and dihydrodipicolinate reductase, respectively) were used to qualify the tissue samples for use in the in situ hybridization application. A paraffin-embedded brain section from the same patient served as a positive control for detection of Zika virus mRNA expression. Individual slides were viewed on an Olympus Eclipse 55i microscope (Shinjuku, Tokyo, Japan) and images acquired on a Nikon DS-Fi1 digital camera (Minato, Tokyo, Japan).
Microscopic examination of the placental disc, using routine hematoxylin-eosin staining, revealed secondary and tertiary chorionic villi that were enlarged, even for 21 weeks' gestation, with focally marked stromal edema (Figure 1, A through C). Some of the enlarged and edematous villi had irregular outlines. Most chorionic villi at all levels demonstrated prominent hypercellularity of the villous stroma. These stromal cells were of variable appearance, from rounded to spindle shaped, and had the morphologic features of villous stromal macrophages (Hofbauer cells). Some of the Hofbauer cells contained clear intracytoplasmic vacuoles. No evidence of ongoing or remote acute or chronic villitis or intervillositis was seen. Focal and mild intravascular karyorrhexis was present in some chorionic villus vessels, which we attributed to recently induced intrauterine fetal demise.16 No villous necrosis, remote necroinflammatory abnormalities, reparative changes, or villous fibrosis was present. The umbilical cord and extraplacental membranes were normal for gestational age—there was no funisitis or chorioamnionitis. No significant hemorrhages, either acute or remote, were present in any of the placental tissues. No abnormalities of the trophoblast were present.
CD163 and CD68 immunostains demonstrated intensely positive staining in the hyperplastic population of stromal cells of both secondary and tertiary villi, confirming their identity as stromal macrophages, or Hofbauer cells. The CD163 highlighted the intensity of the Hofbauer cell hyperplasia—in some villi, the macrophages were tightly packed and confluent (Figure 2, A through F). It was obvious that the enlargement of the villi was due not only to edema, but also to Hofbauer cell hyperplasia, which had also significantly contributed to the greater-than-expected size of the chorionic villi. Ki-67 demonstrated positive intranuclear staining in occasional stromal macrophages (Figure 3, A and B), which was consistent with the presence of a hyperplastic population of Hofbauer cells. Some cytotrophoblast cells also expressed Ki-67 staining, which was expected. Stains for lymphocytes and plasma cells confirmed the impression that there was no villitis.
In situ hybridization using a Zika virus RNA probe demonstrated scattered, strongly positive staining cells within the villous stroma of the chorionic villi (Figure 4), which were presumably Hofbauer cells.
Together with epidemiologic, clinical, and radiologic investigations, autopsy pathology studies have significantly contributed to the consensus conclusion that the Zika virus can produce destructive lesions of the fetal brain and microcephaly.15,17–23 There now remains no doubt that Zika virus can reach the placenta during maternal viremia, be transmitted to the developing fetus, and cause brain damage, microcephaly, other malformations (congenital Zika syndrome) and, in some, cases, perinatal death. In all published autopsies of fetuses and neonates with microcephaly from mothers clinically suspected or confirmed as having Zika virus infection during pregnancy, the virus has been identified in the fetal brain tissue; it has also been identified from other fetal organs and the amniotic fluid.23–25 In vitro studies have provided evidence that Zika virus has an affinity for infecting cells of neural origin.26–28 A murine model of congenital Zika virus transmission has been developed, which has demonstrated fetal demise, brain damage, and Zika virus present in trophoblast, consistent with transplacental infection.29 However, the mouse placenta is chorioallantoic, making it difficult to extrapolate these findings to the human hemomonochorial placenta.
Because of the severity of the autopsy pathology findings in the central nervous system of the fetus, the placental findings in this present case of absence of necrosis or inflammation were unexpected. There was no evidence of ongoing, remote, acute or chronic villitis or intervillositis, either by routine histologic staining or by using antibodies to neutrophils, lymphocytes, and plasma cells. Villitis is typically present in most maternal blood-borne infections that are vertically transmitted, including infections by protozoal (Chagas disease, toxoplasmosis), bacterial (Listeria, syphilis, and gram-negative and gram-positive bacteria), and viral (rubella, cytomegalovirus) agents.30 In addition to producing a maternally or fetally derived inflammatory reaction in the chorionic villi, which may be accompanied by villous necrosis, transplacental transmission of these agents also produces inflammation and cell necrosis in the fetal somatic organs. Zika virus appears to be different from some TORCH agents (such as cytomegalovirus, Treponema pallidum [syphilis], Toxoplasma, and rubella virus) and other hematogenously transmitted infections in transiting from the maternal to the fetal circulatory systems through the placenta, producing a range of tissue-destructive abnormalities in the fetal brain, and yet not eliciting a maternal or fetal inflammatory leukocytic response in the placenta. This is indicative of a highly selective predilection of the virus for cells of the central nervous system.
The most striking observation in the placenta was a prominent and diffuse hyperplasia of Hofbauer cells. These cells are of monocytic derivation and are, along with fibroblasts, a normal component of the stroma of the chorionic villi. Hofbauer cells are of fetal origin, and they first appear in the chorionic villi at the 10th to 18th days of gestation. Because this predates the development of a fetal circulation, they are believed to initially be of fetal mesenchymal origin, derived from monocyte progenitor cells of the hypoblast-derived yolk sac that have migrated to the mesenchymal core of the villi. Later in gestation, it has been suggested that they are derived from a population of recruited fetal monocytes.31,32 Hofbauer cells have many histologic features of macrophages in other organs—they are large (10-30–μm diameter) cells with cytoplasmic processes, and contain large vacuoles, pinocytotic vesicles, and intracytoplasmic granules. They have been characterized as M2/alternatively activated macrophages,33 and their function includes phagocytosis of fluids and apoptotic materials, antigen presentation in response to infectious agents, and possibly an angiogenic role in early placental vasculogenesis, maintenance of placental water balance, and an endocrine function. Because Hofbauer cells can vary in size and shape, they can easily be confused with mononuclear inflammatory cells, and especially plasma cells, leading to an erroneous diagnosis of chronic lymphocytic or plasmacellular villitis (Figure 1, C). In placentas from normal gestations, Hofbauer cells become few in number by the fourth to fifth month of gestation, and immunohistochemistry with antibodies to macrophages/monocytes may be necessary to visualize them.31,32,34 Hyperplasia of Hofbauer cells is abnormal and has been reported to occur as a result of a wide variety of pathologic conditions of pregnancy. These include ascending infections,35 villitis of unknown etiology,36 and maternal blood-borne infections that cause villitis including TORCH infections such as syphilis and cytomegalovirus infection,37,38 and Chagas disease.39 The mechanism(s) by which Hofbauer cell hyperplasia occurs in response to an inciting factor is, at least in part, the result of proliferation of these cells within the chorionic villous stroma. Hofbauer cells have been observed to contain mitotic figures.40 The ability of Hofbauer cells to proliferate within the chorionic villous stroma has been confirmed by using antibodies to cellular proliferation markers such as Ki-67.41,42
In the present case, symptomatic maternal infection began at 11 weeks' gestation while visiting an area endemic for Zika virus, after which the mother was confirmed to have Zika virus infection with both serologic and RT-PCR methods. Her serum remained positive for Zika virus RNA up to the day before the termination of pregnancy at 21 weeks' gestation.15 It appears clear that during the approximate 10-week period from initial symptomatology and prolonged duration of maternal viremia before the demise of the fetus, transplacental transmission of Zika virus to the fetus occurred, resulting in activation, induction of proliferation, and eventually, a prominently hyperplastic population of Hofbauer cells. We can surmise that this placental reaction began soon (rather than later) after the onset of maternal infection and viremia. This is based on the obstetrical ultrasound evidence of an initial decrease in the relative percentile of fetal head growth documented at 16 weeks' gestation, and the extent of Hofbauer cell hyperplasia in the chorionic villi with relatively few stromal macrophages in the proliferative stage of the cell cycle by the time of delivery of the fetus and placenta.
We have demonstrated that 9 weeks after initial maternal symptomatology and prolonged maternal viremia, the placenta still had detectable Zika virus present within stromal (Hofbauer) cells of the chorionic villi together with Hofbauer cell hyperplasia. The significance of the occurrence of Hofbauer cell proliferation and hyperplasia, together with the demonstration of residual virus in these stromal cells following transplacental fetal infection, is especially interesting in light of a recent investigation that has demonstrated that Zika virus can experimentally infect, and subsequently replicate in, human Hofbauer cells. In this study, Quicke et al43 found that Hofbauer cells isolated from term human placentas were permissive to infection by a contemporary strain of Zika virus that is currently circulating in the Americas. Following in vitro infection of Hofbauer cells by Zika virus, viral replication was documented in these cells, which was accompanied by induction of type I interferon and proinflammatory cytokines. Viral infection also resulted in a modest activation of Hofbauer cells in which there was upregulation of retinoic acid–inducible gene I (RIG-I)–like receptor transcription as well as downstream antiviral effector genes, thus indicating that Zika virus infection of Hofbauer cells induced an antiviral response. The finding by these investigators that in vitro infection and replication of Zika virus infection resulted in only minimal cellular necrosis is especially interesting because of our findings of a lack of recent or remote necrosis in the placental chorionic villi of the infected fetus from this report. Also of significance in light of our findings is the recent report by Martines and colleagues44 who describe the immunohistochemical identification of Zika virus antigens within Hofbauer cells in chorionic villi from the placenta of a fetus that was spontaneously aborted at 11 weeks' gestation.
It is still not well understood what role(s) the Hofbauer cell has in facilitating or inhibiting transplacental transmission of infectious agents such as the Zika virus. However, based on the demonstration in this communication of proliferation and prominent hyperplasia of Hofbauer cells in the placenta from a microcephalic fetus infected early in gestation, the identification of residual Zika virus in villous stromal cells, using an RNA probe, and the previously published results of in vitro infection and replication of Zika virus in human Hofbauer cells, it appears highly probable that the Hofbauer cell has an important, or even primary, role in those cases where transplacental transmission of the Zika virus does occur. The unexpected absence in placental tissues of any necrosis or leukocytic response by the mother or fetus to transplacental Zika virus infection is also interesting and of unknown significance.
The authors have no relevant financial interest in the products or companies described in this article.