Maternal SARS-CoV-2, Placental Changes and Brain Injury in 2 Neonates

A 32-week gestational age, appropriately grown, Hispanic male infant was born to a 21-year-old multigravida via Cesarean section. The mother was healthy until 27 weeks of gestation when she was admitted to the ICU for pneumonia and multisystem disease, diagnosed as SARS-CoV-2 by nasopharyngeal swab reverse transcription polymerase chain reaction (NP RT-PCR). Continuous fetal monitoring and ultrasound evaluations remained normal. At 32 weeks of gestation, a multidisciplinary decision was made to perform a Cesarean section for maternal indications after administration of antenatal corticosteroids for fetal lung maturation. The mother was SARS-CoV-2 negative by this time. The infant was born with Apgar scores of 4 and 7 at 1 and 5 minutes with an umbilical cord arterial blood gas pH 7.22, pCO2 97 mm Hg, and base excess +11. He immediately displayed seizure-like activity and poor respiratory effort, which required intubation and assisted ventilation. Continuous, conventional electroencephalography recordings on admission showed frequent multifocal epileptiform discharges consistent with status epilepticus. Initial hematologic indices and serum chemistries were normal. Early hypotension with biventricular systolic depression on echocardiography was treated with vasopressors until day 5. Significant lung disease persisted despite administration of early surfactant and broad-spectrum antibiotics; bilateral confluent densities originally presented on the initial chest radiograph did not completely resolve after full respiratory recovery (day 19). At 24 hours of age, the infant’s NP RT-PCR for SARS-CoV-2 was negative. However, serum coronavirus disease 2019 (COVID-19) IgG and combined (total) IgG, IgM, and IgA reactivity to a recombinant derivative of SARS-CoV2 spike protein were present, and serum inflammatory markers and cytokines were markedly elevated (Table 1). Cerebrospinal fluid (CSF) bacterial and viral studies were negative. Bacterial cultures from blood, urine, and endotracheal sites were also negative, as was herpes simplex PCR in blood and CSF. Chromosomal microarray, newborn metabolic screen, plasma amino acids, acylcarnitines, and urine organic acid screens were also normal. An epilepsy genetic panel was noncontributory. Brain MRI on day 2 showed left germinal matrix and left intraventricular hemorrhage (Fig 1A). Follow-up brain MRI at 10 weeks showed resolution of hemorrhages, but severe parenchymal atrophy (Fig 1B). The infant was discharged after 3 months (43 weeks’ postmenstrual age) with a seizure disorder and acquired microcephaly (head circumference < 2nd percentile compared with 81st percentile at birth). Eventually, the seizures resolved and antiepileptics were discontinued, but the infant was readmitted on several occasions for failure to thrive and respiratory infections. His 12-month neurologic examination remained abnormal, with head lag, truncal hypotonia, increased appendicular tone, hyperreflexia, and delayed developmental milestones. At 13 months of age, he had an emergency department visit for an upper respiratory infection. Three days later, his mother found him unresponsive in bed; paramedics who responded reported finding him in asystolic cardiac arrest, from which he could not be revived. Autopsy showed a significant reduction in brain weight and cerebral white matter volume, enlarged ventricles, extensive gliosis in the cerebrum, cerebellum and brainstem, and evidence of virus throughout the brain (Fig 1, E1–8).

A 39-week gestational age, appropriately grown, Hispanic female infant was born to a 20-year-old primigravida, who reported asymptomatic SARS-CoV-2 positivity (NP RT-PCR) in the late second trimester, followed by a negative test. Her prenatal assessments and ultrasound scans had all been unremarkable. Although asymptomatic, she again tested positive at the time of vaginal delivery, also complicated by clinical chorioamnionitis. The infant was awarded Apgar scores of 4 and 6 at 1 and 5 minutes, respectively, and had an umbilical cord arterial blood gas of pH 7.22, pCO2 47 torr, and base deficit −10. She required nasal continuous positive airway pressure for apnea and was begun on antibiotics for presumptive sepsis. Her physical examination revealed mild hypotonia but was otherwise normal. Initial hematologic indices and serum chemistry were also normal except for mild metabolic acidosis. At 16 hours of age, she developed clinical seizures confirmed on continuous conventional electroencephalography. A 24-hour NP SARS-CoV-2 RT-PCR was negative, but SARS-CoV-2 IgG, combined (total) IgG, IgM, and IgA reactivity to a recombinant derivative of SARS-CoV-2 spike protein, inflammatory markers, and cytokines were markedly elevated (Table 1). There was CSF pleocytosis but no detectable virus or bacteria. Tests for inborn errors of metabolism were negative. Chromosomal microarray showed a copy number variant of uncertain clinical significance. Brain MRI on day 4 showed diffuse restricted diffusion with low apparent diffusion coefficient throughout the supratentorial brain and sparing of the basal ganglia and medial temporal lobes (Fig 1C). She was discharged at 5 weeks of age but has had multiple hospital readmissions for breakthrough seizures and respiratory infections. Repeat brain MRI at 12 weeks of age showed progressive development of severe cystic encephalomalacia (Fig 1D). Her most recent follow-up exam at 1 year of age showed microcephaly (head circumference < 3rd percentile compared with 59th percentile at birth), abnormal neurologic examination with low axial tone, head lag, increased appendicular tone, hyperreflexia, clonus, and significant neurodevelopmental delay with inability to roll over or sit unsupported. The infant is currently in hospice care.

The placentas from both cases showed thrombosis and recanalization of stem villous vessels, stromal fibrosis, and increased stromal karyorrhexis of terminal villi. These widespread changes correspond to high-grade fetal vascular malperfusion (Fig 2A–B).

Immunofluorescence was done to detect the presence of SARS-CoV-2 spike glycoprotein 1 (S1) and nucleocapsid protein in the placenta and case-1 infant’s brain, using Stochastic Optical Reconstruction Microscopy (Nikon Instruments, Inc. Melville, New York, USA). We detected S1 in both case placentas, colocalized with the nucleocapsid protein of the syncytiotrophoblast (Fig 2C). We also identified S1 protein, colocalized with nucleocapsid protein, throughout the brain of case-1 (Fig 1, E5–E8).

Both placentas were also analyzed for plausible inflammatory and modulating factors, which, directly or indirectly, adversely affect fetal central nervous system development, and compared with 2 age- and gender-matched placentas of SARS-CoV-2 negative mothers. Immunofluorescence was performed³  for hypoxia-inducible factor-α (HIF-α), a transcription factor responsive to cellular hypoxia; Nucleotide-binding oligomerization domain and leucine-rich repeat containing receptors pyrin domain containing 1 protein (NLRP1), which is formative in the assembly of inflammasomes; macrophage inflammatory protein 1 β (MIP-1β), a chemotactic cytokine; stromal cell-derived factor 1 (SDF1), an angiogenic and chemotactic chemokine; interleukin 13, an inflammatory cytokine; interleukin 10, an anti-inflammatory cytokine; and human chorionic gonadotropin (hCG), a marker of placental function. HIF-1α, NLRP1, MIP-1β, SDF-1, IL-13, interleukin 10 (IL-10) were all elevated (Fig 2D–O), whereas hCG was markedly decreased in the SARS-CoV-2 positive placentas, compared with controls (Fig 2P–Q).

Exposure to intrauterine inflammation and disrupted placental programming following SARS-CoV-2 in pregnancy can have long-term consequences on the fetus.⁴  To our knowledge, the 2 cases described here represent the first time that neonates born to SARS-CoV-2 positive mothers present with a neonatal clinical course mimicking hypoxic-ischemic encephalopathy of the newborn. There was no preceding clinical event in labor that heralded an acute hypoxic-ischemic event in either case, and the neurologic sequelae extended well beyond the neonatal period. Both placentas and brains showed evidence of SARS-CoV-2 infection and the placentas displayed alterations in inflammatory and oxidative stress markers.

Severe SARS-CoV-2 placental infection can impact neonatal outcome even in the absence of vertical transmission.^4–6  Although SARS-CoV-2 was absent in the nasopharyngeal swabs of our newborns, their elevated SARS-CoV-2 antibodies and proinflammatory mediators, as well as the neuropathology findings from the brain of Case 1, suggest brain injury either caused by the fetal inflammatory response to placental infection and/or the consequence of an undetected in-utero brain infection.

With respect to the timing of infection during pregnancy, both mothers initially tested positive for SARS-CoV-2 in the second trimester. However, both pregnancies continued well into the third trimester without any complications related to pregnancy or fetal compromise as could be assessed by antenatal monitoring. These observations, combined with exclusion of a lengthy list of other possible causes of the clinical and pathologic findings, implicate SARS-CoV-2 as the etiology of the neurologic injury.

The placental histologic findings of thrombosis, loss of stromal vessels, and apoptosis are associated with maternal-fetal vascular malperfusion, placental ischemia, and insufficiency, which result in deterioration of placental function.⁷  Consequences of chronic placental insufficiency may underlie the neonatal course mimicking hypoxic-ischemic encephalopathy of the newborn in the absence of a sentinel event during delivery.⁸ 

Successful pregnancy necessitates synchronized communications between the mother and fetus, and the immune system, and cytokine signaling pathways mediate these communications.⁹  Exposure to intrauterine inflammation and disrupted placental programming following SARS-CoV-2 in pregnancy can have long-term consequences on the fetus.¹⁰  Placental insults causing tissue hypoxia and oxidative stress induce release of proinflammatory mediators and cause endothelial damage, with resulting alterations in placental blood flow and fetal brain injury.¹¹ 

Inflammation of the feto-placental unit leads to development of the fetal inflammatory response syndrome, which causes fetal hypoxia, blood vessel damage, and blood-brain barrier compromise, and leads to white matter injury of the fetal brain.^12–14  In a proinflammatory state, cytokines directly cross the blood-brain barrier causing neuronal injury.^14–16  HIF1-α, a master regulator of oxygen homeostasis, responds to tissue hypoxia and produces placental vascular changes similar to those seen in our cases.¹⁷  Experimental studies show HIF-1α expression is increased in the placenta and developing mouse brain during acute systemic hypoxia.¹⁷  This suggests that elevated placental HIF-1α is an indirect tissue marker for fetal cerebral hypoxia. Oxidative stress, measured by NLRP1, may contribute to neuronal injury.^18–20  Elevated levels of placental proinflammatory NLRP1, MIP1-β, interleukin 13 (IL-13) are also known to impair fetal neurodevelopment.^21,22  Increased myometrial SDF1 recruits immune cells, that in turn produce inflammatory cytokines.²³  The placental hormone hCG plays a neuroprotective role against hypoxic-ischemic neurodegeneration in the developing brain.²⁴  Moreover, hCG promotes uterine angiogenesis and vasculogenesis to ensure adequate blood supply to the placenta.²⁵  In this light, decreased hCG levels suggest placental compromise. The constellation of placental pathology along with placental changes in HIF-1α, NLRP1, MIP1-β, SDF1, IL-13 and hCG suggest that placental SARS-CoV-2 infection triggered inflammatory and oxidative stress injury in the fetoplacental unit, which may have resulted in neurologic damage mimicking hypoxic perinatal brain injury.

We believe the extensive placental changes are sufficient to explain at least the initial neurologic presentation of both infants.^5,26  The autopsy identification of SARS-CoV-2 in the brain of case-1 raises the possibility that viral infection of the brain may have additionally directly contributed to the evolution of this infant’s neurologic injury. In retrospect, testing of amniotic fluid and cord blood samples for SARS-CoV-2 as well as uniformity in testing for inflammatory markers would have been helpful.

In conclusion, our cases demonstrate that midtrimester maternal SARS-CoV-2 infection can infect the placenta and fetal or infant brain and trigger a cascade of inflammatory events in both placenta and fetus. This may be associated with major brain injury and progressive neurologic sequelae in infants beyond the neonatal period. Our cases also highlight the shortcomings of current fetal monitoring for assessment of fetal well-being, especially when the target of injury is the fetal brain. Future longitudinal studies are needed to study the impact of timing of in utero SARS-CoV-2 infection on placental inflammation, as well as the long-term consequences on the developing brain.

COVID-19

coronavirus disease of 2019

CSF

cerebrospinal fluid

hCG

human chorionic gonadotropin

HIF-α

hypoxia-inducible factor-α

IL-10

interleukin 10

IL-13

interleukin 13

MIP1-β

macrophage inflammatory protein 1 βη

NLRP1

NLR pyrin domain containing 1 protein

S1

spike glycoprotein 1

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2

SDF1

stromal cell-derived factor 1