Neutralizing and non-neutralizing antibodies: Zika Virus and antibody dependent enhancement of infection

Although Zika Virus (ZIKV) has been isolated in 1947, until recently ZIKV infection was only associated with relative mild clinical symptoms and sporadic outbreaks in Africa, Asia, and Oceania. Following the emergence of ZIKV in the Americas however, maternal ZIKV infections have been associated with congenital infections of the brain and the CNS as well as with intrauterine growth restriction (IUGR) of foetuses and possibly also with an increased risk of miscarriage. ZIKV is therefore unique among the human pathogenic flavivirus’ since neither Yellow Fever Virus (YFV), Dengue Virus (DENV), West Nile Virus (WNV) or Japanese Encephalitis Virus (JEV) are transmitted trans-placental, thus infecting the embryo or foetus in utero. As described before, the infection of both mice and non-human primates with various ZIKV strains including the original ZIKV MR766 strain as well as strains from Asia (ZIKV FSS13025), Oceania (ZIKV H/PF/2013) and the Americas (ZIKV Paraiba 2015) as well as the infection of human placenta explants (ZIKV MR766 and ZIKV Nica-1/-2 2016) suggest that ZIKV can cross the placenta probably by infecting maternal cytotrophoblast cells (CTB) and maternal decidual fibroblast cells combined with placental injury due to the release of inflammatory cytokines. In addition, it has been proposed that either maternal antibodies against ZIKV or the closely related DENV might promote the entry of ZIKV-IgG complexes into cells in a process known as “Antibody-dependent Enhancement (ADE)”. ADE has been implicated in the development of severe forms of DENV associated hemorrhagic fever (Dengue hemorrhagic fever, DHF, or Dengue Shock Syndrome, DSS) which is generally thought to be caused by cross-reactive but not cross protective antibodies in which the antibodies produced during the infection with one DENV serotype fail to neutralize viral particles of a different serotype but instead facilitate the entry of DENV into cells bearing the Fcγ receptor such as monocyte derived macrophages or monocyte derived dendritic cells. This increases viral replication and thus the viral load.

Both ZIKV PRVABC59 and ZIKV MR766 replicate in a wide variety of cell lines, including LNCaP cells (prostate cancer cell line), ARPE19 (retinal cell line), SF268 (neuronal cell line), RD (muscle cell line), JEG-3 (placental cell line), Caco-2, Hep-2, HLF (pulmonary cell line) and hepatic Huh-7 cells, thus explaining the presence of viral RNA in a wide variety of tissues in animals and humans infected with ZIKV. Furthermore, ZIKV MR766 and ZIKV PRVABC59 can also replicate in DF-1, RK-13, BHK-21, LLC-MK2, and Vero cells, all of which are nonhuman cell lines.  In contrast, human myeloid U937 cells do not support the replication of either ZIKV H/PF/2013 or ZIKV HD78788, with only less than 0.6% of infected cells staining positive for viral antigen at 48 hrs p.i.. Pre-incubation of ZIKV with convalescent serum from DENV patients increases the percentage of cells that stain positive for ZIKV to 36.9% (ZIKV H/PF/2013) and 59.0% (ZIKV HD78788) respectively whereas the percentage of DENV-2 positive cells only increases by 13%, suggesting that ZIKV infection of non-permissive cells can be enhanced by antibodies against DENV-2 which are non-neutralizing. Pretreating DENV-2 and ZIKV MR766 with a DENV-2 derived antibody (4G2) that recognizes the E protein from various Flavivirus’ including WNV, JEV and ZIKV, increases viral replication in FcγR positive THP-1 cells, indicating that ZIKV MR766 infection in the presence of 4G2 can be enhanced. These results indicate that ADE might promote the infection of non-permissive cells with ZIKV either in a strain dependent manner or dependent on the antibody being used since different antibodies might not only bind ZIKV with different affinities but also might recognize different epitopes on the viral surface.

Detailed epitope mapping of 133 antibodies revealed that broadly speaking three different regions are recognized by these antibodies, namely the envelope-dimer epitope (EDE)-1, EDE-2 and the fusion-loop epitope. (FLE). In contrast to EDE-1, binding of antibodies to EDE-2 is dependent on the presence of an N-linked glycan at Asn153.

In the case of DENV, all antibodies bind DENV particles as determined by a capture ELISA assay. In the case of ZIKV however, FLE antibodies only bind to ZIKV H/PF/2013 but not ZIKV HD78788 whereas both are recognized by EDE-1 and EDE-2 antibodies.

Figure: Domains of ZIKV E protein 

Furthermore, monoclonal antibodies to EDE but not FLE can inhibit ADE of ZIKV H/PF/2013 infection in the presence of DENV serum, indicating that DENV antibodies bind the EDE of ZIKV and thus promote viral entry in otherwise non-permissive cell lines.

As described before, placental cells do express significant amounts the viral Axl receptor early and mid-gestation and despite high levels of IFN-l1 low levels of viral replication can be detected. It might therefore be possible that in the presence of DENV antibodies or even antibodies against YFV, JEV and WNV, ADE might increase the entry of ZIKV and thus promote ZIKV replication not only by increased viral entry but also by inhibiting RLR mediated antiviral signalling pathways including inhibiting the production of nitric oxide that otherwise inhibits the viral RNA dependent RNA Polymerase. Further studies using GFP labelled virus particles should clarify if ADE indeed does increase viral entry via the FcγR concomitant with a localisation to IFITM-2/-3 positive endosomes in both placental and non-placental cells such U937 or THP-1 cells. Interestingly, the induction of autophagy with Rapamycin has been reported to decrease DENV-2 replication in U937 cells, suggesting that upon viral entry the majority of viral particles is not degraded in this cell line; therefore in the absence of DENV antibodies, ZIKV might either not be able to enter U937 cells due to the absence of the viral receptor or alternatively viral particles might be degraded, which might be similar to syncytiotrophoblast (STB) cells infected with ZIKV Nica-1/-2 2016.

Figure: Non-neutralising antibodies my mediate viral entry in both Axl negative cells (left) and Axl positive cells (right) by ADE 

Figure: Neutralising antibodies may promote viral entry and viral degradation in absence of Axl
dependent viral entry 

In contrast to non-neutralizing antibodies found in the convalescent sera of DENV patients, ZIKV specific antibodies that recognize the domain III or the fusion-loop motif of the viral E protein neutralize ZIKV H/PF/2013, ZIKV Paraiba 2015, ZIKV P6740 (Malaysia), ZIKV Dakar 41519 and to a lesser extent ZIKV MR766 in an ELISA based assay. More importantly, one antibody, ZIKV-117, prevents intrauterine growth restriction (IUGR) in the offspring of female Ifnar-1 -/- mice sired with male Ifnar-1 +/+ mice if treated prior infection with ZIKV Dakar 41529 as well as increasing the survival of Ifnar-1 -/- mice infected with ZIKV Dakar 41529 if treated either 1 day p.i. or 5 day p.i.. Additionally, viral replication in foetal placental and brain tissue as well as maternal brain and serum is decreased, indicating that ZIKV-117 neutralising mAb decreases viral replication by preventing viral entry via the cellular receptor and by falling to induce ADE. The question remains however if in animal models DENV derived non-neutralizing antibodies do increase viral replication in placental cells.

ADE has been shown to be responsible for severe cases of DHF in neonates that have been exposed to DENV antibodies in utero. Maternal vaccination against DENV, or a previous maternal DENV/ZIKV infection therefore might increase the risk of ZIKV related complications in the neonate despite the absence of congenital infection. Vice versa, neutralizing ZIKV antibodies might increase the risk for severe forms of DHF since one clone of ZIKV neutralizing mAb stains DENV-1, DENV-2 and DENV-4 infected C6/36 cells. The question also remains if neutralizing antibodies in addition to preventing viral entry via the Axl receptor are internalized via the FcγR and degraded in the lysosome. Instead of promoting viral replication, this might initiate the presentation of viral antigens in a MHC Class-I/-II dependent manner.

Further reading

Huang X, Yue Y, Li D, Zhao Y, Qiu L, Chen J, Pan Y, Xi J, Wang X, Sun Q, & Li Q (2016). Antibody-dependent enhancement of dengue virus infection inhibits RLR-mediated Type-I IFN-independent signalling through upregulation of cellular autophagy. Scientific reports, 6 PMID: 26923481 

Nour AM, Li Y, Wolenski J, & Modis Y (2013). Viral membrane fusion and nucleocapsid delivery into the cytoplasm are distinct events in some flaviviruses. PLoS pathogens, 9 (9) PMID: 24039574

Panyasrivanit M, Greenwood MP, Murphy D, Isidoro C, Auewarakul P, & Smith DR (2011). Induced autophagy reduces virus output in dengue infected monocytic cells. Virology, 418 (1), 74-84 PMID: 21813150 

Dai L, Song J, Lu X, Deng YQ, Musyoki AM, Cheng H, Zhang Y, Yuan Y, Song H, Haywood J, Xiao H, Yan J, Shi Y, Qin CF, Qi J, & Gao GF (2016). Structures of the Zika Virus Envelope Protein and Its Complex with a Flavivirus Broadly Protective Antibody. Cell host & microbe, 19 (5), 696-704 PMID: 27158114 

Charles AS, & Christofferson RC (2016). Utility of a Dengue-Derived Monoclonal Antibody to Enhance Zika Infection In Vitro. PLoS currents, 8 PMID: 27660733 

Dejnirattisai W, Supasa P, Wongwiwat W, Rouvinski A, Barba-Spaeth G, Duangchinda T, Sakuntabhai A, Cao-Lormeau VM, Malasit P, Rey FA, Mongkolsapaya J, & Screaton GR (2016). Dengue virus sero-cross-reactivity drives antibody-dependent enhancement of infection with zika virus. Nature immunology, 17 (9), 1102-8 PMID: 27339099 

Chan JF, Yip CC, Tsang JO, Tee KM, Cai JP, Chik KK, Zhu Z, Chan CC, Choi GK, Sridhar S, Zhang AJ, Lu G, Chiu K, Lo AC, Tsao SW, Kok KH, Jin DY, Chan KH, & Yuen KY (2016). Differential cell line susceptibility to the emerging Zika virus: implications for disease pathogenesis, non-vector-borne human transmission and animal reservoirs. Emerging microbes & infections, 5 PMID: 27553173 

Sapparapu G, Fernandez E, Kose N, Cao B, Fox JM, Bombardi RG, Zhao H, Nelson CA, Bryan AL, Barnes T, Davidson E, Mysorekar IU, Fremont DH, Doranz BJ, Diamond MS, & Crowe JE (2016). Neutralizing human antibodies prevent Zika virus replication and fetal disease in mice. Nature PMID: 27819683 

Tabata T, Petitt M, Puerta-Guardo H, Michlmayr D, Wang C, Fang-Hoover J, Harris E, & Pereira L (2016). Zika Virus Targets Different Primary Human Placental Cells, Suggesting Two Routes for Vertical Transmission. Cell host & microbe, 20 (2), 155-66 PMID: 27443522

Savidis G, Perreira JM, Portmann JM, Meraner P, Guo Z, Green S, & Brass AL (2016). The IFITMs Inhibit Zika Virus Replication. Cell reports, 15 (11), 2323-30 PMID: 27268505