Monomeric IgA Antagonizes IgG-Mediated Enhancement of DENV Infection

Dengue virus (DENV) is a prevalent human pathogen, infecting approximately 400 million individuals per year and causing symptomatic disease in approximately 100 million. A distinct feature of dengue is the increased risk for severe disease in some individuals with preexisting DENV-specific immunity. One proposed mechanism for this phenomenon is antibody-dependent enhancement (ADE), in which poorly-neutralizing IgG antibodies from a prior infection opsonize DENV to increase infection of Fc gamma receptor-bearing cells. While IgM and IgG are the most commonly studied DENV-reactive antibody isotypes, our group and others have described the induction of DENV-specific serum IgA responses during dengue. We hypothesized that monomeric IgA would be able to neutralize DENV without the possibility of ADE. To test this, we synthesized IgG and IgA versions of two different DENV-reactive monoclonal antibodies. We demonstrate that isotype-switching does not affect the antigen binding and neutralization properties of the two mAbs. We show that DENV-reactive IgG, but not IgA, mediates ADE in Fc gamma receptor-positive K562 cells. Furthermore, we show that IgA potently antagonizes the ADE activity of IgG. These results suggest that levels of DENV-reactive IgA induced by DENV infection might regulate the overall IgG mediated ADE activity of DENV-immune plasma in vivo, and may serve as a predictor of disease risk.


INTRODUCTION
Dengue virus (DENV) is one of the most widespread vector-borne viral pathogens in the world. Consisting of four antigenically and genetically distinct serotypes (DENV-1, -2, -3, and -4), DENV is transmitted primarily by the tropical and subtropical mosquitoes Aedes aegypti and A. albopictus (1,2). DENV and its mosquito vectors can currently be found across Central and South America, South and South-East Asia, the Western Pacific, and sub-Saharan Africa, meaning 40% of the world's population is currently at risk of exposure and infection (1)(2)(3). Consequently, an estimated 400 million DENV infections are thought to occur every year, resulting in 100 million clinically apparent infections (2). Approximately 500,000 cases per year progress to severe dengue-characterized by thrombocytopenia, vascular leakage and hemorrhage-resulting in nearly 20,000 deaths (4)(5)(6)(7).
A distinct epidemiological feature of dengue as compared to other flaviviral diseases is the increased risk for severe disease upon heterologous secondary infection (8). While the risk factors associated with developing severe dengue upon secondary DENV exposure are complex and incompletely understood, the leading mechanistic explanation for this phenomenon is a process known as antibody-dependent enhancement (ADE) (9,10). ADE is thought to occur when poorly-neutralizing or sub-neutralizing concentrations of DENV-reactive IgG opsonize DENV and facilitate its entry into permissive F c gR-bearing cells (11). Various lines of evidence support the association of ADE with severe dengue, including increased incidence of severe dengue in infants born to dengue-immune mothers (12)(13)(14); increased viremia in interferon receptor-deficient mice or non-human primates passively immunized with anti-DENV antibodies (15,16); increased incidence of severe dengue during the second of sequential heterologous DENV outbreaks; and in patients with a narrow range of preexisting anti-DENV antibody titers (17,18). Furthermore, in-vitro assessments of serum ADE activity in DENVprimed non-human primates have been shown to correlate with viral titers following heterologous attenuated DENV infection (19).
The increased risk of severe dengue upon secondary heterologous infection also presents a challenge to vaccine development as incomplete or waning vaccine-elicited immunity may place recipients at an increased risk of developing severe dengue should they be exposed following vaccination (20). This is most significantly highlighted by the revelation that the only currently US FDA licensed DENV vaccine (Dengvaxia ® ) fails to protect previously DENV naïve individuals from infection, and can increase the risk of hospitalization with virologically confirmed dengue (21)(22)(23). Accordingly, understanding the subtleties of both natural and vaccine-elicited DENV humoral immunity is critical for further our understanding of disease risk and infection-associated immunopathogenesis.
To date, the literature on dengue serology has overwhelmingly focused on the contribution of immunoglobulin isotypes IgM and IgG to functional dengue immunity and infection-associated immunopathogenesis. During both primary and secondary DENV infection, these isotype antibodies follow a highly predictable pattern of induction, with an IgM response preceding the rise of DENV-reactive IgG, and DENV-reactive IgG reaching significantly higher titers during secondary infection (24,25). These characteristics, as well as the assumed importance of IgG-mediated ADE, have left the role of other serum antibody isotypes relatively unexamined. Notably, this includes IgA, the second most prevalent antibody isotype in serum and one that has been suggested to play a unique and non-redundant role in many viral infections (26). Most work on DENV-reactive serum IgA has focused on its potential as a diagnostic tool (27), with a small body of literature examining DENV-reactive serum IgA as a possible correlate of severe disease (28)(29)(30)(31).
Our group and others recently reported that IgA was the dominant isotype-switched antibody expressed by circulating plasmablasts during acute primary DENV infection (32,33). IgA-expressing plasmablasts were also observed in secondary dengue, but constituted a smaller fraction of the total infectionelicited immune response (32,33). Importantly, the IgA antibodies expressed by these plasmablasts exhibited comparable DENV-binding and DENV-neutralization activity to IgG antibodies derived from contemporaneous samples (32). Given the milder symptoms and lower viral burden typically associated with primary dengue relative to secondary dengue, we hypothesized that DENV-reactive IgA may play some role in limiting DENV propagation and potentially the immunemediated enhancement of disease.
To test this hypothesis, we isotype-switched pairs of monoclonal antibodies to show that conversion of IgG to IgA does not impact the ability of a monoclonal antibody to bind whole DENV virions or to neutralize DC-SIGN-dependent DENV infection of a susceptible cell line. However, while DENV-reactive IgG antibodies exhibited potent infection-enhancing activity in in vitro ADE assays, we observed that DENV-reactive IgA is incapable of mediating ADE. Additionally, we observed that adding DENV-reactive monoclonal IgA to either an enhancing concentration of monoclonal IgG or to an enhancing dilution of dengue-immune plasma antagonizes ADE in a dose-dependent fashion. These results shed new light on the role of the IgA component of the humoral response to DENV, and suggest a new avenue of prophylactic and therapeutic approaches to disease.

DENV Binding and Neutralizing Are Unaffected by Antibody F c Isotype
To assess the potential contribution of DENV-reactive IgA to a functional anti-DENV humoral immune response, we synthesized two pairs of previously described DENV-reactive monoclonal antibodies with either an IgG or an IgA Fc domain ( Figure 1A). Both mAbs selected for this analysis were previously determined to bind the fusion loop of the DENV E protein and to react with all 4 DENV serotypes (32). However, VDB33 was initially identified as an IgG clone, while VDB50 was discovered as an IgA clone ( Table 1). This cross-conversion strategy was chosen so as to determine if the native F c configuration of a given antibody influenced its functionality as either an IgG or IgA protein product.
The DENV-binding capacity of the IgG and IgA versions of VDB33 and VDB50 was initially assessed with a DENV virioncapture ELISA. For this analysis, DENV-3 was chosen as the prototypic DENV serotype as previous work demonstrated that the IgG versions of both VDB33 and VDB50 exhibited significant DENV-3 reactivity (32). Consistent with previously published reports, both VDB33 and VDB50 exhibited potent DENV-3 binding activity with VDB33 demonstrating~200 fold higher affinity for DENV-3 than VDB50 ( Figure 1B and Table 2). However, the DENV-binding capacity of the two mAbs was not impacted by their conversion to either an IgG or IgA format   Figure 1B and Table 2). Furthermore, this cross-conversion of VDB33 and VDB50 to either an IgG or IgA format minimally impacted the DENV-3 neutralization activity of the clones when assessed using a flow cytometry-based neutralization assay ( Figure 1C and Table 2). These results indicate that both IgG and IgA isotype antibodies are equally capable of binding and neutralizing DENV, reaffirming that antibody epitope/paratope interactions occur independently of an antibody's F c domain.

DENV-Reactive IgA Is Incapable of Mediating ADE
Having demonstrated that the antigen binding and neutralization capacity of DENV-reactive monoclonal antibodies is negligibly impacted by the isotype of the construct, we endeavored to determine if the infection-enhancing capability of these antibodies was impacted by their isotype conversion. To this end, we utilized a K562-based ADE assay, wherein antibody/ DENV immune complexes were pre-formed and added to the F creceptor expressing K562 cell line to assess the ability of defined antibody complexes to enhance DENV infection. The IgG versions of both VDB33 and VDB50 exhibited potent infection-enhancing activity in the K562 ADE assay, with both antibodies capable offacilitating DENV infection/enhancement in a dose-dependent fashion (Figures 2A-D). Consistent with their relative EC 50 /IC 50 values, VDB33-IgG exhibited notably higher ADE activity than VDB50-IgG, but with the peak of ADE activity occurring at a similar antibody concentration. However, no infection enhancement was observed when the same assay was performed with either VDB33-IgA or VDB50-IgA (Figures 2A-D). This was despite the fact that these IgA isotype antibodies exhibit nearly identical virus binding and neutralization activity as their IgG counterparts, underlining the obligate role of an antibody's F c domain in determining the ADE potential of an antibody.

DENV-Reactive IgA Antagonizes IgG-Mediated Enhancement of DENV Infection
In light of the inability of VDB33-IgA and VDB50-IgA to facilitate ADE of DENV-3, we next endeavored to determine how DENVreactive IgG and IgA behave in a competitive setting. IgG and IgA antibodies are never found in isolation in a dengue immune individual, so determining how these antibodies function in a complex/poly-immune setting is critical for understanding their potential contribution to functional anti-DENV immunity.
To this end, we utilized the same K562 ADE assay as previously described, but used a fractional IgG/IgA replacement strategy wherein the total amount of antibody remained the same across the different titration schemes but the ratio of IgG to IgA was varied from 100:0 to 0:100. The fractional addition of DENV-reactive IgA significantly reduced the ADE activity observed in cultures containing either VDB33-IgG or VDB50-IgG (Figure 3). While both VDB33-IgA and VDB50-IgA were capable of antagonizing IgG-mediated ADE of DENV-3, the highly avid yet non-enhancing VDB33-IgA antibody was capable of dramatically blunting IgGmediated ADE even when used at low fractional concentrations. Of note, the addition of DENV-reactive IgA to these ADE assays does not appear to shift the antibody dilution at which maximal ADE activity is observed for any of the cultures. Rather, the addition of DENV-reactive IgA reduces the magnitude of infection achieved at any given antibody dilution. These results are consistent with IgA actively antagonizing IgG mediated ADE by competing with DENV-reactive IgG for the same viral epitopes.

DENV-Reactive IgA Antagonizes DENV-Immune Plasma Mediated Enhancement of DENV Infection
A limitation of the analysis presented thus far is that all the monoclonal antibodies used in this analysis have the same antigen specificity; namely, the fusion loop of the DENV E protein. Therefore, it is unclear what impact-if any-DENVreactive IgA would have in the presence of a polyclonal IgG repertoire of divergent DENV antigen specificity. Therefore, we endeavored to determine how the presence of either VDB33-IgA or VDB50-IgA impacts the infection-enhancing potential of polyclonal/DENV-immune serum.
Plasma from DENV-immune donors was screened to identify samples with both high DENV-3 reactive IgG titers by ELISA as well as DENV-3 enhancing activity in the K562 ADE assay. Samples from four subjects were selected for additional analysis based on these criteria ( Figure 4A, Supplementary Figures 3, 4). VDB33-IgA or VDB50-IgA were then titrated into cultures containing this enhancing DENV-immune plasma to determine if IgA isotype monoclonal antibodies could antagonize polyclonal enhancement of DENV-3 infection.
Consistent with what was observed with IgG monoclonal antibodies, the addition of VDB33-IgA or VDB50-IgA significantly suppressed ADE-mediated K562 infection with DENV-3 ( Figures 4B, C). The additional of DENV-reactive IgA in these assays suppressed ADE-mediated infection by 75%-90% in a dose-dependent fashion, a result consistent with the concept that IgG antibodies targeting the fusion loop of the DENV E protein are particularly amenable to facilitating ADE activity and are abundant in DENV-immune serum (34,35). Notably, an equal amount of isotype control human serum IgA had no impact on the ADE activity, suggesting that the observed suppression via VDB33-and VDB50-IgA is a function of antigen binding rather than F c -mediated (Supplementary Figure 5).These data also indicate that even modest concentrations of DENV-reactive IgA can significantly antagonize polyclonal IgG-mediated enhancement of DENV infection, signifying that the presence of DENV E reactive IgA (especially fusion loop reactive IgA) has the potential to significantly modulate DENV infection and associated immunopathogenesis.

DISCUSSION
In this study we demonstrate that DENV-reactive IgA monoclonal antibodies can bind and neutralize DENV but are incapable of facilitating ADE of DENV infection in vitro. Furthermore, the presence of DENV-reactive IgA can significant blunt the DENVinfection enhancing activity of both DENV-reactive monoclonal IgG and polyclonal DENV-immune serum in a completive fashion. These results suggest an unappreciated role for DENV-reactive IgA during the humoral response to DENV infection and raise the potential that IgA could act as either a natural or therapeutic regulator of DENV dissemination and infection-attendant inflammation.
Although we have shown that DENV-reactive IgA is capable of disruption IgG mediated ADE, IgA may not be unique in this respect. Indeed, the depletion of IgM from flavivirus-immune serum has been shown to increase antibody-dependent enhancement of Zika virus infection of K562 cells, presumably by removing IgM as an antagonist of IgG-mediated infection enchantment (36). Accordingly, parallel lines of evidence suggest that the ability of DENV-reactive IgG to interact with FcgRs is linked to the infection-enhancing potential of DENV-immune serum andby extensionthe clinical severity of DENV infection. Polymorphisms in FcgRIIaa component of the lowaffinity IgG receptor complexhave been associated with a decreased likelihood of becoming either symptomatically infected or progressing to severe dengue after DENV exposure (37)(38)(39). While the mechanism behind the phenomenon is hasn't been definitively established, it has been shown that at least some of these polymorphisms decrease the affinity of FcgRIIa for IgG (40). Furthermore, post-translational modifications of IgG antibodies have been shown to significant impact their affinity for FcgR complexes and to correlate with dengue severity (41). Most notably, the presence of high levels of afucosylated IgG-a posttranslational modification which increases the affinity of the IgG Fc domain for FcgRIIIa (42)either before or after DENV infection has been associated with increased dengue severity (43)(44)(45). Finally, the presence of serum complementsuch as C1q and C3 -can inhibit IgG mediated ADE both in vitro and in vivo, ostensibly by  interfering with the ability of IgG Fc to interact with FcgR and/or forcing a complement-bound antibody into a configuration that is not amenable to fusion and viral entry (46)(47)(48).
While not found in nature, abolishing the ability of IgG antibodies to interact with FcgR through genetic engineering the Fc portion of IgG (LALA mutation) also ablates many of the infection-enhancing properties of DENV-reactive IgG both in vitro (35) and in vivo (49,50). Collectively, these this evidence underlines the importance of IgG/FcgR interactions the process of ADE, and emphasize the potential diagnostic and therapeutic implications for factors that can disrupt this immunologic nexus.
The data presented herein demonstrate that DENV-reactive IgA is capable of antagonizing IgG mediated enhancement of DENV infection, yet it is still unclear what role this process plays in vivo during natural DENV infection. Several previous studies have noted that high levels of serum IgA are associated with more severe disease following secondary DENV infection (28)(29)(30)(31). However, we and others have observed a higher frequency of IgA expressing plasmablasts following uncomplicated primary DENV infection than following severe secondary DENV infections (32,33). A key takeaway from the analysis performed in this study is that the absolute concentration of a given DENV-reactive antibody isotype may be a incomplete indicator of the infection-enhancing potential of a serum sample. Severe dengue is accompanied by robust production of antibodies of all isotypes, so without additional context the absolute level of any given single antibody isotype may provide an incomplete or misleading impression of immunologic features associated with disease severity.

Viruses
DENV-3 (strain CH53489) propagated in Vero cells were utilized for ELISA, FlowNT50, and ADE assays. Virus for ELISA was purified by ultracentrifugation through a 30% sucrose solution and the virus pellet was resuspended in PBS.

Monoclonal Antibodies and Serum
The variable regions from the heavy and light chains were codon optimized, synthesized in vitro and subcloned into a pcDNA3.4 vector containing the human IgG1 or IgA1 Fc region by a commercial partner (Genscript). Transfection grade plasmids were purified by maxiprep and transfected into a 293-6E expression system. Cells were grown in serum-free FreeStyle 293 Expression Medium (Thermo Fisher), and the cell supernatants collected on day 6 for antibody purification. Following centrifugation and filtration, the cell culture supernatant was loaded onto an affinity purification column, washed, eluted, and buffer exchanged to the final formulation buffer (PBS). Antibody lot purity was assessed by SDS-PAGE, and the final concentration determined by 280 nm absorption. The clonotype information for all monoclonal antibodies generated as part of this study is listed in Table 1. Dengue IgG antibody positive plasma was purchased from SeraCare. Donor ID and batch numbers are shown in Supplementary Table 1. Purified polyclonal IgA from human serum was purchased from a commercial source (Sigma, I4036-1mg).

DENV-Capture ELISA
Monoclonal antibody and plasma DENV-reactivity was assessed using a 4G2 DENV capture ELISA protocol. In short, 96 well NUNC MaxSorb flat-bottom plates were coated with 2 mg/ml flavivirus group-reactive mouse monoclonal antibody 4G2 (Envigo Bioproducts, Inc.) diluted in borate saline buffer. Plates were washed and blocked with 0.25% BSA + 1% Normal Goat Serum in PBS after overnight incubation. DENV-3 (strain CH53489) diluted in blocking buffer was captured for 2 hr, followed by extensive washing with PBS + 0.1% Tween 20. Serially diluted monoclonal antibody or plasma samples were incubated for 1 hr at RT on the captured virus, and DENVspecific antibody binding quantified using anti-human IgG HRP (Sigma-Aldrich, SAB3701362), anti-human IgA HRP (Biolegend, 411,002), or anti-human IgM HRP (SeraCare, 5220-0328). Secondary antibody binding was quantified using the TMB Microwell Peroxidase Substrate System (KPL, cat. #50-76-00) and Synergy HT plate reader (BioTek, Winooski, VT). Monoclonal antibody binding data were analyzed by nonlinear regression (One site total binding) to determine EC50 titers in GraphPad Prism 8 (GraphPad Software, La Jolla, CA). End-point titers of DENV-reactive plasma samples were determined as the reciprocal of the final dilution at which the optical density (OD) was greater than 2× of a control flavivirus naïve serum.

Neutralization Assay
Neutralizing titers of monoclonal antibodies and heatinactivated plasma were assessed using a flow cytometry-based neutralization assay in U937 cells expressing DC-SIGN as previously described (51,52). Four-fold dilutions of antibody or sera were mixed with an equal volume of virus diluted to a concentration to achieve 10%-15% infection of U937-DC-SIGN cells in the absence of antibody. The antibody/virus mixture was incubated for 1 h at 37°C, after which an equal volume of medium (RPMI-1640 supplemented with 10% FBS, 1% penicillin/streptomycin, 1% l-glutamine (200 mM) containing 5 × 10 4 U937-DC-SIGN cells was added to each well and incubated 18-20 hr overnight in a 37°C, 5% CO2, humidified incubator. Following overnight incubation, the cells were fixed with IC Fixation Buffer (Invitrogen, 00-82222-49), permeabilized using IC Permeabilization Buffer (Invitrogen, 00-8333-56) and immunostained with flavivirus group-reactive mouse monoclonal antibody 4G2 (Envigo Bioproducts, Inc.), and secondary polyclonal goat anti-mouse IgG PE-conjugated antibody (#550589, BD Biosciences). The percentage of infected cells were quantified on a BD Accuri C6 Plus flow cytometer (BD Biosciences). Data were analyzed by nonlinear regression to determine 50% neutralization titers in GraphPad Prism 8 (GraphPad Software, La Jolla, CA).

ADE Assay
In vitro antibody-dependent enhancement (ADE) of DENV-3 infection was quantified as previously described (32,53). Fourfold serial dilutions of antibody or heat-inactivated sera were incubated with virus (in sufficient amounts to infect 10%-15% of U937-DC-SIGN cells) at a 1:1 ratio for 1 h at 37°C. This mixture was then added to a 96-well plate containing 5 × 10 4 K562 cells per well in duplicate. Cells were cultured for 18-20 hr overnight in a 37°C, 5% CO2, humidified incubator. Processing and quantification continued as outlined in the FlowNT50 methods.

Statistical Analysis
All statistical analysis was performed using GraphPad Prism 8 Software (GraphPad Software, La Jolla, CA). A P-value < 0.05 was considered significant.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
ADW, AR, ST, TE, MM, JC, HF, GG, and ATW contributed to conception and design of the study. ADW and HSF performed analysis. ATW and ADW wrote the first draft of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.