Adverse Outcomes of Pregnancy Associated Zika Virus Infection

Abstract

The spread of Zika virus to the Americas was accompanied by surge in the number of infants with CNS abnormalities leading to a declaration of a health emergency by the WHO. This was accompanied by significant responses from governmental health agencies in the US and Europe that resulted in significant new information the described the natural history of this perinatal infection in a very short period of time. Although much has been learned about Zika virus infection during pregnancy, limitations of current diagnostics and the challenges for accurate serologic diagnosis of acute Zika virus infection has restricted our understanding of the natural history of this perinatal infection to infants born to women with clinical disease during pregnancy and to Zika exposed infants with obvious clinical stigmata of disease. Thus, the spectrum of disease in infants exposed to Zika virus during pregnancy remains to be defined. In contrast, observations in informative animal models of Zika virus infections have provided rational pathways for vaccine development and existing antiviral drug development programs for other flaviviruses have resulted in accelerated development for potential antiviral therapies. This brief review will highlight some of the current concepts of the natural history of Zika virus during pregnancy.

Introduction

In late 2014, cases of an acute onset and self-limited, exanthematous illness characterized by arthralgias, pruritus, and less commonly, conjunctivitis were noted in Northeast Brazil in areas where Dengue virus (DENV) and Chikungunya virus (CHIKV) infections were endemic ^1–3. Shortly thereafter in the spring of 2015, the Brazilian Ministry of Health reported that Zika virus (ZIKV) was circulating in this region and in other regions of Brazil and potentially responsible for clusters of this newly recognized acute exanthematous illness ^1,3. Following these reports, a number of centers in Northeast Brazil reported ZIKV activity and in one report, attack rates of clinical illnesses consistent with ZIKV infection were reported to be as high as 8.2/1000 in children to as low as 3.8/1000 in adults ³. Initially, the emergence of ZIKV infection in Brazil was also associated with an increase in cases of a Guillian-Barre like syndrome; however, by late summer in 2015 an increase in the number of infants born with microcephaly, including severely microcephalic infants (>3SD below mean head circumference for age) with unusually severe decreased growth and development of the cerebrum, and redundant scalp skin ¹. Early case studies strongly suggested a link between ZIKV infection during early pregnancy and the delivery of an infant with severe microcephaly and neurological deficits ⁴. In November of 2015, the Brazilian Ministry of Health, the Pan American Health Organization issued alerts about the possible association between ZIKV infections in pregnancy and microcephaly in the offspring of affected pregnancies and in early 2016, the World Health Organization (WHO) declared ZIKV a health emergency. Amidst a flurry of international conferences and meetings, several large natural history studies of ZIKV infections were organized by the NIH and CDC to gather sufficient number of cases and controls to definitively address questions of causality of neurologic damage in offspring of women infected with ZIKV during pregnancy. Importantly, these studies were sufficiently powered to comprehensively define the spectrum of abnormalities in infants infected in-utero. These studies continue and with expected enrollment, should address key questions in the natural history of ZIKV during pregnancy in countries in South America and the Caribbean. In this brief overview, we will provide some of the key epidemiological characteristics of the ZIKV outbreak in South America and the Caribbean, the spectrum of disease in the infant infected in-utero, potential mechanisms of disease in the infected fetus, and the current approach to diagnosis of this infection in pregnant mothers and their offspring.

Epidemiological Observations of the ZIKV outbreak in Brazil and northern South America

ZIKV was first isolated in 1947 from a Rhesus macaque monkey that served as a sentinel monkey in a Rockefeller Foundation supported field station that was established for studies of Yellow Fever in the Zika forest in Gambia ⁵. Subsequently, additional ZIKV isolates were collected from mosquitos in the same area. In 1954, the first ZIKV infection in humans was reported in a patient in Nigeria ⁶. There appeared to be only limited ZIKV activity outside of Africa such that the first case of human ZIKV infection occurring outside of Africa was described in a patient in Indonesia ⁷. Serological studies carried out in the 1950s and 1960s suggested that ZIKV was endemic in all of Africa as well as in several Asian countries, although the quantifying the seroprevalence of ZIKV infection is problematic because of the cross reactivity between antibodies to other Flaviviruses and ZIKV. This remains an ongoing technical issue in the serodiagnosis of ZIKV infections in areas of the world where other Flavivirus infections are endemic ². Definitive information detailing the emergence of ZIKV infection in a population was first reported from studies carried out in Yap, an island in Micronesia located in the Western Pacific ². Although DENV infection was initially suspected as the etiology of this acute febrile illness associated with rash and arthralgias, local physicians viewed the presenting symptomatology of this infection as atypical for DENV infections and studies carried out on specimens sent to the CDC demonstrated ZIKV in about 14% of those specimens tested by PCR ^8,9. It was estimated that over 70% of the population was infected with ZIKV during the 3 month period of the outbreak, and of ZIKV infected individuals, about 18% exhibited clinical symptoms compatible with ZIKV infection ^8,9. Subsequently, in 2013 in French Polynesia, an outbreak of ZIKV resulted in an estimated 39,000 cases of ZIKV infection, although this is likely an underestimate of the number of cases as many infected individuals with mild or asymptomatic infections were not tested for ZIKV infection ^2,10,11. Interestingly the duration of this outbreak was also finite and reported to be about 21 weeks ². After this outbreak, ZIKV infections were reported throughout the Pacific Islands, including the Chilean Easter Islands ¹⁰. The origin(s) of ZIKV that was associated with the initial outbreak of ZIKV infections in Pacific Islands are uncertain but phylogenetic analysis of isolates from patients infected with ZIKV in YAP revealed a lineage that was most consistent with ZIKV isolates from Cambodia ¹².

Although the first descriptions of the acute exanthematous illness in Brazil in 2014 were likely ZIKV infections, it was not until the spring of 2015 when the first case of ZIKV infection was confirmed in the northeastern Brazilian state of Bahia ¹. Subsequently, ZIKV infections were reported in almost all states in Brazil and by December 2015, it was estimated that between 500,000-1,000,000 people had been infected ¹³. In late 2015, ZIKV infections were confirmed in Colombia and in the Caribbean, including Puerto Rico (Pan American Health Organization. 2015. Epidemiological update. Zika virus infection. 16 October 2015. Pan American Health Organization, Washington, DC). Although the source of ZIKV that resulted in the Brazilian outbreak has not been definitively identified, some authorities suggest that ZIKV was imported into Brazil following international sporting competitions in 2014 that included participating teams from Polynesia ^14–18. Brazil represented an ideal environment for the emergence of ZIKV as both species of mosquito vectors, Aedes aegypti and Aedes albopictus were present throughout Brazil and mosquito control was limited in many areas of the Northeast of the country ¹⁷. Shortly after its emergence in Brazil, this newly introduced arbovirus spread rapidly through South America and the Caribbean. Imported cases of ZIKV infection were also reported in several European countries and the US in travelers returning from endemic areas ¹⁹. ZIKV infections continue to be reported in Puerto Rico, albeit at a lower rate than during 2016, and 3 cases of ZIKV infection presumably acquired through mosquito exposure have occurred in the US mainland ²⁰ (cdc.gov/zika/reporting/Nov2017). In addition, cases of presumed sexually transmitted ZIKV virus infection have been reported in the US ^21,22.

Significant interest developed in ZIKV as a cause of birth defects following reports from Brazil of the sudden increase in infants born with microcephaly, many with seemingly stereotypic cranial abnormalities not commonly seen following other causes of microcephaly, including typical findings associated with congenital infections. The spike in the incidence of microcephaly was first reported by health care workers in the early fall 2015 in the Northeast region of Brazil and by November of 2015 the Brazilian Ministry of Health suggested a possible association between ZIKV infections secondary to an estimated 20 fold increase in the incidence of microcephaly in this region²³, ¹³. Shortly thereafter a number of governmental health agencies including the Pan American Health Organization, the CDC, and the European Center for Disease Prevention and Control also posted alerts describing the association between ZIKV infection during pregnancy and microcephaly. Since this time there has been considerable effort by several health agencies to organize natural history studies of ZIKV infections during pregnancy to define and quantify relationship(s) between ZIKV during pregnancy and adverse outcomes of pregnancy. The goals of these studies is to define the spectrum of maternal disease and outcomes of pregnancy following ZIKV infection as well as long term outcomes of infants infected with ZIKV in-utero or during early infancy.

ZIKV Infections During Pregnancy: Clinical Presentations

ZIKV infections during pregnancy have been estimated to result in a symptomatic exanthematous illness in about 20% of women based on studies of the ZIKV outbreak in YAP and in French Polynesia ^8,24,25. Symptoms most frequently described have included a pruritic rash, fever, arthralgias, conjunctivitis, and headache. It is important to note that based on reports of the outbreak in the Pacific islands and in South America, the clinical course of ZIKV infections during pregnancy does not appear to differ significantly from that described in non-pregnant women. However, the impact of ZIKV on the outcomes of pregnancy cannot be overstated.

The association between ZIKV infection during pregnancy and the increase in the incidence of microcephaly in infants born in northeast Brazil surfaced in both the scientific and lay press in late 2015. However, it is of interest that a retrospective analysis of birth defects in infants born following the ZIKV outbreak in French Polynesia between 2013-2014 revealed several unique abnormalities in brain development in offspring born to women exposed to ZIKV during pregnancy ²⁶. These authors reported that approximately 8,700 cases of suspected ZIKV infections occurred in French Polynesia between 2013-2014 and that during the following year an increase in the number of infants with brain malformations were reported ²⁶. A retrospective analysis of the outcome of 4,787 pregnancies during this ZIKV outbreak described 33 cases of brain malformations or symptomatology ascribed to CNS damage ²⁶. Of these, 19 were reported with atypical congenital cerebral malformations that included microcephaly with loss of brain parenchyma in 5 infants in which ZIKV RNA was demonstrated in stored amniotic fluid from 4/5 of these pregnancies ²⁶. In addition, three additional infants with microcephaly and imaging findings of parenchymal brain disease with significantly altered neurodevelopment were described, but samples required for laboratory diagnosis of ZIKV were not available ²⁶. Perhaps most interestingly, 6 newborn had normal head circumferences but abnormal head imaging that included ventriculomegaly and loss of corpus callosum, finding that suggested a wide spectrum of CNS damage could be associated with presumed ZIKV infection during pregnancy ²⁶. Finally, a group of 5 newborns in this cohort exhibited functional evidence of brainstem dysfunction ²⁶. Although retrospective in design, it is important to note that many of the newborn findings in this cohort of infants were consistent with the early case reports and clinical findings in affected infants in larger clinical series that have reported in studies from Brazil ²⁷. Thus, shared clinical manifestations of CNS damage in infants born in French Polynesia between 2014-2015 and infants born in Brazil in 2015-2016 correlated with outbreaks of ZIKV in these previously unexposed populations and strongly argued for a direct linkage between CNS damage in infants born to women exposed to ZIKV infection during pregnancy.

Initial case reports from Brazil that described cases of microcephaly in newborn infants that were associated with cranial malformations that included findings previously described in infants with the fetal brain disruption sequence (FBDS), severe microcephaly with near absence of posterior fossa, redundant scalp skin secondary to microcephaly and loss of brain parenchyma, and skull collapse ²⁸. From the early case reports and from larger series, the terminology of congenital Zika syndrome (CZS) was proposed to identify infants exposed in-utero to ZIKV ²⁸. These authors proposed five criteria that were characteristic of severe congenital ZIKV infection and included; (1) severe microcephaly with partial collapse of the skull, (2) thin cerebral cortices with subcortical calcifications, (3) macular scars with pigmented retinal findings, (4) congenital contractures, and (5) early hypertonia with evidence of extrapyramidal involvement ²⁸. These findings are notable because they are rarely if ever seen in other congenital infections. Other authors have also described abnormal ultrasound findings in the fetus including cerebral atrophy, ventriculomegaly, cerebellar hypoplasia, and arthrogryposis ^29,30. Congenital arthrogryposis has been described by several authors and represents a presentation unique to congenital ZIKV infections when viewed in the context of clinical findings of other congenital infections ³¹. Fetal and postnatal imaging of infants from this same case series also demonstrated loss of brain parenchyma, abnormal cortical development ranging from disorganization of the gyri to lissencephaly, subcortical calcifications, and abnormalities in the cerebellum, brain stem and basal ganglia ²⁹. Significant abnormalities in hindbrain structures included brain stem calcifications, reduced size of brainstems, loss of pons, and enlargement of 4^th ventricle ³². Interestingly, the subcortical calcifications, an imaging finding thought to be characteristic of ZIKV infection, slowly resolved in several infants without documented improvement in their neurological status ³³. Abnormal findings in the eyes from infants with presumed congenital ZIKV infections are frequent on the order of 30-70% and perhaps are the most common clinical finding in infants born to women infected with ZIKV during pregnancy ^34,35. Commonly reported ocular findings include pigmented retinopathy that is often focal, chorioretinal atrophy, and optic disc hypoplasia ^34,36–40. Anterior or posterior uveitis has not been described in most series. The finding of chorioretinal atrophy has been argued to be unique in congenital ZIKV. An abbreviated summary of clinical and imaging findings of infants born to women exposed to ZIKV during pregnancy is provided in Table 1.

Table 1.

Frequently reported findings in infants born to women exposed to ZIKV during pregnancy

Clinical Findings
Head	microcephaly; redundant scalp skin; biparietal depression with prominent occiput; overlapping sutures; characteristics of fetal brain disruption sequence 26,28,141–145.
Neuromuscular	arthrogryposis; hypertonia, hypotonia; clonus, hyperreflexia; aspiration syndromes; swallowing difficulties 28–31,142,146–148.
Visual	microphthalmia; macular chorioretinal atrophy; chorioretinal atrophy; pigmented retinopathy; optic disc atrophy 34,35,37,39,149–155.
Auditory	sensorineural hearing loss 156.
Other	craniofacial disproportion; failure to thrive 28.
Imaging Findings
Brain and spinal cord	loss of cortical parenchyma (cortical thinning); ventriculomegaly; polymicrogyria; cerebellar hypoplasia; cerebral cortical dysplasia;lissencephaly; absence of corpus callosum; subcortical calcifications; cortical migration abnormalities, brain stem calcifications; basal ganglia calcifications; brain stem hypoplasia; reduced thickness of spinal cord 28,29,33,46,47,157–172.

Finally, it should be noted that the term congenital Zika syndrome (CZS) represents useful terminology for description of the most severely affected infants; however, it fails to define spectrum of disease associated with ZIKV infection in-utero much like the term Cytomegalic Inclusion Disease used to describe clinical findings associated with perhaps 5% of infants with congenital cytomegalovirus infections ⁴¹. Consistent with the likelihood of an expanded spectrum of findings in infants with congenital ZIKV infections is the report of abnormal ophthalmologic findings in 8/24 (33%) of infants without CNS findings who were born to women with confirmed ZIKV infections during pregnancy ³⁵. Thus, it is an almost certainty that the spectrum of disease associated with infants infected in-utero and perhaps in the perinatal period with ZIKV remains to be fully defined ⁴². Thus, a more accurate and inclusive term such as congenital ZIKV infection should be considered.

ZIKV Infections During Pregnancy: Risks of Fetal Infection and Disease.

Limited data is available that quantifies the absolute risk of intrauterine transmission of ZIKV following infection during pregnancy. Similarly, the impact of the gestational age at the time of intrauterine transmission of ZIKV and magnitude of ZIKV replication in the fetus on the phenotypes of CNS damage have been estimated but not quantified. Several explanations likely account for these large gaps in our understanding of natural history of maternal and congenital ZIKV infections and include the lack of reliable diagnostic assays for ZIKV infections, the limited precision of case definitions of maternal ZIKV infection, particularly in areas with ongoing arbovirus activity, and finally, the current lack of understanding of the full spectrum disease associated with congenital ZIKV infections. Yet in the presence of these limitations, several studies have provided solid estimates of the course of ZIKV infections in pregnant women and their offspring, two of which enrolled a sufficient number of subjects to allow extension of the primary data. In the largest reported prospective study carried out during the ZIKV outbreak in the Americas, investigators in Brazil enrolled pregnant women with new onset rash between September 2015 and May 2016 and followed them through pregnancy ²⁷. Upon enrollment, acute ZIKV infection was defined by qualitative nucleic acid testing (RT-PCR) of specimens of blood and urine. Of 345 women enrolled, 182 (53%) were diagnosed as having an acute ZIKV infection and of these 134 were followed through pregnancy ²⁷. This study provided several important results including; (i) a descending maculopapular rash, conjunctival injection, and myalgias were more frequently reported in ZIKV infected women than in controls, some of whom were infected with other flaviviruses, (ii) fetal losses were similar in both ZIKV infected and control patients, perhaps secondary to the presence of acute chikungunya virus infections in the control cohort, (iii) adverse outcomes of pregnancy were significantly higher in women with acute ZIKV infections compared to controls (46.4% vs. 11.5%), and of the live born infants from the ZIKV infected group, 49/117 (42%) were found to have abnormal clinical and/or imagining exams, and (iv) adverse outcomes occurred regardless of the trimester of infection with 55%, 52%, and 29% of the adverse outcomes following first, second, or third trimester ZIKV infections respectively ²⁷. Importantly, of the infants with abnormal clinical exams in the first month of life, most had abnormalities in the CNS that included cerebral calcifications, ventriculomegaly, hypoplasia of different regions of the brain, and some 60% of these infants had grossly abnormal neurologic exams ²⁷. Consistent with other reports, these investigators noted that disproportionate microcephaly was only observed in 4/117 (3.4%) of infants and that these infants were born to women with acute ZIKV infection in the first trimester of pregnancy ²⁷. This important study illustrates several critical issues surrounding our current understanding of the impact of maternal ZIKV infection on the outcomes of pregnancy including the lack of definitive data on the absolute risk of intrauterine ZIKV transmission because in this study, only women with symptomatic infection were enrolled. Furthermore, even though the diagnosis of ZIKV was made by detection of ZIKV RNA, the outcomes of pregnancy were not stratified as a function of viral load and/or duration of viremia/viruria in the ZIKV infected cohort. Thus, the risk of transmission and adverse outcomes following ZIKV infection in pregnancy could be overestimated in this study secondary to an enrollment bias, suggesting that the true risk of intrauterine transmission and adverse outcomes following ZIKV infection during pregnancy remain to be defined. However, this study demonstrated the importance of ZIKV infection as an etiology of adverse outcomes of pregnancy and clearly demonstrated that ZIKV infection acquired at nearly any time during pregnancy could result in severe damage to the CNS of the developing fetus. Finally, the adverse outcomes of pregnancy in this study did not document ZIKV infection in affected infants raising the possibility that mechanisms of disease other than direct ZIKV mediated damage of the developing fetus such as ZIKV mediated placental dysfunction could have contributed to the clinical findings in these infants.

A second study that was performed by the Centers for Disease Control involved follow-up of 1297 pregnancies with possible recent ZIKV infection and/or exposure that were identified through a CDC sponsored ZIKV registry. Of these, 972 women completed pregnancy and birth defects were reported in 51 (5%) of fetuses or infants of women with possible recent ZIKV infection during pregnancy ⁴³. When pregnancy outcomes of women with confirmed ZIKV infection during the first trimester of pregnancy were analyzed, 9/60 (15%) of fetuses or infants had birth defects ⁴³. In pregnant women with ZIKV infection confirmed by nucleic acid amplification or serology, the rate of birth defects increased to 10% (24/250) ⁴³. Among the 24 fetuses and infants with birth defects, 18 (75%) had brain abnormalities or microcephaly ⁴³. These results contribute additional data to our understanding of the natural history of ZIKV during pregnancy and although this study population likely suffers selection bias as well the assignment of cases based on current ZIKV diagnostics, the estimates of impact of ZIKV infection on the outcome of pregnancy, particularly as a cause of CNS damage in the fetus and newborn infant, is consistent with that described by Brasil ²⁷.

ZIKV Infections During Pregnancy: Clues to Mechanisms of Disease in the Developing Fetus.

Early descriptions of the severe structural abnormalities of the brain, particularly the cerebrum, in infants born to women with ZIKV infections during pregnancy suggested that ZIKV infections associated with severe microcephaly occurred early in pregnancy during cortical neurogenesis and likely followed lytic infection of neural stem cells and neuroprogenitor cells. Loss of these cell types early in the developmental program of the cortex could be expected to lead to the findings described in several case series that included significant loss of brain parenchyma as well as a loss of intracranial pressure resulting in overlapping sutures with collapse of the skull. These findings are also described in the fetal brain disruption sequence (FBDS), a clinical description of pathologic findings in infants with severe malformation of the cerebrum from a variety of etiologies, including intrauterine infections ^44,45. In addition, findings of polymicrogyria and lissenencephaly in some infants were also consistent with an insult to the brain early in development, a mechanism that is consistent with the higher incidence of clinical findings of severe microcephaly in infants born to women with ZIKV infections in the first trimester ^26,27. In some cases, infants with findings of severe microcephaly exhibited decreasing head circumferences in-utero suggesting ongoing destruction of the brain parenchyma ⁴⁶. Other findings in ZIKV infected infants included cerebellar hypoplasia, dystrophic calcifications, including subcortical calcifications, hypoplasia of the corpus callosum, and abnormalities in the basal ganglia and brainstem ^26,27,32. Several of these later findings have been described in infants born to women infected in the second and third trimesters of pregnancy.

Clinical finding of congenital arthrogryposis as well as isolated joint contractures have been described in 5-10% of infants born to women infected with ZIKV during pregnancy ³². As noted above, this clinical finding has not been well described with other congenital infections, and thus represents a clinical presentation unique to this congenital infection, perhaps secondary to the loss of motor neurons and/or connections between the motor cortex and motor neurons. Abnormal imaging of the spinal cord and brainstem has been reported in these infants and abnormal EMGs have been reported in a small number of infants with arthrogryposis ^31,47. Examination of the spinal cords from two infants with congenital arthrogryposis associated with maternal ZIKV infection revealed damage to the cord, including neuronal loss and calcifications ⁴⁸. Similarly, loss of corticospinal tracts, loss of motor neurons, and gliosis were reported in another autopsy series ³². These authors made an intriguing observation that in one case, damage was limited to neurons derived from the neural tube and not neurons in the dorsal root raising the possibility of distinct neurotropism for ZIKV; however, more recent studies in experimental models have demonstrated ZIKV infection of peripheral neurons, including neurons of the DRG and of neural crest origin ^32,49,50. Finally, small animal models of ZIKV infection of the developing CNS have also demonstrated damage to the spinal cord and arthrogryposis ⁵¹.

Lastly, in addition to the myriad of clinical presentations that have been describe in infants born to women exposed to ZIKV during pregnancy, fetal loss has also been frequently associated in women infected with ZIKV during pregnancy. The rate of fetal loss following ZIKV exposure during pregnancy has not been defined definitively secondary to a number of factors including inaccurate case ascertainment and confounding obstetric conditions in maternal populations, including infections with other arboviruses. In one of the larger case series, Brasil reported fetal loss in 9/125 (7.2%) ZIKV virus infected women and 4/61 (6.6%) in the control, ZIKV negative cohort ²⁷. Interestingly, in both cohorts, about one-half of the fetal losses occurred during the first trimester and in the ZIKV negative control group, two of the women were infected with Chikungunya virus ²⁷. Thus, ZIKV virus infection during pregnancy appears to be associated with fetal loss but to date, has not been rigorously quantified in the maternal population from areas with increased ZIKV activity.

Mechanisms of Damage to the Developing CNS in the ZIKV Infected Fetus

As noted previously, histopathologic studies of CNS disease from fetuses and newborn infants have provided some insight into potential mechanisms of CNS damage associated with ZIKV infection ^32,52–56. Although these studies have for the most part described limited numbers of patients who were often enrolled secondary to clinically apparent manifestations of ZIKV infections and importantly, have likely described cummulative histopathologic features of CNS damage that could be distant from the acute infection, each of these studies point to direct ZIKV induced destructive lesions of neuronal tissue as a primary mechanism of disease. Specific findings have included mild lymphohistiocytic infiltration, reactive astrocytosis, multifocal dystrophic calcifications, and apoptosis, including activated caspase-3 positive cells ^32,53. The presence of inflammatory infiltrates in the brains of ZIKV infected fetuses and newborns has raised the possibility that host derived inflammatory responses could also contribute to CNS damage ^57–59. ZIKV RNA has been detected in neuronal cells as well as infiltrating inflammatory cells in the brains of fetus and infants infected in-utero, although in some series ZIKV was detected in tissues from a limited number of patients ^58,59. Importantly, detection of ZIKV in tissues specimens utilizing immunochemistry to detect ZIKV antigens has been reported by several investigators to result in non-specific findings and more definitive evidence of ZIKV has been obtained with in-situ hybridization for the detection of ZIKV RNA.

Studies in experimental animal models including rhesus macaques and mice have clearly demonstrated the capacity of ZIKV to infect neural stem cells and neuroprogenitor cells and have recapitulated damage to the developing CNS in both rodents and non-human primates following intrauterine infection ^60–64. Consistent with in-vivo evidence of neurovirulence of ZIKV, in-vitro models have demonstrated infection of neural stem cells and neuroprogenitor cells with disruption of cellular physiology and cell death. A recent report described increased susceptibility of committed postmitotic neuroprogenitor cells to ZIKV infection suggesting that susceptibility to CNS infection may increase with increasing gestation ⁶⁵. Specific mechanisms of disease following ZIKV infection have included histopathologic evidence of infected cell apoptosis and from in-vitro studies, the induction of signaling, and loss of cell-cell communications. The in-vivo neurovirulence of ZIKV was initially suggested to be secondary to the use of the AXL cellular protein as a ZIKV specific receptor ^66,67. Because AXL is robustly expressed on neural stem cells and neuroprogenitor cells, ZIKV could readily target cells of neuronal origin. However, recent studies have demonstrated that AXL is not required for ZIKV infection of susceptible neural stem cells ⁶⁸. Although animal models and in-vitro systems have provided consistent findings that ZIKV infection of the developing CNS can result in a destruction of neuronal tissue, the extent of tissue damage in some cases also has prompted consideration of other mechanisms of damage including loss of vasculature leading to ischemic damage to the CNS and immunopathology from host inflammatory responses ³². In brain organoids, ZIKV virus infection induced expression of TLR3, a pathogen-associated molecular pattern receptor, and the authors of this study suggested that inflammatory responses following ZIKV infection could be a mechanism of neuronal cell damage and loss during ZIKV infection ⁶⁹. Similarly, studies in animal models of ZIKV infections and observations in infants with ZIKV infections have provided mechanism that could be independent of the loss of neuronal cells and have included infection of vascular endothelium in the brain and retina ^70–72. Finally, in-vitro studies have suggested that ZIKV infection can induce premature differentiation of neuroprogenitors cells thus accounting for smaller brains and microcephaly in ZIKV infected infants ⁷³.

Prior to the ZIKV outbreaks in Pacific Islands and more recently in the Americas, ZIKV infections in Asia and Africa were not associated with neurologic disease, perhaps secondary to the lack of accurate laboratory diagnostics for ZIKV infection and the presence of confounding infections and environmental insults that could have limited identification of ZIKV as a cause of perinatal neurological disease. However, the possibility that the genetic evolution of ZIKV as it spread from Asia into the Pacific Islands and subsequently into the Americas resulted in a more neurovirulent strain of virus has been suggested in two recent reports ^74,75. In the study of Yuan, et.al., the authors described the acquisition of a single posttranslational modification on the envelope protein of the ZIKV circulating in Brazil that was associated with enhanced neurovirulence as compared to ancestor Zika strains from Cambodia, the origin of ZIKV circulating in South America ⁷⁴. In addition, this finding provides an explanation for the extended spectrum of disease in the Asian lineage of ZIKV as compared to the African lineage ⁷⁴. Similarly, the second report defined the loss of a glycosylation site on the ZIKV envelope protein that resulted in increased ZIKV neuroinvasiveness ⁷⁵. Lastly, it is important to note that although investigators have correctly focused on the capacity of this virus to infect and damage cells of the CNS, ZIKV can replicate in a number of cell types in-vivo and in-vitro and, more importantly, the analysis of tissues from fatal cases of ZIKV infections and non-human primate models of ZIKV infections have demonstrated ZIKV RNA in a variety of organs including the kidneys, liver, and spleen ^58,76. In addition, in-vitro ZIKV has been shown to replicate in a variety of primary human cells including dermal fibroblasts, cytotrophoblasts, decidual cells, Hofbauer cells, and Sertoli cells from the testis ^77–80. Thus, even though ZIKV is frequently referred to as a neurotropic virus in the literature, the virus exhibits very broad tropism for cells of many lineages in-vivo and in-vitro and its promiscuous tropism must be considered in the interpretation of findings from both human tissue and animal models of ZIKV disease.

Laboratory Diagnosis of ZIKV Infection

Although ZIKV can be isolated from infected individuals and propagated in tissue culture, current diagnostic approaches rely on detection of host antibody responses to ZIKV and detection of ZIKV RNA primarily by nucleic acid amplification methodologies. Development of more definitive diagnostics for identification of ZIKV infected individuals is a recognized priority but several hurdles remain in the serological diagnosis of acute or past ZIKV infection. Similarly, even though assays employing nucleic acid amplification have extraordinary sensitivity, the duration of ZIKV RNA in accessible fluids such as blood and urine in ZIKV infected individuals has not been well defined except in small series of selected patients such as those with demonstrable symptoms of acute ZIKV infection. The diagnosis of congenital ZIKV infection potentially could rely on serological testing if for example virus shedding in urine and/or blood is limited to only a subpopulation of infants infected in-utero such as those with the most severe symptomatology. As an example, one group of investigators reported that detection of IgM anti-ZIKV antibodies in newborn infants was strongly correlated with congenital ZIKV infection and the detection of IgM anti-ZIKV antibodies in the CSF could be considered as evidence of CNS involvement ⁸¹. Thus, many aspects of the virology of this maternal/infant infection remain to be defined before a diagnostic algorithm can be developed that will provide a definitive laboratory diagnosis of maternal or congenital ZIKV infection in the newborn infant.

Serological assays of both acute and past ZIKV infection have been plagued by the cross-reactivity between ZIKV and other flaviviruses that are often circulating simultaneously, including Dengue virus, West Nile virus and in some areas, Yellow Fever virus. Genetic analysis of multiple isolates of these viruses have revealed greater than 50% identity at the amino acid level between ZKIV and Dengue virus, West Nile virus and somewhat lower between ZIKV and Yellow Fever virus ⁸². Thus, there is considerable cross-reactivity of antibodies reactive with these closely related viruses ^9,83,84. Importantly, this cross reactivity extends to two of the most immunogenic ZIKV proteins, the envelope protein, E, and the non-structural protein, NS1. Antibody cross reactivity between Dengue virus and ZIKV has been well described, including studies of the E protein that have clearly defined the structural basis for cross reactivity between anti-ZIKV and anti-Dengue virus antibodies ^84–86. The serological diagnosis of ZIKV in areas with circulating Dengue virus has been and remains problematic ^9,83,87–89. As a result, testing for acute ZIKV infection has relied on both the presence of antigen binding IgG (or IgM) antibodies and functional assays which quantify the amount of both ZIKV and Dengue virus neutralizing antibodies in the same serum specimen. An increased titer of ZIKV neutralizing antibodies as compared to Dengue virus neutralizing antibodies in the same specimen has been used as evidence of a probable ZIKV infection. Variations on this approach have greatly simplified the assay formats, but these assay continue to be plagued by indeterminate results as well as potential false positive results secondary to the cross-reactivity present in an individual serum specimen ^90–93. Recently, investigators have described an assay based on competitive inhibition of binding of a ZIKV NS1 specific monoclonal antibody that appears to have increased sensitivity and specificity for the serological detection of a ZIKV specific response, except in the first 10 days following ZIKV infection ⁹⁴.

The gold standard for the diagnosis of acute ZIKV infection is the detection of ZIKV RNA in a relevant clinical specimen, usually by nucleic acid amplifications technologies. ZIKV nucleic acids have been amplified from whole blood, urine, saliva, and serum from acutely infected patients ^18,95–97. It has been reported that urine provides a more reliable specimen for the detection of ZIKV RNA both during and following acute infection ^96,98,99. Other sources of ZIKV RNA include saliva from infected patients that in one study was shown to more reliably detect infection than blood ¹⁸. Conversely, other studies have suggested that whole blood could be the optimal patient specimen for detection of ZIKV in that ZIKV RNA was detected in whole blood for up to two months after infection ¹⁰⁰.

A wide array of conditions and primer/probe sets for PCR based assays have been published with a similarly wide variance in their specificity and level of sensitivity. Some of the more widely employed assays have been shown to lack sensitivity, thus creating a window for detection of viremia and/or viruria in patients acutely infected with ZIKA. In contrast, other assays that employ an RNA amplification step prior the conversion of the RNA to a DNA template for PCR have been shown to have markedly improved sensitivity, albeit with the added cost of instrumentation and sample processing ¹⁰¹. Several critical questions remain surrounding the identification of the most sensitive source of ZIKV RNA during different times of infection and the most consistent and specific conditions for amplification of ZIKV RNA. Thus, the optimal diagnostic approach will remain a work in progress until the virology of ZIKV is defined in sufficiently large populations with varied phenotypic expression of infection and presumably, similar variations in the magnitude and duration of virus replication/shedding.

Finally, numerous reports have described the use of ZIKV reactive monoclonal antibodies (mAbs) or polyvalent antisera for immunohistochemical or fluorescent antibody detection of ZIKV antigens in tissue specimens. Similar problems surround results from these approaches, particularly the non-specific reactivity of the some of the more widely available ZIKV reactive antibodies. In contrast, analysis of tissue specimens using in-situ nucleic acid hybridization for detection of ZIKV nucleic acids appears to provide specificity and sensitivity but with increased technical demands ^58,59,76.

Treatment and Prophylaxis: Vaccine Development

Development of antiviral drugs for treatment and/or prophylaxis of ZIKV infection has been accelerated because of established drug development programs for treatment of Dengue virus infections ^102,103. Two approaches have been commonly employed. The first is screening of small molecules and nucleoside analogs in assays targeting functional enzymatic activities of non-structural proteins such as NS3 and NS5 ^102–106. The second approaches involves testing licensed drugs, particularly those licensed for treatment of hepatitis C virus, for activity against ZIKV. This latter approach described as repurposing existing drugs offers an efficient and cost effective strategy for the discovery of effective treatments ^51,107. An example of this approach is the demonstration of the activity of sofosbuvir, a licensed direct acting Hepatitis C antiviral drug, against ZIKV both in-vitro and in-vivo in a murine model of ZIKV infection ^108,109. Similarly, the malaria drug, chloroquine, has been shown to block ZIKV infection both in-vitro and in-vivo, presumably through its inhibition of processes in endocytosis ^110–112. In some cases, repurposed drugs have entered testing in clinical trials for treatment of Dengue virus infections ^113,114.

Even though there is considerable interest in the development of drugs for treatment of flavivirus infections, establishing efficacy in a clinical trial is confounded by identifying quantifiable endpoints secondary to characteristics of flavivirus infections, including the lack of defined relationships between clinical outcomes and readily measurable clinical parameters such as fever and laboratory parameters such as viremia. Recent studies in rhesus macaques provided insight into issues that surround the design of an informative clinical trial for testing antivirals for treatment of flavivirus, specifically ZIKV infections ¹¹⁵. In this experimental model, ZIKV was shown to have a 4 hour eclipse period in-vivo (time to production of infectious virus after initial infection) and after about 5 hrs, a single infected cell produced approximately 3.5 logs of infectious virus ¹¹⁵. Modeling the efficacy of an antiviral drug that inhibits virus replication demonstrated that treatment at the time of infection would clear viremia by 3 days after infection whereas beginning treatment at 2 days post infection provided no benefit in the changing the virus load in infected animals ¹¹⁵. These data demonstrate a major hurdle that must be overcome if treatment is initiated at the time of symptoms of a flavivirus infection. Finally, the major public health threat of ZIKV infection is the impact of this infection on the outcome of pregnancy, thus treatment will involve administration of an antiviral compound to pregnant women. Treatment of pregnant women with antiviral agents, particularly nucleoside analogs, remains problematic and safety profiles of candidate antiviral drugs must exceed drugs that are licensed for use in non-pregnant hosts.

In contrast to limitations associated with treatment of pregnant women with antiviral drugs, the use of biologics, specifically anti-ZIKV antibodies, offers an approach to limit the magnitude and duration of viremia and likely modify ZIKV infection of specific tissues. Several potent ZIKV neutralizing mAbs have been developed and in some cases, tested in experimental models of ZIKV infection ^116–119. These antibodies could be used as prophylaxis in pregnant women during ZIKV outbreaks and potentially as treatment in women with acute ZIKV infection. Importantly, these mAbs have provided evidence that prophylactic vaccines can likely be developed and have also identified characteristics of vaccine induced antibody responses that could be expected to provide protection from ZIKV infection ¹²⁰.

A number of different candidate prophylactic vaccines have been described and some have been scheduled for phase I trials in human volunteers ¹²¹. Vaccines have been developed utilizing multiple approaches, including (i) inactivated whole ZIKV, (ii) virus like particles derived from expression of subgenomic fragments of ZIKV, (iii) recombinant VSV expressing ZIKV envelope protein, (iv) recombinant adenovirus expressing ZIKV envelop protein, (v) mRNA based vaccines, and (vi) DNA vaccines ^122–128. Robust ZIKV neutralizing antibody responses were induced by each of these candidates suggesting that the development of a prophylactic ZIKV vaccine will be forthcoming.

A continuing controversy in the pathogenesis of flavivirus infection, specifically DENV, is the importance of antibody dependent enhancement of infection (ADE), a phenomenon first reported in-vitro and subsequently shown to contribute to severe DENV infections in individuals infected with DENV a second time. Although the discussion of this critically important topic in is beyond the scope of this chapter, its potential importance to ZIKV vaccine development cannot overstated as a contentious debate continues between investigators with interests in DENV vaccine development and deployment. Evidence that ADE during DENV infection leads to more severe diseased has been a paradigm in this area of research and recently studies in a large pediatric cohort in Nicaragua provided evidence that a quantitative relationship between protective and non-protective levels of DENV binding antibodies likely will explain the impact of ADE on DENV pathogenesis ¹²⁹. Studies in experimental models have demonstrated that ADE could be demonstrated following ZIKV infection in NHP previously infected with DENV, but that preexisting DENV seroimmunity shortened the duration of ZIKV viremia ¹³⁰. Studies in murine models of ZIKV infection have demonstrated that preexisting DENV binding antibodies can lead to in-vitro evidence of ADE and, as described in one report resulted in more significant disease in ZIKV infected mice ^131,132. Antibodies produced following a ZIKV infection can also result in ADE leading to more severe DENV infection in NHP with increased levels of DENV viremia, neutropenia, and expression of pro-inflammatory response, characteristics of disease following ADE during DENV infection ¹³³. As noted above, this aspect of the pathogenesis of flavivirus infections requires careful definition prior to widespread deployment of ZIKV vaccines into populations in areas of circulating DENV, However, findings reported by Katzelnick, et.al. combined with recently reported findings that non-neutralizing but DENV binding antibodies can activate Fcy Ilia receptors has suggested a mechanism of disease associated with ADE and thus, could provide a measureable surrogate of ADE in vaccine induced responses that could be identified prior to clinical testing of candidate vaccines ^129,134.

Waning of ZIKV Outbreak in the Americas and Projections for the Contribution of ZIKV to Adverse Outcomes of Pregnancy.

Following the large number of cases of ZIKV infections in South and Central America in 2015 and early 2016, several models suggested spread of this epidemic into the Caribbean and states in the US bordering the Gulf of Mexico. Although these models correctly predicted the spread of ZIKV to these areas, the magnitude of the outbreaks were smaller compared to the outbreak reported in Brazil and the incidence of adverse outcomes of pregnancy were correspondingly less than observed in Brazil. Furthermore, the incidence of ZIKV virus infection has dropped dramatically in South and Central America, perhaps secondary to development of ZIKV specific immunity in a large percentage of the population secondary to the high attack rate in the initial outbreak ^135,136. The decrease in new cases of ZIKV infections has resulted in decreasing enrollments for ongoing natural history studies and could potentially limit field testing of candidate ZIKV vaccines. However, several models have suggested that ZIKV will become established in these areas as an endemic infection and thus, continue to contribute to adverse outcomes of pregnancy ^137–140. These projections strongly argue for continued surveillance of ZIKV as a cause of adverse outcomes in pregnancy and to commit the resources required to more fully define the natural history of this viral infection during pregnancy.