Small-Animal Models of Zika Virus

Abstract

Zika virus (ZIKV) infection can result in serious consequences, including severe congenital manifestations, persistent infection in the testes, and neurologic sequelae. After a pandemic emergence, the virus has spread to much of North and South America and has been introduced to many countries outside of ZIKV-endemic areas as infected travelers return to their home countries. Rodent models have been important in gaining a better understanding of the wide range of disease etiologies associated with ZIKV infection and for the initial phase of developing countermeasures to prevent or treat viral infections. We discuss herein the advantages and disadvantages of small-animal models that have been developed to replicate various aspects of disease associated with ZIKV infection.

ACUTE VIRAL DISEASE

There are many advantages to modeling viral disease in mice and other small-animal species. Various advantages and disadvantages of rodent species that have been used to replicate disease manifestations of Zika virus (ZIKV) disease are included in Table 1. This article will summarize the major findings of work performed in various small-animal models to characterize disease outcomes after challenge with ZIKV.

Table 1.

Comparison of Advantages and Disadvantages of Small-Animal Species Used to Model Zika Virus Infection and Disease

Variable	WT Mice	Ab-Induced Immune Deficient Mice	IFN Pathway KO Mice	STAT2-/- Hamsters	Guinea Pigs
Advantages	Readily available, inexpensive, normal immunity, well characterized, diverse immune profiles	Readily available, induced immune deficiency as needed, high virus replication	Age-dependent disease, diverse disease manifestations, lethality, high virus replication	Intermediate sensitivity, persistent infection of testes, lower mortality rate	Readily available, similar placentation to human, naturally susceptible to infection, normal immunity
Disadvantages	Naturally resilient, low virus replication, requires high virus inoculum, difficult infection routes	Ab expense/availability, reversion to WT and clearance of virus, little overt disease	Limited availability, may require in-house colony, susceptible to other pathogens, abnormally severe disease	Lack of reagents, limited availability, underlying polyomavirus, variable disease manifestation	Lack of reagents, large size, not well characterized, expensive to house, somewhat difficult to work with

Abbreviations: Ab, antibody; IFN, interferon; KO, knockout; WT, wild type.

Initial studies investigating ZIKV shortly after its discovery in the mid 1900s included challenge of various mouse strains, as well as a human challenge model [1–4]. These early studies demonstrated that mice were somewhat refractory to ZIKV infection [5], so work during the 20th century was limited to around 40 published reports. Emergence of ZIKV in the early 21st century resulted in an increased effort to learn more about this virus, with around 2800 papers being published (as of August 2017). One of the main goals was to develop small-animal models for use in delineating the infection cycle, identifying consequences of virus infection, and discovering antiviral countermeasures. A brief summary of the work to develop mouse models of acute ZIKV infection are included in Table 2.

Table 2.

Mouse Models of Acute Zika Virus (ZIKV) Infection and Disease

Mouse Strain(s)	Age, wk	ZIKV Strain, Dose, Route	Pathological Finding		Reference
			Neurological	Other
Porton	0, 4	MR 766, titration, intracerebral	Not reported	LD50 of 107.2 pfu in 1–2-d-old mice and 106.4 pfu in adult mice	[5]
IFNAR-/-or AG129	5, 11	Cambodian, 105 pfu, intraperitoneal	Brain virus 3, tremors 6 d after virus injection	Viremia detected 2 d after virus injection, virus in spleen and testis	[8]
AG129	3–4 or 8	H/PF/2013, 100–5 pfu, subcutaneous, 105.3 pfu, intraperitoneal	Increased virus level in brain, neurodegeneration	Viremia detected 2 d after virus injection, myofiber necrosis, inflammatory cell infiltration, nuclear rowing	[6]
AG129	8	MR 766, 101–5 pfu, intraperitoneal	Paralysis, conjunctivitis, neurodegeneration, viral antigen in brain and spinal cord, acute encephalitis	IFN-γ and IL-18 levels increased in serum, viral RNA detected in brain, spleen, liver, kidney	[9]
IFNAR-/-	5–6	MP1751, 106 pfu, subcutaneous	Perivascular cuffing, polymorphonuclear cells in gray/white matter near blood vessels, neurodegeneration in hippocampus	Viral RNA detected in brain and spleen 3 d after virus injection, apoptosis in spleen, hematopoiesis in liver	[84]
AG129	8–10	P 6–740, 103.0 pfu, subcutaneous	Paralysis, conjunctivitis, hyperexcitability, seizure, tremor, ZIKV Ag in neurons and astrocytes, encephalitis, and myelitis	Viral RNA level peaked 5 d after virus injection; virus level in testis peaked 7 d after virus injection, including in Leydig cells; viral RNA shedding in urine	[7]
IFNAR-/- or Irf3-/-Irf5-/-Ifr7-/-	5–6	H/PF/2013 or MR766, 102 ffu, subcutaneous (footpad)	ZIKV RNA in brain and spinal cord, higher paralysis incidence after intravenous delivery	Viral RNA detected 2–6 d after virus injection, 106 ffu virus in testes of IFNAR-/- mice	[12]
C57BL/6 + Ifnar Ab	5	Dakar, 106.4 pfu, subcutaneous or intraperitoneal	Paralysis 8 d after virus injection after intraperitoneal challenge, viral RNA by in situ hybridization, neurodegeneration in cerebellum and hippocampus	Viral RNA detected in brain, kidney, skeletal muscle, and spleen 3 d after virus injection; increased glial fibrillary acidic protein level in cerebrum, brain titer 15 d after virus injection	[17]
C57BL/6, IFNAR-/-	0–1	PRVABC-59, 103.3–4.3 pfu, subcutaneous	Wide array of neurologic symptoms, virus in neurons	Increased inflammatory gene expression, T-cell– driven immunity, viral load in brain and spleen	[85]
STAT2-/-, IFNAR-/-	5–6	MR-766, 103 pfu, subcutaneous	Viral RNA in brain, hind limb paralysis, elevated cytokine levels in brain	Viral RNA detected in various tissues, upregulated IFN expression	[13]
BALB/c + immunosuppression with dexamethasone	6–8	PRVABC-59, 106.5 pfu, intraperitoneal	Neurons in hippocampus stain positive for ZIKV NS1	Orchitis	[18]
Swiss	0	SPH 2015, dose not provided, intracerebral or subcutaneous	Neurological disease and death of neurons, gliosis	Atonic urinary bladder	[19]
STAT2-/-	5–8	PRVABC-59, 103.7 pfu, intradermal	Elevated ZIKV replication in spinal cord	Enhanced disease severity after convalescent sera treatment	[23]

Abbreviations: Ag, antigen; ffu, focus-forming units; IFN, interferon; IL-18, interleukin 18; LD₅₀, median lethal dose; NS1, nonstructural protein 1; pfu, plaque-forming units.

Experiments conducted during the last several years often used immune-compromised mouse strains, which were permissive for virus infection and displayed various neurologic signs of disease, including some that replicated severe disease manifestations in humans. The AG129 mouse strain, which lacks receptors for interferon α/β (IFN-α/β) and IFN-γ were susceptible to infection after inoculation with various strains of ZIKV injected by various routes [6–9]. Virus replicates in a wide range of tissues, and viral RNA can be detected in the serum during the course of infection, with timing depending on the strain and route of virus infection. Relevant disease signs include lacrimation and neurologic involvement. The severity of disease observed in this mouse strain underscores the importance of the IFN response in controlling ZIKV infection.

Other IFN-pathway-knockout mice also show varying degrees of disease severity after infection with ZIKV and provide a suitable model for virus replication in tissues. Infection of IFNAR^-/- mice, lacking IFN-α/β receptors, with ZIKV results in age-dependent severity, with mortality occurring in mice that are around 3 weeks of age or younger [8]. The presence of IFN-γ receptors provides intermediate protection in 5-week-old mice and complete survival in mice that are 11 weeks old. Another knockout strain, IFNGR^-/- mice, lack IFN-γ receptors and are susceptible to ZIKV infection. These mice were used to compare the pathogenicity of various ZIKV infectious clones [10]. Mosquito transmission of ZIKV to IFNAR^-/- mice has helped identify the vector competency for several virus strains [11]. STAT2^-/- knockout mice and mice lacking IFN response factors 3, 5, and 7 are also susceptible to intravenous infection with ZIKV [12, 13].

Histologic analysis of AG129 mice infected with ZIKV identified neurons throughout the central nervous system that were positive for ZIKV antigen, with outcomes including neurodegenerative multifocal neutrophilic encephalitis and myelitis [7]. The observation of ZIKV in motor neurons in the ventral horn of the spinal cord was similar to disease observed after infection with West Nile virus [14]. Astrocytes were also heavily infected in various regions of the brain and spinal cord after ZIKV infection, which was similar to pathology observed after challenge of mice with Venezuelan equine encephalitis virus, an alphavirus [15]. Interestingly, infected AG129 mice displayed rear limb myofiber degeneration and necrosis with inflammatory cell infiltration in the absence of hind-limb paralysis, suggesting that this virus infects muscle cells [6]. Viral antigen positivity, however, was not directly observed in the muscle tissue. Direct intraocular inoculation with ZIKV results in infection of the cells lining the blood-retinal barrier and causes chorioretinal atrophy [16].

Since wild-type mice generally display only a very transient viral infection after ZIKV challenge, various methods are used to increase the susceptibility of wild-type, immunocompetent strains. Mice can be treated with function-blocking antibodies (MAR1-5A3) targeting the IFN-α/β receptors to increase the susceptibility of the mice to infection following challenge [17]. Viral RNA was detected in the cerebrum and hippocampus in mice treated with function-blocking antibody. Dexamethasone immunosuppression renders BALB/c mice susceptible to ZIKV infection, resulting in detectable virus titer in various tissues and lethality [18]. Infection of neonatal mice shortly after birth also results in morbidity and mortality, with more-severe disease and death observed in younger mouse pups [19]. Infected neonates developed neurologic complications, including tremors, seizure, hyperactivity, and limb collapse, with detectable virus in the brain as late as 15 days after infection.

Preexisting flavivirus antibodies are implicated in worsened disease during infection, a concept known as antibody enhancement of disease. This phenomenon is well-known for dengue virus (DENV), where antibodies to one strain bind at low levels to an incoming heterotypic strain, resulting in an increased uptake of virus by Fc receptor–bearing cells of the immune system. Infection in this manner effectively increases the host cells available for viral replication and thereby increases the antigen load, host response, and immunopathology of the infection [20]. This has been demonstrated in immunodeficient mouse models of DENV infection [21, 22]. Enhanced disease in ZIKV-infected mice after treatment with immune serum–containing antibodies to DENV or West Nile virus has been demonstrated [23]. Because of the overlap in sequence between DENV and ZIKV, it is possible that antibody enhancement of disease might also play a role in ZIKV pathogenesis, as has been demonstrated in rodent models. However, there is no clinical evidence to support this idea, and care should be taken not to overinterpret data from rodent models. Additionally, mounting data from epidemiological studies and studies in nonhuman primates suggest that antibodies to DENV play more of a protective role, rather than a role in disease enhancement [24, 25]. Previous exposure to DENV did not correlate with worsened outcome in ZIKV infection during pregnancy [26]. Further epidemiological data are needed to provide further insight into this subject.

Aside from mouse models, other small-animal species have been used to model ZIKV, including chickens and guinea pigs. Infection of chicken embryos with ZIKV results in virus infection of the developing nervous system, causing fetal demise at higher virus challenge doses and a microcephaly-like phenotype at lower doses, replicating some aspects of congenital infection [27]. Immunocompetent guinea pigs infected with ZIKV developed a transient viremia, had detectable viral RNA in whole blood and serum, had an increase in cytokine and chemokine levels, developed infection of various tissues, and developed neutralizing antibodies [28]. These additional models of ZIKV infection and disease may provide useful systems for the evaluation of countermeasures or in disease characterization.

CONGENITAL INFECTION

Intrauterine exposure of a developing fetus to ZIKV infection can result in debilitating manifestations in the fetus, the most severe and obvious of which is microcephaly. An increase in the incidence of microcephaly during the recent ZIKV outbreak in Brazil eventually led to the discovery of the causative role of the virus in developmental abnormalities of fetuses after exposure in utero. Aside from microcephaly, other disease manifestations have been reported, including smaller birth weight, brain abnormalities despite normal head size, hearing loss, optic nerve hypoplasia, joint and bone deformities (arthrogryposis), and many other less apparent effects of ZIKV congenital exposure [29]. A significantly higher fetal mortality rate was associated with ZIKV infection among pregnant women in Brazil, compared with that due to other etiologies (approximately 5.1% vs 1.4%; P < .0001) [30]. This section will focus on various consequences of intrauterine exposure to ZIKV that have been replicated in small-animal models, as well as those aspects of congenital infection that are dissimilar between rodents and humans.

In attempting to recapitulate disease associated with intrauterine infection in small-animal models, various laboratories have independently developed mouse models of congenital infection (Table 3). Despite the use of a diverse range of ZIKV strains, as well as differences in the dose, route, and timing of infection during gestation, some consistent consequences of infection have been delineated in rodents. Virus was generally detected in the placenta [31–34], while fetal infection was dependent on the timing of maternal challenge. Specifically, lower fetal infection rates were generally observed when the female was challenged at gestational stages after embryonic day 12 (as discussed below). Another commonly observed consequence was intrauterine growth restriction and spontaneous abortion of developing fetuses [31, 35, 36]. These findings are similar to outcomes in natural congenital infection [30] and can be used to further characterize congenital infection, as well as to identify countermeasures to inhibit or prevent fetal infection. The period between embryonic days 3 and 14, when pregnant mice have been challenged with ZIKV across several studies, corresponds to gestation days 4–48 during the first trimester in humans.

Table 3.

Rodent Models of Congenital Infection With Zika Virus (ZIKV)

Strain(s), Rodent	Origin(s), Virus Strain(s)	Infection Route(s), Dose(s)	Embryonic Day(s)	Major Findings	Reference
IFNAR1-/- × C57BL/6, mouse	French Polynesia, H/PF/2013	Subcutaneous (footpad), 103 ffu	6.5, 7.5	Fetal and placental infection, IUGR, fetal demise, placental and fetal brain apoptosis	[31]
C57BL/6, mouse	French Polynesia, H/PF/2013	(MAR1-5A3 Ab E5.5) subcutaneous (footpad), 103 ffu	6.5, 7.5	Placental and fetal infection, IUGR	[31]
SJL, mouse	Paraiba, Brazil, 2015	Intravenous/103, 1010.6, or 1012 pfu/mL	10–13	IUGR, upregulation of apoptosis genes, cortical malformations	[35]
ICR, mouse	SZ01	Lateral ventricle (injection of fetus), 105.8 pfu/mL (1 µL)	13.5	Brain replication in ventricular subventricular zones, cortical thinning, infection of neural progenitor or intermediate progenitor cells	[45]
C57BL/6, mouse	SZ01	Intrauterine (fetal brain injection), 105.5 pfu	13.5	ZIKV infection of placenta and fetal brain, reduction of cortex founder cells	[33]
C57BL/6, IRF3/7-/-, and IFNAR-/-, mouse	Cambodia FSS13025	Intravaginal, 104.4–105.7 pfu	4.5–8.5	IUGR, fetal demise, ZIKV infection of fetal brain (based on RNA detection and electron microscopy findings)	[36]
CD-1, mouse	2010 Cambodia, 2015 Brazil, 2015 Puerto Rico, 1968 Nigeria	Intrauterine, intraperitoneal /106 TCID50	10, 14	High aborted fetus rate (30%–45%), virus in fetus/ dam, and increased interferon expression 48 h after virus injection w/infection at embryonic d 10 (not embryonic d 14); neuroinflammation and cortical thinning in neonatal brain	[43]
STAT2-/-, hamster	Malaysia, P 6–740	Subcutaneous/102.7 CCID50	8.5	Virus in fetal brain, placental pathology, live births	[32]
FVB/NJ, C57BL/6, mouse	Bahia, Brazil, HS-2015-BA-01	Intravenous (jugular)/105 pfu	5.5, 7.5, 9.5	Infection of placenta/fetus, pathology of fetal brain, fetal demise, arthrogryposis	[34]

Abbreviations: CCID₅₀, 50% cell culture infectious dose; ffu, focus-forming units; IUGR, intrauterine growth restriction; pfu, plaque-forming units; TCID₅₀, 50% tissue culture infectious dose.

The interaction between virus in maternal blood and fetal tissues occurs in the placenta. Although the placental structure of mice is quite distinct from that of humans, there are various similarities that are found between the two that make the mouse a reasonable model for various aspects of human placentation, including notably the lining of fetal structures with syncytiotrophoblast cells that contact maternal blood [37]. Trophoblast cells of the mouse conceptus will differentiate, forming a branched villous structure, which invades the placental wall decidua beginning around embryonic day 8.5. The placenta of mice is considered functional at approximately 10 days after coitus [38–40].

The timing of virus challenge during pregnancy is important in the context of congenital infection of the fetus in rodent models. Congenital infection of mice with other flaviviruses was demonstrated when virus challenge occurred between embryonic days 5 and 12, but detection of virus or viral antigen in the fetus was less likely as development proceeded past embryonic day 12 [41, 42]. Similar observations have been reported with ZIKV infection in mouse models [34, 43]. Although peripheral challenge of pregnant dams with ZIKV after embryonic day 12 results in virus infection of the placenta, infection at this time does not typically result in fetal infection in rodents. The timing of reduced congenital virus infection in rodents appears to be associated with placental development. This is somewhat dissimilar to congenital infection in humans, where ZIKV infection during the second and third trimesters can result in transmission to the fetus [44]. This observation could indicate partial control of the dissemination of virus to the fetus at the placental barrier, or it could suggest other mechanisms involved with transplacental infection.

A functional placenta is present in mice around embryonic day 10. Challenge of mice between embryonic days 4 and 14 results in productive placental infection in mice. The detection of virus in the placenta after challenge from embryonic days 4 to 14 suggests a high degree of susceptibility of this organ [43]. After the development of a functional placenta in mice, transplacental transmission of ZIKV and other flaviviruses to the fetus is reduced unless the fetuses are challenged directly [33, 43, 45]. The lack of infection at later times in mice may suggest the placental barrier is preventing transplacental movement of ZIKV in the mouse. However, this is dissimilar to congenital infection of human fetuses, where brain lesions have been observed in newborn babies after ZIKV infection during the second and third trimesters of pregnancy [46]. However, these second and third trimester infections of human placentas appear to be limited primarily to Hofbauer cells, and the inflammatory villitis that is seen during first trimester placental infections is reportedly absent at later infection times [47]. Indeed, Hofbauer cells isolated from human placentas have been the focus of several studies, which demonstrate susceptibility of these cells to ZIKV [47–49]. In contrast, infection of the placenta during the first trimester of human pregnancy results in infection in a wider array of placental cells. Mesenchymal cells and cytotrophoblasts support ZIKV replication, along with Hofbauer cells. This gestationally dependent infection of placental cells was further supported by a study that demonstrated expression of genes for various entry factors used by ZIKV in early stage–type trophoblast cells [50]. These cells were highly susceptible to infection in cell culture, while cells isolated from term placentas did not express these genes and were relatively resilient to ZIKV infection. While mice and humans may differ in some aspects of congenital infection, challenge of pregnant mice during early gestation (ie, before embryonic day 10) models many aspects of natural infection.

Most of the studies have terminated at some point during gestation, and few live births have been recorded. The consequences of intrauterine exposure to ZIKV on the development of neonates will be an important step in future research to further identify consequences of congenital infection. Depending on the strain of virus used, the immune state of the dam, and the timing of infection, the aborted fetus rate can be relatively high [43], which would result in fewer births. Disease manifestations in females just after birth may cause the dams to neglect or cannibalize pups [32], further complicating studies designed to characterize the effect of intrauterine exposure to ZIKV on the development of neonates. Future studies should include live births to determine the suitability of mice to model the effects of intrauterine infection on developing neonates.

SEXUAL TRANSMISSION

A somewhat unanticipated consequence of ZIKV infection is sexual transmission. Evidence of non–mosquito-borne transmission was available as early as 2008, when a returning scientist that was infected with ZIKV in Senegal transmitted the virus to his wife, which was suspected to be through sexual contact [51]. Closely related flaviviruses, such as DENV, are not known to be transmitted through sexual contact. It has been well documented that men who were exposed to ZIKV in areas of endemicity could further spread the virus to their sex partners, which may or may not include the manifestation of disease. Various clinical studies have provided evidence of sexual transmission of ZIKV from male to female partners, from female to male partners, and from male to male partners [52–56], although the great majority of transmission cases are from male to female partners. Transmission has been reported to occur several weeks after returning from areas of endemicity, with the majority of cases occurring within 3 weeks of return. Viral RNA was detected in the semen of infected males up to 6 months after disease onset [57], although very few studies have evaluated the duration of persistence of ZIKV in semen. While little is known about the localization of virus in the reproductive tract of infected men, mouse models have provided insight into which cell types are infected and potential mechanisms of sexual transmission (Table 4).

Table 4.

Rodent Models of Zika Virus (ZIKV) Sexual Transmission and Infection of the Male Reproductive Tract

Model	Virus Origin, Strain	Infection Route(s)/Dose(s)	Major Findings	Reference
C57BL/6, (MAR1-5A3), Rag1-/-, and Axl-/- mice	Dakar 41519 (ma) H/PF/2013	Subcutaneous (footpad)/106.0 or 103.0 pfu	Persistence of ZIKV in testicles/epididymides (21 d after virus injection), reduced testis size, low testosterone level, infected spermatogonia, seminiferous tubule degradation	[59]
IFNAR-/- mice	MEX2-81	Subcutaneous (footpad)/105 pfu	Presence of viral RNA in testicles/epididymides 21 d after virus injection, testicular atrophy, reduced testosterone level	[60]
IFNAR-/- mice	ZIKA-SMGC-1	Intraperitoneal/103.0 pfu, intratesticular/102.6 pfu	Atrophy of the reproductive tract, virus in testicles/epididymis but not seminal vesicle or prostate	[61]
AG129 mice	Puerto Rico, PRVABC59	Intraperitoneal/103 pfu	ZIKV RNA/virus observed in semen, demonstration of sexual transmission of virus and fetal infection, virus/inflammation in testicles/ epididymides	[58]
STAT2-/- hamsters	P 6–740	Subcutaneous/102.7 CCID50	Infection of Sertoli cells and spermatogonia	[32]
BALB/c mice (dexamethasone)	Puerto Rico, PRVABC59	Intraperitoneal/106.5 pfu	Pathology of seminiferous tubules, lymphocytic infiltration (12–14 d after virus injection), reduced viral load in testicles/epididymides and prostate after interferon treatment	[86]
AG129 and LysMCre+ IFNAR mice	Cambodia, FSS13025	Intravaginal/105 or 106 ffu	Estrus cycle–dependent susceptibility of females	[62]
IFNAR-/- mice	Puerto Rico, PRVABC59	Rectal/106.5 pfu	Nonlethal, virus in rectum, testis, brain, and spleen on d 21; inflammation; and splenomegaly	Martinez et al, unpublished data

Abbreviations: CCID₅₀, 50% cell culture infectious dose; ffu, focus-forming units; pfu, plaque-forming units.

Many published reports have demonstrated ZIKV infection of the male reproductive tract of rodents, primarily indicating detection of infectious virus or viral RNA in the testes of infected males [7–9, 12]. The testes support very high levels of virus, particularly in immunocompromised mice. Virus has been localized in various cells of the testis, including Leydig cells and spermatogenic precursors, and in epithelial cells of the epididymis [32, 58]. Interstitial inflammation and inflammatory cell infiltrates were observed. Necrotic cells that stained positively for viral antigen were observed in the lumen of the epididymides [59], representing a potential source of virus for transmission. Other severe disease manifestions, such as orchitis and testicular atrophy, have been observed in mouse models [60].

The implications of testis infection observed in mice may not be directly translatable to human infection. For example, orchitis, or inflammation of the testicle, is a very painful condition in people and is generally the consequence of bacterial infection, including sexually transmitted infections. This is commonly observed in immune-suppressed mice infected with ZIKV [18, 61]. If a common consequence of ZIKV in infected men was orchitis, the pain would be intense, and they would surely seek medical attention. Orchitis is likely not a commonly occurring consequence of ZIKV infection of the male reproductive tract.

Intravaginal infection of female mice with ZIKV can cause a systemic infection and may also be transmitted to offspring [58, 62]. The estrous cycle may influence the susceptibility of females, as AG129 mice that were inoculated intravaginally with ZIKV during a hormonally induced estrus-like phase did not die of viral infection and had a relative lack of virus replication in various tissues, while those that were in an induced diestrus-like phase displayed virus replication and mortality [62]. Transmission from infected males to naive females has been demonstrated in AG129 mice, including transmission from vasectomized mice, despite a lower virus load as compared to intact males [58]. Males also had detectable virus in the semen several weeks after virus challenge, modeling persistent virus present in men infected naturally with ZIKV. Rectal inoculation of male mice results in systemic spread of virus to various tissues, including the testes, and supports observations of male-to-male transmission (Martinez et al, unpublished data).

USE OF MODELS IN THE DEVELOPMENT OF COUNTERMEASURES

As there are currently no Food and Drug Administration (FDA)–approved drugs to treat acute flaviviral diseases, it is unknown how direct-acting antiviral compounds would affect infection and disease outcomes in people infected with ZIKV. There is potential for antiviral treatment to reduce disease burden and further spread of the virus if therapy is initiated soon after the onset of clinical disease or if prophylactic treatment, in the context of a widespread viral outbreak, is used. In regard to the treatment of Zika, antiviral agents could have use in preventing fetal transmission, clearing sequelae from the testes, or reducing viral load during acute infection to reduce or inhibit further transmission.

Various nucleoside analogs, including 7-deaza-2ʹ-C-methyladenosine, BCX4430, sofosbuvir, and NITD008, have been shown to be active against ZIKV in cell culture and in mouse models [7, 9, 63, 64]. These viral RNA-dependent RNA polymerase inhibitors delay or prevent mortality, reduce virus titer in relevant tissues, and improve disease severity in infected mice. Although these broad-spectrum antivirals did not completely eliminate disease, efficacy in knockout mouse models was nevertheless impressive because of the acute susceptibility of these mice to ZIKV infection and would likely fare better against natural infection in an immunocompetent host. Sofosbuvir is an FDA-approved drug used to treat chronic hepatitis C virus infection [65]. If this compound has clinical efficacy against ZIKV, the approval process would be truncated because the compound is well characterized for use in humans. BCX4430 has been shown to be effective against a wide range of viruses of human concern [66–68], and clinical trials have been initiated. NITD008 was initially identified as a potential antiviral to treat DENV but had toxicity after long-term treatment [69]. Short-term treatment, applicable to treatment of acute arboviral diseases, did not result in appreciable toxicity, and further clinical investigation may be warranted. Broad-spectrum activity of NITD008 has also been observed in various animal models [69, 70].

Indirect-acting antiviral agents have also been identified to have activity against ZIKV in mouse models. The compound 25-hydroxycholesterol is an enzymatic product of cholesterol-25-hydroxylase. Treatment with 25-hydroxycholesterol reduced the ZIKV RNA level in the serum by blocking viral entry in mice and macaques, including reduction of the incidence of microcephaly in a congenital model [71]. Azithromycin, a macrolide antibiotic that is FDA approved for use, including during pregnancy, prevented ZIKV production and viral-mediated cell death in primary human brain tissue by reducing viral proliferation and cytopathic effects in glial cells and astrocytes [72]. However, activity has not yet been demonstrated in a small-animal model.

Antibody therapy has been used to control virus infection and disease during outbreaks [73]. Antibodies targeting the envelope protein and, in particular, domain III have been shown to include potent neutralizers that would be suitable for use in therapy [74, 75]. Efficient neutralizing antibodies have been isolated from the serum of people who have been infected with ZIKV, revealing the importance of antibodies targeting the envelope protein in the context of clearance of ZIKV infection [76]. Some of these antibodies have shown activity in mouse models and are effective in preventing or reducing disease severity after ZIKV infection, including prevention of congenital infection after treatment of pregnant dams [76, 77]. This is consistent with previous studies with the related West Nile virus, where prevention of congenital infection was observed in rodent models treated with pooled human immune serum [78].

Care must be taken in regard to use of antibody therapy in the context of ZIKV infection, as treatment with convalescent plasma from donors possessing antibodies specific for flaviviruses, including DENV and West Nile virus, may enhance disease severity [23]. Antibodies containing mutations in the Fc receptor region maintain the specificity and neutralizing capabilities of the antibodies but remove the possibility for negative enhancement interactions of antibodies with immune cells. Antibodies with an inactivated Fc receptor have been shown to be effective in small-animal models [76].

Vaccines are very important in controlling acute arboviral diseases and have demonstrated efficacy in substantially reducing the disease burden of yellow fever virus, the archetypical flavivirus [79]. Various types of vaccines have been developed to protect against ZIKV (Table 5), and additional studies are being published at a rapid pace. Inactivated or attenuated viruses may also elicit long-term immunity, as is the case with the attenuated 17D yellow fever virus vaccine. Inactivated and modified live-attenuated viruses (which have a 10-nucleotide deletion in the 3ʹ untranslated region) have shown promise in preventing ZIKV infection in mouse models [80, 81]. Development of vaccines that target the envelope protein have been prominent and show promise in various animal models [82, 83]. As with antibody therapy, an important question in regard to vaccine development is whether the antibodies elicited by vaccination might serve to enhance infection with ZIKV. Some reports of enhancement in small-animal models have been reported [23].

Table 5.

Various Anti–Zika Virus (ZIKV) Countermeasures Tested in Small-Animal Models

Countermeasure, Mouse Strain (Age)	Virus Strain	Treatment Protocol(s)	Outcome(s)	Reference
Antiviral therapy
AG129 (8–14 wk)	MR 766	7 DMA, oral, 50 mg/kg/d, once daily × 10 beginning -1 h	Reduced viral RNA level by 0.5–1.3 log10, delayed mean time to death to 15–23 d after virus injection	[9]
AG129 (8–10 wk)	P 6–740	BCX4430, 300 mg/kg/d, intramuscular, twice daily × 7 d beginning 0–7 d after virus injection	Reduced viral RNA level by approximately 2 log10 5 d after virus injection, delayed or prevented death, reduced disease severity, survival after rechallenge	[7]
IFNAR-/- (4 wk)	GZ01/2016	NITD008, 50 mg/kg, oral, 4, 24, 48, 72, and 96 h after virus injection	Reduced viral RNA level 2.6-fold 2 d after virus injection, 50% survival, reduced disease signs	[63]
C57BL/6 + Ifnar1 mAb	DAKAR 41519	Sofosbuvir, 33 mg/kg/d, oral × 7 d, beginning 24 h after virus injection	Improved survival, extended mean time to death	[64]
BALB/c, IFNAR-/- (3–4 wk)	GZ01/2016	25-Hydroxycholesterol, 50 mg/kg, intraperitoneal × 7 d, beginning -12 h	Reduced viral RNA level in serum (BALB/c), improved survival and reduced viral RNA level in various tissues (IFNAR-/-)	[71]
BALB/c (6–8 wk) + DEX	PRVABC-59	Pegylated interferon alfa 2b, 103.3 IU, subcutaneous, every 96 h 1–9 d after virus injection; interferon beta 1b, 105.2 IU/dose, intraperitoneal, every 48 h 1–9 d after virus injection	Reduced viral loads in various tissues, 100% survival, no prominent inflammation in any of the tissues tested	[18]
Antibody therapy
Pregnant ICR mice E13.5	SZ01	Convalescent sera, 100 µL, intraperitoneal, 1 d (E14.5) and 2 d (E15.5) after virus injection)	nAb titer of 161, reduced caspase 3 level in cortex, rescued cortical plate thinning	[77]
IFNAR-/-	H/PF/2013	Human mAb from DENV-infected patients, 10 µg, intraperitoneal, -1 and 9 d after virus injection	mAb EDE1 C10 neutralized ZIKV, 100% survival	[87]
C57BL/6 + Ifnar1 mAb	ZIKV- Dakar	mAb ZIKV-117 from human antisera, intraperitoneal, 6.7 mg/kg 1 d after virus injection or 16.7 mg/kg 5 d after virus injection	100% survival (1 d after virus injection) or 70%–75% survival (5 d after virus injection); reduced transmission, pathology, and mortality	[76]
Ifnar1 -/- female × WT male	ZIKV- Dakar (E6.5)	mAb ZIKV-117, dams treated with 6.7 mg/kg 1 d after virus injection or 16.7 mg/kg 5 d after virus injection	4–5-log10 reduced virus level in fetal and maternal tissues, protection in pregnancy model due to neutralization	[76]
Pregnant C57BL/6 mice	ZIKV- Dakar (E5.5)	mAb ZIKV-117, dams treated with 16.7 mg/kg 1 d after virus injection	Reduced virus in dams, placenta, and fetus; improved fetal and placental disease severity; prevented vertical transmission	[76]
Vaccine
AG129 (4–6 wk)	FSS13025 or MR 766	FSS13025 was administered subcutaneously. Inactivated MR 766 virus, vaccine, 10 µg/dose, 0 and 21 d, intramuscular	100% survival in MR 766 and FSS13025 infection, absence of detectable viremia in serum up to 6 d after virus injection	[81]
BALB/c, C57BL/6 or SJL (4 wk)	ZIKV2015, PRVABC-59	DNA prM-Env or DNA Env vaccines, 50 µg, intramuscular, -28 d after virus injection	Reduced viral RNA level, correlation of efficacy with nAb, Ab transfer improves	[82]
IFNAR-/-	FSS13025	10-del ZIKV, route, -4 wk	High nAb titer, robust T-cell response, 100% survival, undetectable viremia	[80]
CD-1 (1 d)	FSS13025	10-del ZIKV, as above	100% survival
AG129 (6 wk)	P6-740	prM-Env mRNA, -6 wk	High nAb, 100% survival	[83]
C57BL/6 + Ifnar1 mAb (8 or 18 wk)	Dakar 41519	Modified prM-Env mRNA, -6 wk	High nAb, 100% survival
C57BL/6, BALB/c	MR-766	VSV-ZprME, VSV-ZENV, intravenous or intramuscular, -3 wk	Offspring born to vaccinated females were protected from challenge on postnatal d 7	[88]

Abbreviations: conv, convalescent serum; Env, envelope protein; ffu, focus-forming units; mAb, monoclonal antibody; mRNA, messenger RNA; nAb, neutralizing antibody; pfu, plaque-forming units; RDRP, RNA-dependent RNA polymerase; VSV, vesicular stomatitis virus.

CONCLUSIONS

Small-animal models are important in delineating the consequences of ZIKV. The use of various rodent species and strains, including genetic knockouts, has provided useful information to help us better understand disease as a result of infection with this virus. Small-animal models are useful for the evaluation of antiviral agents and vaccines during preclinical studies. This includes the usefulness of these models for evaluating countermeasures against congenital and sexual transmission, which is very difficult in larger-animal models. Rodent models appear to be more susceptible to enhancement by antibodies to other flaviviruses, compared with primates, and may not be the best model system for humans. Robust infection of rodents typically requires knock down of the IFN response to cause disease, which has required the use of immunodeficient rodent strains and includes more-severe disease pathology. Continued efforts will provide information to aid in the development of countermeasures to reduce the disease burden of this emerging virus, and it is anticipated that improved small animals will be developed that better replicate the features of natural ZIKV infection and disease.

Notes

Supplement sponsorship. This work is part of a supplement sponsored by the National Institute of Allergy and Infectious Diseases (NIAID), part of the National Institutes of Health (NIH).

Potential conflicts of interest. Both authors: No reported conflicts of interest. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.