Passive immunotherapy of viral infections: 'super-antibodies' enter the fray

Key Points Antibodies have been used for over a century prophylactically and, less often, therapeutically against viruses. 'Super-antibodies' — a new generation of highly potent and/or broadly cross-reactive human monoclonal antibodies — offer new opportunities for prophylaxis and therapy of viral infections. Super-antibodies are typically generated infrequently and/or in a limited number of individuals during natural infections. Isolation of these antibodies has primarily been achieved by large-scale screening for suitable donors and new single B cell approaches to human monoclonal antibody generation. Super-antibodies may offer the possibility of treating multiple viruses of a given family with a single reagent. They are also valuable templates for rational vaccine design. The great potency of super-antibodies has many advantages for practical development as therapeutic reagents. These advantages can be enhanced by a variety of antibody engineering technologies. Supplementary information The online version of this article (doi:10.1038/nri.2017.148) contains supplementary material, which is available to authorized users.

The use of antibodies to ameliorate the adverse clini cal effects of microbial infection can be traced back to the late 19th century and the work of von Behring and Kitasato on the serum therapy of diphtheria and tetanus (reviewed in REFS 1,2). In these settings, the antibodies act to neutralize bacterial toxins. Therapies followed in which serum antibodies were targeted directly against bacterial and then viral pathogens. For viral pathogens, enriched polyclonal IgG molecules from immunized animals were shown to be effective in prophylaxis, and even prophylaxis after exposure, for a number of viruses, including hepatitis A virus, hepatitis B virus, hepatitis C virus, herpes simplex virus, measles virus, rabies virus, respiratory syncytial virus (RSV), smallpox virus and varicella zoster virus. In general, the effective ness of antibody preparations declined with the dura tion of infection such that they were often regarded as poor therapeutic options. Of course, the major antiviral strategy of the 20th century was vaccination.
Over the latter part of the 20th and early part of the 21st century, there have been major developments in our understanding of antibodies and our ability to manipulate them. The advent of hybridoma technol ogy in 1976 provided a reliable source of mouse mono clonal antibodies (mAbs), the first impact of which was not on antibody therapy but on the characterization of cells through the definition of cell surface markers. Broad implementation of mAbs in therapy had to wait until the development of humanized mouse antibodies and then the generation of fully human antibodies by vari ous techniques described below. Such antibodies have been largely applied in the fields of oncology and auto immunity. Only a single antiviral mAb, the RSVspecific antibody palivizumab, is in widespread clinical use. The reasons for this have been discussed elsewhere 1,3-6 , although perhaps the most important reasons are the fairly high cost of the production of mAbs, the difficul ties of administration and a belief that antibodies are largely effective only in a prophylactic setting, which can be achieved for many viruses by vaccination.
However, as we discuss here, an increasing number of antiviral antibodies with quite remarkable properties in terms of potency and/or cross reactivity with other viruses or strains of the same virus are being isolated. These socalled superantibodies are changing our understanding of what we can hope to achieve with anti bodies against microbial infection in the clinic. Increased potency can greatly reduce the unit costs of treatment, make alternative routes of administration feasible and extend the effective halflife of the antibody. Increased cross reactivity can allow us to consider targeting multi ple viruses with single antibodies. Antibody engineering can impact both potency and cross reactivity and can greatly extend the halflife of superantibodies.
In this Review, we discuss how new approaches have fuelled the identification of superantibodies, where and how such antibodies may be best applied and future directions for the field.

Super-antibody discovery
Many acute viral infections induce robust neutralizing antibody responses in the large majority of individuals. In general, these viruses show little evidence for evasion of antibody responses, and we have referred to them as 'evasion lite' viruses 7 . Typically, they either display limited antigenic variability in their surface protein (or proteins) or show considerable variability but never theless express immunodominant, conserved epitopes. Examples of viruses in this category include measles virus, poliovirus, chikungunya virus and RSV [8][9][10][11] . The life cycle of these viruses presumably does not dictate immune evasion. For these types of viruses, the iso lation of superantibodies from immune donors has been achieved in a fairly straightforward manner 8,9,[12][13][14] . However, some viruses have evolved mechanisms to evade effective neutralizing antibody responses (termed 'evasion strong' viruses) and induce such responses at much lower levels. Effective responses in the context of infection with highly antigenically variable viruses refers not only to their neutralization potency but also to their effectiveness against diverse circulating global isolates, often referred to as breadth. For these viruses -includ ing HIV, influenza virus, Ebola virus and Lassa virus -only a proportion of infected individuals, sometimes quite small, will generate broad and potent neutralizing antibody responses [15][16][17][18][19][20][21] . Furthermore, within these indi viduals, potent broadly neutralizing antibody (bnAb) specificities generally constitute only a small fraction of the antigenspecific memory B cell pool.
For example, only a small percentage of HIVinfected individuals develop broad and potent serum responses over time, and B cell cloning efforts have demonstrated that bnAbs generally comprise <1% of the HIV enve lope (Env)specific memory B cell repertoire 22 . Although there are probably multiple factors that contribute to the low abundance of bnAbs within these individuals, the intrinsic nature of the viral Env protein likely has a key role. The HIV Env protein has evolved a multitude of mechanisms to evade bnAb responses, including decoy forms of Env, enormous antigenic variability, an evolving glycan shield, immunodominant and variable epitopes and poorly accessible conserved epitopes 23 . Furthermore, most HIVspecific bnAbs incorporate unusual features for epitope recognition, such as uncommonly long (or short) complementarity determining region 3 (CDR3) loops, insertions and deletions, tyrosine sulfation and extensive somatic hypermutation, that likely also con tribute to the rarity and delayed development of bnAbs during natural infection 24-26 .
In the case of influenza virus infection, the vast majority of neutralizing antibodies elicited by infec tion or vaccination bind to variable epitopes within the haemagglutinin (HA) globular head of the viral parti cles and display strainspecific neutralizing activity 27 . Influenza virusspecific bnAbs typically target the con served HA stem, but this region has variable immuno genicity 28-31 . The relatively low frequency of bnAbs against the HA stem is perhaps due to the typically tight packing of HA trimers on the virus surface, which may limit antibody accessibility to this region. For Ebola and Lassa viruses, extensive glycosylation on the surface Env proteins results in masking of conserved neutral izing epitopes 15,32 . In cases where superantibodies are present at low frequency within immune repertoires, largescale donor screening and highthroughput B cell isolation platforms have proved to be critical for the dis covery of superantibodies. Over the past several years, technological advances in these two areas have led to the identification of large numbers of superantibodies, mostly from infected individuals, against a plethora of viral pathogens.
Large-scale donor screening. In the case of HIV, which has served as a prototype virus for many studies in this field 33 , systematic selection of donors with broadly neutralizing serum responses has proved to be critical for the identification of superantibodies. Before 2009, the HIV field had been operating with a handful of bnAbs, all of which were limited either in breadth or in potency 34 . A number of factors complicated the iden tification of bnAbs, including the inefficiency of tradi tional approaches to mAb discovery, the small fraction of B cells that secrete bnAbs and the limited availability of samples from donors who had developed broad and potent neutralizing serum responses. Beginning in 2005, the problem of limited samples was addressed by estab lishing donor screening programmes to identify HIV infected individuals with broadly neutralizing serum responses to serve as source material for the generation of bnAbs 18,19,35,36 . In one of the largest studies, ~1,800 HIVinfected individuals from Australia, Rwanda, Uganda, the United Kingdom and Zambia were screened for broadly neutralizing sera using a reduced pseudo virus panel representative of global circulating HIV iso lates 19 . A subset of individuals, termed 'elite neutralizers' , was identified that exhibited exceptionally broad and potent neutralizing serum responses and was therefore prioritized for bnAb isolation.
Over the past 8 years, mining of these and sim ilar samples has led to the identification of dozens of remarkably broad and potent HIV superantibodies 37-41 . Careful selection of donors with desirable serum profiles has also enabled the isolation of rare super antibodies to influenza virus, RSV, human metapneumovirus (HMPV), rabies virus and Zika virus 12,[42][43][44] . For exam ple, the paninfluenza A virusneutralizing mAb FI6 and the RSV and HMPV crossneutralizing mAb MPE8 were isolated from donors who were selected on the basis of their strong heterotypic serum responses 12,42 . Similarly, two panlyssavirusneutralizing mAbs, called RVC20 and RVC58, were isolated from the memory B cells of four donors who exhibited potent serumneutralizing activity against multiple lyssavirus species 43 .

High-throughput human B cell isolation technologies.
Human antiviral neutralizing mAbs have been isolated using various different technologies, including combina torial display libraries, human immunoglobulin trans genic mice and single B cell isolation methods (FIG. 1). Although all of these technologies have proved valuable for mAb generation, the recent burst in superantibody Human antibody heavy-chain and light-chain genes are amplified by reverse transcription (RT)-PCR, and antibody fragments are displayed on the surface of a particle or cell in which the antibody genes are found (such as phage, yeast or mammalian cells [142][143][144][145] ). Successive rounds of enrichment are performed to select for clones that bind to the target antigen. Genes encoding antibodies of interest are cloned into human IgG expression vectors to produce monoclonal antibodies (mAbs). b | Human immunoglobulin transgenic mice are generated by introducing human immunoglobulin loci into the mouse genome 146,147 . Upon immunization, the transgenic mice produce fully human antigen-specific antibodies. discovery has primarily been driven by advances in single B cellbased methods. There are several possi ble reasons for this, including inefficiencies in combi natorial library generation and interrogation (leading to the loss of rare clones), altered binding characteristics of antibody fragments produced in heterologous expres sion systems (for example, Escherichia coli or yeast), constraints on the generation of suitable recombinant antigens for immunization or library selections, the loss of native heavychain and lightchain pairing during immune library generation and inherent differences between the adaptive immune systems of humanized mice and humans 45 . Several technological breakthroughs in the B cell cloning arena have been most critical in fuelling superantibody identification. One of these advances came in 2009, when direct functional screening of thousands of B cell clones from an HIV elite neutralizer led to the isolation of two superantibodies, PG9 and PG16, that were approximately an order of magnitude more potent than firstgeneration bnAbs 37 . Notably, PG9 and PG16 bind poorly to recombinant Env pro teins and thus would not have been identified without direct functional screening of B cell supernatants. To date, many dozens of HIVspecific bnAbs targeting a diverse range of epitopes have been identified using functional screening approaches 39,46-48 . Highly potent superantibodies to RSV, HMPV, Lassa virus and human cytomegalovirus (HCMV) have also been dis covered using highthroughput functional screening technologies 12,13,49,50 (TABLE 1).
Similar to PG9 and PG16, these superantibodies were isolated by screening B cell supernatants on the basis of their capacity to neutralize infection in vitro and were subsequently found to react poorly with currently available recombinant Env proteins. In the case of RSV, this approach led to the isolation of the highly potent mAb D25, which binds to an epitope that is exclusively expressed on the prefusion conformation of RSV fusion glycoproteins (F proteins) 49,51 . An engineered variant of mAb D25 (MEDI8897), which exhibits 50-100 times greater neutralization potency than palivizumab, is now being tested in clinical trials for the prevention of RSVassociated disease in highrisk infants. A second RSV prefusion F proteinspecific superantibody, which crossneutralizes several different paramyxoviruses, including HMPV, was also isolated by screening B cell supernatants for neutralizing activity 12 (TABLE 1).
In the case of HCMV, direct functional screening enabled the isolation of highly potent superantibodies specific for conformational epitopes within the gH-gL-UL128-UL130-UL131A pentamer complex, which was not previously known to be a target for neutralizing anti bodies 13 . Direct functional screening approaches have also led to the discovery of potent superantibodies to Middle East respiratory syndrome coronavirus (MERS CoV), Ebola virus, influenza virus, chikungunya virus, rabies virus and the poxvirus family 9,42,43,52-56 . Notably, in the case of MERSCoV, only 1 B cell culture out of 4,600 screened showed neutralizing activity 53 . Similarly, the pan influenza A virusneutralizing mAb FI6 was isolated by testing 104,000 plasma cells from eight immune donors 42 . Finally, only 2 of 500 mAbs that were selected on the basis of their ability to neutralize rabies virus showed crossneutralizing activity against multiple lyssavirus species 43 . These examples clearly illustrate that exhaustive interrogation of immune repertoires is often required for the identification of rare crossneutralizing superantibodies.
A second breakthrough in the HIV antibody field followed the development of technology for antigen specific single B cell sorting 57-59 . This approach, coupled with the use of rationally designed Env probes, allowed for the discovery of two new potent HIVspecific bnAbs that target the conserved CD4 binding site 60,61 (TABLE 1). Following this discovery, several other potent bnAbs against the CD4 binding site were isolated through the use of similar approaches 38,62,63 . Recently, advances in the generation of recombinant nativelike HIV Env trimers have enabled the identification of exceptionally potent 'PG9class' bnAbs 40 .
Many HIV superantibodies have now been gen erated with the use of single B cell sorting technol ogy 38,40,60,62,63 . The use of fluorescently labelled probes to sort antigenspecific memory B cells has also ena bled the discovery of highly potent superantibodies to Ebola virus, RSV, human papilloma virus (HPV), Zika virus and influenza virus 11,30,31,44,64,65 . In the case of Ebola virus, a largescale single B cell cloning effort led to the isolation of several hundred mAbs specific for Ebola virus envelope glycoprotein (GP), two of which showed potent panEbola virusneutralizing activity and protec tive efficacy 64,66 . A similar effort in the RSV field allowed for the isolation of several prefusion Fproteinspecific mAbs that show over 100 times more potent neutral izing activity than palivizumab 11 . In addition, multiple groups have used clever dualantigen labelling strategies to identify potent bnAbs to HIV, influenza virus, Ebola virus and HPV 30,31,40,65,67,68 . The structures of several super antibodies bound to their viral targets are shown in FIGURE 2. Finally, a recent report showed that bnAbs to HIV can be readily elicited in cows through the use of a single Env trimer immunogen and that this induc tion depends on the long heavychain CDR3 loops of the bovine immunoglobulin repertoire 67 . It is possible that this repertoire may provide advantages in generating superantibodies against other pathogens.
Vaccination or infectioninduced antibodysecreting cell (ASC) responses have also proved to be a rich source of antigenspecific antibodies. Following early studies that showed that a transient but large population of ASCs appears in peripheral blood 5-7 days after tetanus toxoid booster vaccination 69 , it was shown that influ enza virus vaccination produced a similar ASC response and that the large majority of mAbs cloned from these cells bound with high affinity to influenza virus, provid ing a proof of concept that the ASC response could be exploited to rapidly generate antigenspecific antibodies against any immunizing antigen 70 . To date, plasmablast cloning has led to the isolation of mAbs against many different viruses, including dengue virus, Zika virus, HIV, influenza virus, vaccinia virus and rotavirus 29,71-76 .
One of the advantages of the plasmablast approach is that antigen baiting is not required for B cell sorting, thereby allowing for the isolation of antibodies that target epitopes that are poorly presented on recom binant antigens. For example, in the case of dengue virus, potent bnAbs targeting Edimerdependent epitopes were isolated using this approach 72 . However, it is important to emphasize that the ability to isolate superantibodies using this method will depend on several factors. First, during a primary infection, the ASC population will be mainly composed of activated, lowaffinity naive B cells (rather than affinitymatured memory B cells), making the possibility of identifying superantibodies extremely unlikely. Second, in the context of a booster vaccination or secondary infec tion with an antigenically similar virus, most of the plasmablast response will be directed against immuno dominant epitopes, which in many cases are not tar geted by effective neutralizing antibodies. In such cases, exhaustive cloning, production and characterization of the plasmablastderived mAbs would likely be required to identify rare superantibodies. By contrast, secondary infection with an antigenically related but sufficiently divergent virus can drive the preferential expansion of B cells that target highly conserved epitopes, as exemplified by the unusually high frequency of bnAbs induced in donors who were infected with the novel H1N1 influenza virus in 2009 (REFS 29,75). In principle, one could use this type of approach for the generation of superantibodies in humanized mice or other animal models by using suitably designed immunogens and immunization regimens. In response to the 2014-2015 MERSCoV outbreak, two different groups illustrated the power of their mAb discovery platforms by isolating highly potent MERS specific mAbs, producing the mAbs in gram quanti ties and testing the lead mAbs in animal models at an unprecedented speed 53,77 . In one of these studies, a sin gle highly potent MERSCoVneutralizing mAb was identified from the memory B cells of a convalescent donor through the use of a highthroughput functional screening approach 53 . This mAb, called LCA60, showed protection both before and after exposure in a mouse model of MERSCoV infection. Importantly, this process took the authors only 4 months from the initial B cell screening to the development of a stable cell line that produces the neutralizing mAb at 5 g l -1 . In the second study, human immunoglobulin transgenic mice were immunized with the MERSCoV spike protein and then used to generate a panel of potent MERSCoVspecific neutralizing mAbs within several weeks 77 . The authors also quickly generated a humanized mouse model of MERSCoV infection, which was used to demonstrate the therapeutic efficacy of their mAbs.
In a third study, vaccination of transchromosomal cows engineered to produce fully human IgG mole cules with MERSCoV was shown to yield high serum titres of MERSCoVspecific neutralizing antibodies 78 . Importantly, administration of the purified polyclonal transchromosomal bovine human IgG to mice either 12 hours before or 24 and 48 hours after MERSCoV infection resulted in a significant reduction in viral lung titres. Transchromosomal bovines have also been used to rapidly generate polyclonal neutralizing antibodies to Hanta virus, Venezuelan equine encephalitis virus and Ebola virus [79][80][81] , demonstrating the feasibility of using this platform to rapidly generate therapeutics to com bat emerging viral threats. The antibodies arising from trans chromosomal cows are polyclonal to date, but there is potential for mAb isolation.
Between 2015 and 2016, several groups responded to the 2014-2015 Ebola virus outbreak by swiftly gener ating highly potent Ebola virus GPspecific neutral izing mAbs from the memory B cells of convalescent donors 54-56,64,66 . Many of these mAbs showed potent therapeutic efficacy against either Ebola or Bundibugyo virus after exposure in animal models and a subset showed protective efficacy against multiple Ebola virus strains 54,55,64,66 . Similarly, several groups have recently reported on the isolation of potent Zikavirusspecific neutralizing mAbs from human donors 14,44,82,83 . Notably, one of these neutralizing mAbs, called ZIKV117, showed protection against Zika virus after exposure in both pregnant and nonpregnant mice 82 .
In certain cases, the availability of super antibodies that target highly conserved epitopes may shorten time lines further by bypassing the need for mAb discovery. For example, it was recently shown that a subset of dengue virusspecific mAbs potently cross neutralizes Zika virus [84][85][86] . These bnAbs -perhaps carrying Fc mutations that ablate Fc receptor binding to avoid the potential for antibodydependent enhancement 14could immediately be used for prophylaxis for preg nant women living in Zikavirusendemic regions. Notably, cocktails of superantibodies targeting dif ferent epitopes, or bispecific or trispecific superan tibody constructs, will likely be required to prevent neutralization escape [87][88][89][90][91][92][93] .

Antibodies in prophylaxis and therapy
Antibodies can function against viruses by several mechanisms, primarily divided into activities against Nature Reviews | Immunology free virus particles and activities against infected cells. Neutralization, measured in vitro as the abil ity of an antibody to prevent viral entry into target cells without a requirement for involvement of any other agents, is an activity against free virions that has been most correlated with protection in vivo.

Activities against infected cells generally depend on
Fc effector functions and involve host effector cells. They include antibody-dependent cellular cyto toxicity (ADCC), complement-dependent cytotoxicity and antibody dependent cellular phagocytosis (ADCP). It is presumed that these activities are likely to be impor tant in antibody based therapy. Because neutral ization frequently correlates with the ability to bind to native structures on the virion surface, it can give some indi cation of the ability of antibodies to mediate effector activities such as ADCC and ADCP. The potency of super antibodies is then often estimated from neu tralization measurements, although ultimately it is of course in vivo activity that is crucial.

Prophylaxis.
Vaccination is the most effective and low cost method of preventing viral disease. However, the development of effective vaccines against many impor tant viral pathogens -including HIV, RSV, hepatitis C virus, dengue virus and HCMV -has been met with limited success. Furthermore, vaccine development is a long complex process, often lasting 10-15 years, making immunization an impractical means of pro tecting individuals from newly emerging viral threats unless panvirus family vaccines can be developed. For example, if antibodies can be identified that potently neutralize existing strains of Ebola virus, or even Ebola and Marburg filoviruses, then one could anticipate that a vaccine templated from the antibodies would also be effective against emerging strains of Ebola virus. However, at the current time, passive antibody prophy laxis represents a promising alternative to vaccination for a number of viral infections. Currently, three purified polyclonal hyperimmune globulins derived from human donors immune to hepa titis B virus, HCMV or varicella zoster virus are on the market for the prevention of serious diseases associated with these viruses. A rabiesvirusspecific immune globulin, combined with vaccination, is also available for prophylaxis after exposure. In 1998, palivizumab -a humanized mAb that targets the RSV F proteinbecame the first antiviral mAb approved by the US Food and Drug Administration. Palivizumab soon replaced the RSV hyperimmune globulin (RespiGam) for the prevention of severe RSVassociated disease in high risk infants. However, although palivizumab is more specific and 50-100 times more potent than RespiGam, the cost associated with the required dosing makes its use impractical for all infants 94 . A secondgeneration RSVspecific mAb, which shows up to 50 times greater neutralization potency than palivizumab and contains substitutions in the Fc domain that extend its serum halflife, is currently in phase II clinical trials for the pre vention of severe RSVassociated disease in all infants 95 (TABLE 2).
Although palivizumab is the only commercially available mAb for the prevention of a viral disease, there are multiple antiviral mAbs in preclinical and clinical development that have shown efficacy prior to expo sure in animal models. For example, the potent MERS CoVspecific mAbs described above were shown to prophylactically protect humanized mice against MERS associated disease. Similarly, mAbs to chikungunya virus, influenza virus, HIV and Ebola virus have shown potent prophylactic efficacy in animal models 9,85,[96][97][98][99][100][101][102] . Recently, a broadly neutralizing antiZika virus mAb (ZIKV117) was shown to protect against maternal-fetal transmission in a mouse model of Zika virus infection 82 . If this observation translates to humans, prophylaxis with ZIKV117 or similar neutralizing mAbs may be a promising means of protecting atrisk pregnant women against Zika virus infection and fetal transmission.
In the case of HIV, multiple studies have shown that passively administered neutralizing mAbs pro vide protection against intravenous, vaginal, rectal and oral challenge in nonhuman primate and mouse models [99][100][101][103][104][105][106] . A large ongoing study (the Antibody Mediated Prevention (AMP) study will assess the abil ity of the VRC01 mAb specific for the CD4 binding site to decrease the risk of HIV acquisition in humans. Although animal studies have provided proof of prin ciple that a vaccine capable of inducing sufficient titres of bnAbs could prevent the establishment of HIV infec tion in humans, and the AMP study will investigate this directly, the design of immunogens that efficiently elicit these rare antibodies remains a formidable challenge.
To bypass the challenges associated with active vaccination against HIV, a number of groups have proposed alternative strategies on the basis of vec tormediated antibody gene transfer to express bnAbs in vivo 107 . Unlike traditional passive immunization, which would require longterm repeated treatment with bnAbs, vectored immunoprophylaxis involves only a single injection and enables continuous and sustained delivery of antibodies. In 2009, pioneering work demonstrated that vectormediated delivery of antibodylike molecules can provide vaccinelike pro tection against simian immuno deficiency virus (SIV) challenge in nonhuman primate models 108 . Subsequent studies have shown that vectored immunoprophylaxis is also compatible with fulllength IgG molecules and CD4like molecules [109][110][111] . If the preclinical results in mice and macaques translate to humans, vectored antibody gene delivery strategies could provide an alternative form of prophylaxis against HIV and other challenging vaccine targets, such as hepa titis C virus, pandemic influenza virus and malaria. Recently, non viral vector nucleic acid delivery technologies have also been developed to obviate the potential safety issues associated with viral vectormediated delivery, such as longterm persistence and potential viral DNA integra tion into the host genome [112][113][114][115][116][117] . In a recent study, it was shown that the administration of lipid encapsulated nucleosidemodified mRNAs encoding the heavy chain and lightchain genes of the broadly neutral izing HIV1specific antibody VRC01 to humanized mice resulted in high serum antibody concentrations and protection against intravenous HIV1 challenge 114 . Similar proofofconcept studies have also been per formed using synthetic DNA plasmid mediated anti body gene transfer 112,113 . In one such study, synthetic DNA plasmids encoding crossneutralizing anti dengue virus antibodies were delivered to mice by electro poration and resulted in biologically rele vant levels of serum antibody 112 . Importantly, a single intramuscular injection of plasmid DNA conferred protection against severe dengue disease in a mouse model. Although several technical challenges remain to be addressed, such as enhancing in vivo antibody expression levels and reducing the potential for immunogenicity, these studies demonstrate the feasibility of utilizing plasmid DNA and modified mRNAbased antibody delivery technologies for passive immunotherapy.
Therapy. Conventional wisdom says that antibodies are effective if present before or shortly after viral exposure, but their effectiveness declines markedly once infection is established. For example, the anti RSV antibody palivizumab is effective in the clinic prophylactically but not therapeutically 118 . However, there are indications that the dogma may be chal lenged by superantibodies. An example is the ability of a new generation of bnAbs against HIV to strongly impact ongoing infection in animal models 87,119,120 in which an earlier generation of lesspotent mAbs had very limited effects 121 . This ability likely reflects the increased neutralization potency of the super antibodies as well as the increased breadth of neutral ization that may restrict virus escape pathways 119 . A number of superantibodies are now being evaluated in humans for their activities against established HIV infection [122][123][124][125][126] (TABLE 2). Initial results are interesting, providing for example an indication of enhanced immune responses following bnAb administration 127 . The emerging results will be followed closely, includ ing in the context of combining bnAbs with drugs and other antiviral agents to attempt HIV cure.
For other viruses, clear evidence of a strong thera peutic effect for superantibodies has not been gath ered yet. Several cases such as antibody treatment of rabies virus and Junin virus infections 43,128 are probably better interpreted as prophylaxis after exposure rather than therapy for an established infection. Two prom ising examples of possible therapy are the successful treatment of Ebolavirusinfected or Lassavirus infected monkeys with mAbs once symptoms have appeared 89,129 . Unfortunately, no definitive evidence of the effectiveness of mAbs in Ebolasymptomatic or Lassasymptomatic humans yet exists.
Camelidderived singledomain antibodies (sdAbs), which contain a single heavychain variable domain, represent a promising new class of antibodybased therapeutics for RSV and other viruses that cause lower respiratory tract infections [130][131][132][133] . Because of their small size and high solubility and stability, sdAbs can be rap idly delivered to the site of infection via inhalation. Notably, a neutralizing antiRSV sdAb (ALX0171) that targets an epitope overlapping that bound by palivi zumab recently showed a trend towards a therapeutic effect in a phase I and IIa clinical trial based on reduced viral loads and clinical symptoms in hospitalized RSV infected infants. Prefusion Fproteinspecific sdAbs that show up to 180,000 times greater neutralization potency than ALX0171 have recently been identified and may offer even greater therapeutic benefit 133 .

Practical considerations
Intuitively, the enhanced potency of super antibodies is immediately recognized as beneficial in antibody prophy laxis and therapy. However, there are also a number of additional effects from this enhanced potency that may not be instantly appreciated and that can be further strengthened by antibody engineer ing. For example, enhanced potency means that less antibody needs to be used, and this can enable easier to develop, lowconcentration subcutaneous adminis tration rather than the use of more difficulttodevelop, high concentration sub cutaneously administered formulations or less convenient (lowconcentration) intravenous administration. Enhanced potency also means that the lifetime of an effective antibody follow ing administration is extended, thereby requiring fewer administrations to maintain a useful protective or thera peutic effect. Antibody engineering can also extend its halflife considerably 95,[134][135][136][137][138][139] so that for the most potent superantibodies, one could envisage requiring administrations perhaps only every 3-6 months for effectiveness. Antibody engineering can also deliver greater effectiveness through enhanced Fc effector function 140,141 .

Conclusions
The deployment of antibodies as antiviral agents has progressed through a number of stages over the years, corresponding to increasing levels of potency of the reagent administered. Passive immunotherapy began with immune serum over a century ago, then progressed to polyclonal antibodies, then mAbs and now into highly potent human mAbs dubbed super antibodies. Thanks to research that has been primarily carried out in the field of cancer research, technologies have been developed to endow these superantibodies with enhanced in vivo function.
Will superantibodies change the landscape of antiviral prophylaxis and therapy? The answer to this question will depend on a number of factors: first, the rapidity of development of antiviral vaccines (vaccines will likely remain the least expensive and most effec tive antiviral measure, but some viruses such as HIV present a large challenge to vaccine development); sec ond, the effective cost of antibody treatment, which incorporates not only manufacturing cost but also the durability of the administered antibody and the route of administration; and third, the success of anti bodies in the treatment of established viral infections. Particularly in the therapeutic setting, the answers can only be obtained with clinical trials using the best superantibodies available.   108-111 (2017). This article describes the reliable generation of broadly neutralizing HIV antibodies by simple immunization in cows, suggesting that this animal has special value in super-antibody generation.