Context.—As the number of Zika virus (ZIKV) infections continues to grow, so, too, does the spectrum of recognized clinical disease, in both adult and congenital infections. Defining the tissue pathology associated with the various disease manifestations provides insight into pathogenesis and diagnosis, and potentially future prevention and treatment, of ZIKV infections.
Objective.—To summarize the syndromes and pathology associated with ZIKV infection, the implications of pathologic findings in the pathogenesis of ZIKV disease, and the use of pathology specimens for diagnosis of ZIKV infection.
Data Sources.—The major sources of information for this review were published articles obtained from PubMed and pathologic findings from cases submitted to the Infectious Diseases Pathology Branch at the Centers for Disease Control and Prevention.
Conclusions.—Pathologic findings associated with ZIKV infection are characteristic but not specific. In congenital Zika syndrome, tissue pathology is due to direct viral infection of neural structures, whereas in Guillain-Barré syndrome, pathology is likely due to a postviral, aberrant host-directed immune response. Both fetal and placental pathology specimens are useful for ZIKV diagnosis by molecular and immunohistochemical assays; however, the implications of ZIKV detection in placentas from second- and third-trimester normal live births are unclear, as the potential postnatal effects of late gestational exposure remain to be seen.
As the first infectious agent to be associated with an epidemic of birth defects in more than 50 years,1 Zika virus (ZIKV) has raised concern among parents and within communities worldwide.2 After its identification in the mid-20th century, it spread at first slowly and largely unnoticed through Africa and Asia, but is now rapidly spreading across the Americas, with active mosquito-borne transmission reported in more than 50 countries and territories as of August 1, 2016, and with travel-associated cases identified in other regions (Figure 1).3–10 As such, the World Health Organization11 has declared this pandemic a Public Health Emergency of International Concern, and pregnant women have been advised to avoid travel to affected regions.12
Zika virus is an RNA virus of the family Flaviviridae, and is closely related to other members of this family, including dengue, yellow fever, West Nile, and Japanese encephalitis viruses.13 It was first isolated from a sentinel rhesus macaque from the Zika forest of Uganda in 1947, and was largely unknown until its spread to Micronesia in 2007, when an outbreak of acute, self-limiting disease including primarily rash, fever, arthralgia, and conjunctivitis affected nearly 200 people.14 In 2013–2014, it gained more attention during an outbreak in French Polynesia affecting an estimate of approximately 28 000 people with a similar acute and self-limiting disease, but with a small minority of those also exhibiting more severe, neurologic signs.15–18 It wasn't until 2015, however, with its spread through the Pacific Islands to the Americas and its association with a large outbreak of severe congenital anomalies, particularly microcephaly, that Zika became a household name. As the geographic reach of the virus expands, so too does our understanding of the variety and severity of its clinical manifestations.
Understanding the disease biology of ZIKV is crucial for effective intervention.19 To this end, correlating clinical manifestations with underlying tissue pathology can provide critical insights into disease pathogenesis.20,21 Despite the massive amount of rapidly accruing literature about ZIKV, descriptions of pathology related to ZIKV infections have been limited.22,23 This is due largely to the fact that most infections are asymptomatic, and of the remainder, most have mild and self-limiting disease that does not warrant tissue sampling for histopathologic evaluation.14 Even for the small proportion of infections with severe clinical manifestations, pathologic changes in tissues may be mild or nonspecific, and not readily attributable to ZIKV infection.14–26 Here, we summarize the pathology for the major clinical syndromes heretofore associated with ZIKV infection, and discuss their implications for disease pathogenesis. Described pathologic findings include those from the current literature and from evaluation of cases submitted to the Infectious Diseases Pathology Branch at the US Centers for Disease Control and Prevention for diagnostic consultation, as part of a routine public health service provided by the Infectious Diseases Pathology Branch. As such, institutional review board review was not required for the testing described herein. Permission to present individual case information was obtained from the patients or their families.
ZIKA SYNDROMES AND PATHOLOGY
Although limited, available data suggest that approximately 80% of ZIKV infections are asymptomatic.14 When symptoms occur, they typically comprise a mild, self-limited denguelike syndrome of less than 2 weeks' duration.26 Low-grade fever, rash, arthralgias, and bilateral nonpurulent conjunctivitis occur most commonly, with retro-orbital pain, myalgias, lymphadenopathy, sore throat, aphthous ulcers, and genitourinary and gastrointestinal symptoms less commonly reported (Figures 2 and 3).14,27–32 The rash is usually pruritic, erythematous, and maculopapular, and spreads in a descending fashion from face to extremities (Figure 3, A through C).33,34 Histologically, the skin shows nonspecific changes of mild to moderate perivascular lymphocytic dermal inflammation (Figure 3, D).35 Arthralgias frequently involve the joints of the hands and feet, and may be accompanied by periarticular edema. Leukopenia, thrombocytopenia, and elevated liver enzymes may be present.36–38 Distinguishing Zika fever from dengue and chikungunya infections, which are common in these same geographic regions, is not possible clinically; however, ZIKV may be more commonly associated with peripheral edema and conjunctivitis, and less commonly with leukopenia and thrombocytopenia, than are these other viruses.39 Rarely, severe and even fatal ZIKV disease may occur, particularly in the settings of severe thrombocytopenia or underlying disease.40,41
Noncongenital Neurologic Syndromes
Neurologic sequelae to Zika fever have been recognized in a small proportion of cases. The ZIKV outbreak in French Polynesia in 2013–2014 included more than 8000 suspected infected patients, with 74 having neurologic symptoms and more than half of those being diagnosed with Guillain-Barré syndrome (GBS).17 Guillain-Barré syndrome is a postinfectious autoimmune polyradiculoneuropathy that occurs following an antecedent antigenic stimulus, most commonly a viral or bacterial infection. It is characterized by initial weakness and motor dysfunction with decreased or absent deep tendon reflexes, typically progressing from involvement of the legs to the upper extremities and trunk muscles and, in some cases, the cranial nerves.17,19,42 In French Polynesia, GBS symptoms including generalized muscle weakness, inability to walk, and facial palsy occurred in patients a median of 6 days after symptoms of Zika fever. Electrophysiologic studies in most patients showed acute motor axonal neuropathy, with distal nerve segment conduction disturbances that improved at 4-month follow-up. A case-control study identified neutralizing antibodies against ZIKV in 100% of GBS patients compared with only 56% of controls, providing strong evidence of an association between GBS and recent ZIKV infection.17,43
Concomitantly with the spread of ZIKV to the Americas, there has been an increase in observed GBS cases in Brazil, Colombia, El Salvador, Venezuela, Martinique, and Suriname.36,42–46 Although a causal association between ZIKV infection and GBS has not been clearly established, preliminary findings in the municipality of Salvador, Brazil, between May and July 2015 indicate a more than 4-fold increase over the expected incidence of GBS and a strong correlation with symptoms of Zika fever (more than 90% of patients) prior to GBS onset. Of 50 patients diagnosed with ZIKV-related GBS, all had flaccid weakness in the lower extremities, and more than half also had upper extremity or cranial nerve involvement. Three fatalities occurred; no autopsies were performed (James Sejvar, MD, written communication, July 6, 2016). Three ZIKV-associated GBS deaths have also been reported in Colombia.47 To our knowledge, there are no published reports to date on the histopathology of ZIKV-associated GBS; however, histopathologic evaluation in one case showed peripheral nerve demyelination and mononuclear inflammation with axonal degeneration (Figure 3, E through G). These findings are consistent with typical GBS histopathology and are thought to be due to a cross-reactive immune response to foreign (infectious agent) and host neural antigens, rather than a consequence of direct neural infection.48,49 Further histopathologic characterization and attempted ZIKV detection in neural tissues from ZIKV-associated GBS patients would provide additional insight into whether this same mechanism appears to be consistently involved in these cases, or if direct ZIKV infection of the peripheral nervous system may also play a role.
Neurologic sequelae other than GBS reportedly associated with ZIKV infection are rare and include acute myelitis and meningoencephalitis. Myelitis with cervical spinal cord edema and thoracic cord lesion were diagnosed by imaging studies in a 15-year-old girl with high ZIKV RNA load detected by reverse transcription–polymerase chain reaction (RT-PCR) in serum, urine, and cerebrospinal fluid.50 Meningoencephalitis was diagnosed by magnetic resonance imaging in an 81-year-old man with cerebrospinal fluid pleocytosis and RT-PCR and culture positive for ZIKV.51 Testing for other infectious agents was negative for both patients. Both recovered with treatment; the man's treatment regimen was not specified, but the girl's myelitis responded to corticosteroids, suggesting that, as with GBS, the lesion may have been due largely to an aberrant immune response rather than direct viral infection of the spinal cord.
Congenital Zika Syndrome
The rise in incidence of microcephaly months after the rise in ZIKV infections in Brazil alerted health officials to the potential association of ZIKV with birth defects.19 There are now estimates of up to 1.3 million ZIKV infections in Brazil, associated with 1434 cases of confirmed microcephaly or other congenital abnormalities, including 208 cases with laboratory evidence of ZIKV infection.45,52 Furthermore, retrospective analysis of the French Polynesia outbreak, including laboratory testing of amniotic fluid from pregnancies with abnormal cerebral ultrasounds, also showed a strong link between fetal neurologic abnormalities and ZIKV infection.53–55 Epidemiologic, clinical, and laboratory evidence for ZIKV causality has been established, and a broader spectrum of abnormalities has emerged, composing congenital Zika syndrome (CZS).1,56 According to the World Health Organization, in addition to microcephaly, CZS may include “craniofacial disproportion, spasticity, seizures, irritability and brainstem dysfunction including feeding difficulties, ocular abnormalities, and findings on neuroimaging such as calcifications, cortical disorders, and ventriculomegaly. . . . Genitourinary, cardiac, and digestive systems can also be affected.” 56 The majority of infants with CZS are born to mothers with symptomatic Zika fever in the first trimester24,45 ; in one study, up to 70% of mothers of microcephalic infants had rash in the first trimester.57
Microcephaly is defined as occipitofrontal head circumference 2 or more SD below the mean for sex and gestational age at birth, and is categorized as severe if head circumference is 3 or more SD below the mean (Figure 4, A and B).26,58 It may be detected by prenatal ultrasound; however, both false-negative and false-positive diagnoses may occur, particularly early in pregnancy.19,59 Ultrasound and/or autopsy may reveal widespread brain calcifications (periventricular, cerebral parenchyma, thalami, and basal ganglia), cortical thinning with ventricular dilation, simplified gyral patterns (eg, lissencephaly, pachygyria, agyria), and cerebellar hypoplasia. Thinning of the pons and brainstem, dysgenesis or agenesis of corpus callosum, and holoprosencephaly are also reported. The frontal and parietal lobes and lateral ventricles are predominantly affected, and calcifications are present in a bandlike periventricular or corticomedullary pattern (Figure 4, C). Extracranial abnormalities may also be seen (Figure 4, D). Histopathology of the brain consistently shows cortical thinning with neuronal apoptosis/necrosis and punctate to coarse, and sometimes neuron-shaped, calcifications in cortex and subcortical white matter (Figure 5, A through D). Periventricular and white matter rarefaction with macrophagic infiltrates, perivascular lymphocytic infiltrates, microglial proliferation and activation, and glial nodules are variably reported (Figure 5, E). Viral inclusions have not been documented.26,58,60–64 Spinal cord pathology is described for one case and includes Wallerian degeneration of the descending motor tracts, with sparing of dorsal tracts.62 Immunohistochemistry for ZIKV shows labeling of viral antigen in glial cells, endothelial cells, and degenerating cortical neurons and associated with calcifications (Figure 5, F and G).24,64 Transmission electron microscopy showed spherical 40-nm enveloped viral particles and viral factories within neurons in the brain of an aborted 22-week-gestation fetus with CZS, indicating active viral replication in these cells (Martines, 2016, manuscript in preparation).
Zika virus–associated microcephaly and other brain abnormalities are usually part of generalized intrauterine growth restriction, and are commonly accompanied by congenital contractures (arthrogryposis) due to lack of fetal movement secondary to neurologic abnormalities (Figure 4, D).22,26,34,60,64,65 Oligohydramnios or anhydramnios, hydrops fetalis, pulmonary hypoplasia, cryptorchidism, and fetal death are also reported.34,64,66 In surviving microcephalic infants, ocular abnormalities are also common, with alterations in the neural components of the eye occurring in up to 55% of infants in some reports.67,68 Macular retinal and optic nerve changes are most consistently reported on funduscopic examination, with macular pigment mottling, foveal reflex loss, chorioretinal atrophy, and optic nerve hypoplasia with optic disc cupping.67–70 Lens subluxation, bilateral iris coloboma, vascular abnormalities, cataract, intraocular calcifications, and microphthalmia are also reported.67,71 Histopathologic correlates of ocular changes are not thoroughly described in the literature to date, but demonstration of ZIKV antigen in the retina by immunohistochemistry indicates direct infection of ocular structures (Figure 5, H).
Placental pathologic lesions in CZS are frequently absent, and documented to date predominantly for maternal infection in the first trimester.24,57,64 When present, changes are mild and nonspecific and comprise chronic placentitis, with chronic villitis or increased Hofbauer cells and patchy perivillous fibrin and mononuclear cells. Villous immaturity, edema, hypervascularity, stromal fibrosis and calcification, focal syncytiotrophoblast necrosis, and mild lymphocytic deciduitis may also be present (Figure 6, A through C).24,64 Immunohistochemical (IHC) labeling of viral antigen in Hofbauer cells, fetal endothelium, and maternal leukocytes has been shown only in first-trimester placentas, despite viral RNA being detectable in placental tissue even from full-term pregnancies (Figures 2 and 6, D).24,64,72 Second- or third-trimester placentas negative by IHC and positive by RT-PCR for ZIKV RNA may have calcifications and perivillous fibrin deposition, but these features are indistinguishable from stillbirths due to other causes and otherwise normal late-term placentas.26,73
Zika virus likely exists in sylvatic and urban cycles, with monkey and human hosts, respectively, and the possibility of other, as yet unidentified, animal reservoirs. Transmission occurs primarily through the bite of infected Aedes mosquitoes (Aedes aegypti and Aedes albopictus), though it has also been documented through sexual contact and suspected through blood transfusion and animal bite (Figure 2).19,30,74–76 Zika virus RNA has been detected in semen, saliva, urine, breast milk, and donated whole blood.16,31,77,78 Of utmost concern given the current outbreak are maternal transmission routes, for which there is evidence of both transplacental and perinatal transmission, with ZIKV detection in amniotic fluid and placental and fetal tissues, and potential for postnatal transmission through breast milk.26,54,62–66,72,79,80
Pathology and other studies have shown that ZIKV's congenital neurologic effects are clearly correlated with direct fetal infection and neurotropism of the virus, as opposed to placental insufficiency or fetotoxic placental immune response resulting from placental infection.64,81 The primary potential routes of fetal infections in general include maternal viremia with transfer from maternal blood across the syncytiotrophoblast to fetal vessels in the placenta or direct transfer at the implantation site via extravillous trophoblasts, and ascending infection from the genitourinary tract with transfer across the fetal membranes.82 Immunohistochemical detection of ZIKV in maternal leukocytes, Hofbauer cells, and fetal endothelial cells, along with demonstration of ZIKV replication in primary human placental macrophages in vitro, suggests transfer across the syncytiotrophoblast as the major route of fetal ZIKV infection.24,64,83 Once ZIKV gains access to the fetal circulation, widespread dissemination is presumed, with subsequent viral replication and persistence in permissive tissues. In a review of 10 cases with congenital ZIKV infection, ZIKV was detected by IHC in a variety of embryonic tissues (brain, kidney, lung, liver, heart, skin) in pathology specimens from first-trimester embryos (n = 3), but was localized by IHC to only central nervous system (CNS) tissues in fetuses after the first trimester (n = 7) (Martines, 2016, manuscript in preparation). Though limited, these findings support initial ZIKV dissemination in the fetus, with specific tropism for, and persistent replication in, neural tissues.
In vitro and animal studies further demonstrate ZIKV neurotropism and CNS damage. ZIKV infects and replicates in human neural progenitor cells in vitro and in human stem cell culture systems that model 3-dimensional brain structure (neurospheres and cerebral organoids).84–86 Furthermore, infection results in morphologic and architectural alterations to these primitive brainlike structures that are reminiscent of those seen in the brains of ZIKV congenitally infected microcephalic infants, including cell death and thinning of the ventricular zone–like layer, with dilation of ventricular structures.85,86 Mouse models show CNS tropism and damage preferentially in young or immunodeficient animals.87–90 Infection of immunodeficient pregnant dams results in viral detection in placenta and fetal brain, sometimes with reduced brain size and cortical thinning, similar to CZS.91–94 Studies in infected pregnant monkeys are ongoing, and show preliminary ultrasonographic evidence of decreased fetal head circumference associated with maternal ZIKV infection.91
The timing of maternal infection (first trimester) associated with CZS and the preferential susceptibility of neural progenitor cells in vitro and the CNS of young and fetal animals in vivo suggest that the overall mechanism of neural damage relates to arrest of brain growth, rather than damage and collapse of preformed neural structures.95 More mature neurons in vitro, older experimental animals, later-gestational-age human fetuses, and adults appear more resistant to neural infection, suggesting that neural maturation confers resistance to ZIKV.84,86–88 Noncongenital neurologic syndromes may therefore be due to overactive or aberrant host-directed immune responses rather than direct viral infection of neural cells. That this is the classically understood mechanism behind GBS, the most common noncongenital neurologic ZIKV-associated syndrome, supports this hypothesis. Further, ZIKV and human peptide commonalities have been found, indicating potential targets for such an immune response.96
ZIKV is unique among flaviviruses, and among arboviruses in general, in its strong predilection for maternofetal transmission and adverse fetal outcomes. Microcephaly specifically has not been reported for other flaviviral infections during pregnancy.11 Dengue virus infection is associated with much less severe adverse outcomes (preterm birth and low birth weight), if any at all, and these outcomes may most often be due to placental infection and insufficiency rather than vertical transfer, though congenital infection with infant death has been documented.80,97 Destructive brain lesions in one case of congenital West Nile virus infection were reported.98 Severe outcomes, including fetal encephalopathy and miscarriage, are occasionally reported with congenital chikungunya virus infection.99,100 In dengue, poor outcomes have been linked to prior infection with another dengue virus serotype, potentially through antibody-dependent enhancement of infection.101,102 It remains to be determined whether prior or concurrent dengue virus infection, which is likely in ZIKV-affected areas, might also predispose to GBS or congenital infection in patients with ZIKV infection.11,43,101,102
Characterizing and monitoring the ZIKV pandemic requires the ability to recognize ZIKV-associated clinical syndromes and accurately attribute them to ZIKV infection, which can be difficult. Zika fever shares an overlapping clinical syndrome with other cocirculating viruses, particularly dengue and chikungunya viruses, and the features of CZS—namely microcephaly—can also be associated with many other infectious (eg, TORCH agents: Toxoplasma, syphilis, varicella-zoster, rubella, cytomegalovirus [CMV], and herpes simplex viruses) and noninfectious (eg, genetic, metabolic, environmental, and toxic) causes.58 Coinfections are also possible.103–105 The neuropathology of early congenital infection is generally similar (brain growth arrest with the constellation of abnormalities as described for CZS), regardless of the infection(s) responsible; however, subtle differences may help to guide the differential diagnosis.106 In addition to ZIKV, microcephaly is particularly common with congenital CMV, Toxoplasma, rubella, and herpes simplex virus infections; calcifications are most common with CMV and Toxoplasma, and Toxoplasma frequently has a nodular distribution of lesions and hydrocephaly.31,61,75 Ocular abnormalities are also common in CMV and Toxoplasma.26,70 Although ZIKV appears to have a more specific tropism for, and marked potential for destruction of, the CNS, TORCH infections may more frequently be associated with extraneural abnormalities.66,71 Rubella, CMV, and herpes simplex virus infections may result in auditory deficits, and rubella commonly has cardiac anomalies.26,57,107 The full spectrum of CZS has not yet been described, and additional overlapping, or potentially distinguishing, features may emerge. Histopathologic findings in CNS and placenta can be nonspecific for all congenital infections, and as reported in some ZIKV cases thus far, congenital rubella infections may also have a notable lack of inflammation in tissues.64,108
Because clinical and pathologic features of ZIKV are nonspecific, laboratory confirmation of ZIKV and exclusion of other viruses are equally imperative; however, this can also be challenging because of cross-reactivity of the more readily available immunoglobulin M enzyme-linked immunosorbent assays with other flaviviruses. Additional confirmatory testing (plaque reduction neutralization test) is needed to distinguish ZIKV-specific antibodies, but this testing can only be performed in select reference laboratories, and interpretation can also be challenging in patients with prior flaviviral exposure.8,109 Zika virus can be distinguished by viral isolation or RT-PCR with sequencing performed on serum, plasma, whole blood, amniotic fluid, urine, saliva, semen, and tissues. Although viremia typically lasts for less than 1 week, it has been documented at up to 10 weeks after symptom onset in pregnant women.8,27,65,110 Furthermore, RT-PCR positivity has been documented up to 2 months after symptom onset in whole blood, up to 26 days in urine, and up to 18 weeks in amniotic fluid.65,78,111 For pathology specimens specifically, RT-PCR can be performed on fresh or formalin-fixed, paraffin-embedded fetal and placental tissues, and viral antigen can be demonstrated by IHC in formalin-fixed, paraffin-embedded tissues.64,72,112
Definitively confirming or excluding ZIKV infection in surviving neonatal infants can be challenging because testing at the time of delivery may be beyond the window of reliable detection by both serology and RT-PCR assays, especially for first-trimester maternal infections. Testing of full-term placental tissues from surviving infants by RT-PCR can be used to support maternal and fetal serology results. However, testing of placental tissue cannot definitively confirm or exclude infant infection specifically, because positive results could be due to persistent viral RNA in maternal and/or fetal portions of the sample, and negative results could be due to testing outside the window of detection or nonrepresentative sampling. Though not confirmatory for congenital ZIKV infection in asymptomatic neonates at the time of birth, positive RT-PCR results in placental tissues might prove to be prognostically useful if a correlation with late-onset neurologic disease is identified in the future.
The ultimate scale and impact of the ZIKV pandemic remain to be seen.113 It is likely that microcephaly and other severe abnormalities recognized at birth represent only the severe end of a spectrum of neurologic, and potentially also extraneural, manifestations of congenital ZIKV infection.11,71 Similar to other maternal infections (eg, TORCH agents) and fetotoxic insults, exposure in the latter half of pregnancy, after cortical organization is largely completed, may result in less overt neurologic or other impairments that may be detected only later in life.95 Long-term follow-up of infants with maternal ZIKV infection, regardless of whether or not they show signs of CZS at birth, is therefore crucial to fully characterizing this syndrome (Figure 2). Further investigations to elucidate the specific mechanisms of placental tropism and neurotropism, and to evaluate for potential viral factors or cofactors that may potentiate the effects of congenital ZIKV infection, are also critical to fully understanding CZS.11,19 The continued development and refinement of animal models will undoubtedly be crucial in these endeavors and in therapeutic development efforts.
Alongside research efforts, controlling the virus and preventing its further geographic spread is of immediate importance. Besides individual avoidance of mosquitos and non–vector-borne exposure, improvements in systematic mosquito population control and development of a vaccine for ZIKV may hold the most promise for extinguishing the pandemic. We must simultaneously begin the tasks of addressing the short-term and long-term care needs of children with CZS and its lifelong impairments and the accompanying social and economic impacts of this disease.
We thank Tadaki Suzuki, MD, PhD, Department of Pathology, National Institute of Infectious Diseases, Tokyo, Japan, for technical assistance with the figures. We thank Cynthia Moore, MD, PhD, Division of Congenital and Developmental Disorders, Centers for Disease Control and Prevention, for assistance in obtaining clinical photographs.
The authors have no relevant financial interest in the products or companies described in this article.
The findings and conclusions herein are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.