crossmark

Pan-ebolavirus and Pan-filovirus Mouse Monoclonal Antibodies: Protection against Ebola and Sudan Viruses Frederick W. Holtsberg,a Sergey Shulenin,a Hong Vu,a Katie A. Howell,a Sonal J. Patel,a Bronwyn Gunn,b Marcus Karim,b Jonathan R. Lai,c Julia C. Frei,c Elisabeth K. Nyakatura,c Larry Zeitlin,d Robin Douglas,a Marnie L. Fusco,e Jeffrey W. Froude,f Erica Ollmann Saphire,e Andrew S. Herbert,f Ariel S. Wirchnianski,f Calli M. Lear-Rooney,f Galit Alter,b John M. Dye,f Pamela J. Glass,f Kelly L. Warfield,a M. Javad Amana Integrated BioTherapeutics, Inc., Gaithersburg, Maryland, USAa; Ragon Institute of Massachusetts General Hospital, Massachusetts Institute of Technology, and Harvard, Boston, Massachusetts, USAb; Albert Einstein College of Medicine, New York, New York, USAc; Mapp Biopharmaceutical, San Diego, California, USAd; Department of Immunology and Microbial Science, The Scripps Research Institute, La Jolla, California, USAe; U.S. Army Medical Research Institute of Infectious Diseases, Frederick, Maryland, USAf

ABSTRACT

The unprecedented 2014-2015 Ebola virus disease (EVD) outbreak in West Africa has highlighted the need for effective therapeutics against filoviruses. Monoclonal antibody (MAb) cocktails have shown great potential as EVD therapeutics; however, the existing protective MAbs are virus species specific. Here we report the development of pan-ebolavirus and pan-filovirus antibodies generated by repeated immunization of mice with filovirus glycoproteins engineered to drive the B cell responses toward conserved epitopes. Multiple pan-ebolavirus antibodies were identified that react to the Ebola, Sudan, Bundibugyo, and Reston viruses. A pan-filovirus antibody that was reactive to the receptor binding regions of all filovirus glycoproteins was also identified. Significant postexposure efficacy of several MAbs, including a novel antibody cocktail, was demonstrated. For the first time, we report cross-neutralization and in vivo protection against two highly divergent filovirus species, i.e., Ebola virus and Sudan virus, with a single antibody. Competition studies indicate that this antibody targets a previously unrecognized conserved neutralizing epitope that involves the glycan cap. Mechanistic studies indicated that, besides neutralization, innate immune cell effector functions may play a role in the antiviral activity of the antibodies. Our findings further suggest critical novel epitopes that can be utilized to design effective cocktails for broad protection against multiple filovirus species. IMPORTANCE

Filoviruses represent a major public health threat in Africa and an emerging global concern. Largely driven by the U.S. biodefense funding programs and reinforced by the 2014 outbreaks, current immunotherapeutics are primarily focused on a single filovirus species called Ebola virus (EBOV) (formerly Zaire Ebola virus). However, other filoviruses including Sudan, Bundibugyo, and Marburg viruses have caused human outbreaks with mortality rates as high as 90%. Thus, cross-protective immunotherapeutics are urgently needed. Here, we describe monoclonal antibodies with cross-reactivity to several filoviruses, including the first report of a cross-neutralizing antibody that exhibits protection against Ebola virus and Sudan virus in mice. Our results further describe a novel combination of antibodies with enhanced protective efficacy. These results form a basis for further development of effective immunotherapeutics against filoviruses for human use. Understanding the cross-protective epitopes are also important for rational design of pan-ebolavirus and pan-filovirus vaccines.

T

he Filoviridae family consists of a single marburgvirus species with Marburg virus (MARV) and Ravn virus (RAVV), as well as five ebolavirus species, Ebola virus (EBOV), Sudan virus (SUDV), Bundibugyo virus (BDBV), Reston virus (RESTV), and Taï Forest virus (TAFV) (1, 2). Filoviruses cause lethal hemorrhagic fever in humans and nonhuman primates (NHPs), with case fatality rates of up to 90% (3, 4). EBOV has caused the majority of filovirus hemorrhagic fever outbreaks, including the 2014 outbreak in West Africa with more than 27,000 cases and 11,000 deaths (5). However, other members of Filoviridae have also caused human epidemics, including seven outbreaks of SUDV (6), two outbreaks of BDBV (6), and 12 outbreaks of MARV (7). RESTV has not caused disease in humans, but its recent detection in pigs has raised concern about the potential emergence of ebolaviruses in the human food chain (8). Thus, there is urgent need for the development of broadly protective filovirus therapeutics, as the nature of future outbreaks cannot be predicted. Recent reports indicate that monoclonal antibodies (MAbs)

266

jvi.asm.org

against the filovirus glycoproteins (GP) represent effective postexposure treatments for Marburg virus and Ebola virus hemorrhagic fever (9–17). However, nearly all efficacious ebolavirus an-

Received 26 August 2015 Accepted 5 October 2015 Accepted manuscript posted online 14 October 2015 Citation Holtsberg FW, Shulenin S, Vu H, Howell KA, Patel SJ, Gunn B, Karim M, Lai JR, Frei JC, Nyakatura EK, Zeitlin L, Douglas R, Fusco ML, Froude JW, Saphire EO, Herbert AS, Wirchnianski AS, Lear-Rooney CM, Alter G, Dye JM, Glass PJ, Warfield KL, Aman MJ. 2016. Pan-ebolavirus and pan-filovirus mouse monoclonal antibodies: protection against Ebola and Sudan viruses. J Virol 90:266 –278. doi:10.1128/JVI.02171-15. Editor: A. García-Sastre Address correspondence to M. Javad Aman, [email protected]. F.W.H., S.S., and H.V. contributed equally to this work. For a companion article on this topic, see doi:10.1128/JVI.02172-15. Copyright © 2015, American Society for Microbiology. All Rights Reserved.

Journal of Virology

January 2016 Volume 90 Number 1

Cross-Reactive Mouse-Derived Filovirus Antibodies

tibodies are species specific, and the majority of them target EBOV. The primary amino acid sequence of GP shows nearly 30% identity between EBOV and MARV and 56 to 65% identity between the various ebolavirus species. Despite this homology, no cross-reactive antibodies have been described so far that would show cross-neutralization or cross-protective efficacy against multiple filovirus species. Over the past few years, major progress has been made toward development of effective immunotherapeutics against EBOV. These studies indicated that antibodies targeting certain key epitopes within EBOV GP act synergistically to enhance the therapeutic effects. Two antibody cocktails, MB-003 (11) and ZMab (18), were first reported to show significant postexposure efficacy in NHPs. Upon systematic evaluation of various cocktails, two components of ZMab (4G7 and 2G4) and one component of MB003 (13C6) were combined into a novel cocktail, referred to as ZMapp, which demonstrated 100% efficacy when administered as late as 5 days postinfection in NHPs (14). Recent studies using single-molecule electron microscopy revealed that 13C6 binds to a region on the top of the GP trimer known as the glycan cap, while 4G7 and 2G4 target a conformational epitope consisting of the GP2 subunit and the N terminus of the GP1 subunit within the base of the GP trimer (19) that overlaps with the site targeted by KZ52, a potent neutralizing antibody identified from an Ebola virus disease (EVD) survivor (20, 21). This antibody cocktail identifies these two epitopes as key sites of vulnerability on EBOV (19). However, it is not clear whether cross-reactive antibodies targeting these two sites can be developed. In fact, another MAb (16F6) targeting the same region within SUDV GP has been identified (22), but both 16F6 and KZ52 are strictly species specific. Thus, effective pan-ebolavirus or pan-filovirus antibodies, if possible, are likely to target novel epitopes. Here we report a set of four mouse MAbs reactive to novel conformational epitopes conserved across different species of ebolavirus as well as a MAb broadly reactive to all filoviruses, including MARV. The pan-ebolavirus MAbs showed significant protective efficacy against EBOV in a mouse infection model. One of the antibodies potently neutralized both EBOV and SUDV and showed protection against both viruses in mice, the first demonstration of a single antibody providing in vivo protection against two ebolavirus species. Combination of two of these antibodies targeting a new epitope near or on the glycan cap and another potentially involving GP2 showed enhanced efficacy in mice, suggesting that effective pan-ebolavirus therapeutic antibody cocktails can be developed. Our studies further suggest that, besides neutralization, Fc-mediated effector functions play a role in protective efficacy of the antibodies. MATERIALS AND METHODS Production and purification of filovirus glycoproteins. Plasmids for the following glycoprotein (GP) constructs were generated: GP⌬TM (GP ectodomain lacking the transmembrane region) for EBOV (Mayinga) and SUDV (Boniface) containing amino acids 1 to 627 with a hemagglutinin (HA) tag (YPYDVPDYA) followed by a factor 10a cleavage site (IEGRG AR); EBOV GP⌬muc (amino acids 1 to 311 followed by an aspartic acid and linked to residues 464 to 637); and SUDV GP⌬muc (residues 1 to 313 linked to residues 474 to 640) were generated and expressed in 293T cells. GP⌬TM proteins were purified from the supernatants as described previously (20). Briefly, proteins expressed in 293T cells were purified using Q-Sepharose Fast Flow resin. Eluates were pooled and passed over a GE S-200 HR column. S-200 fractions containing GP⌬TM were pooled and

January 2016 Volume 90 Number 1

dialyzed against phosphate-buffered saline (PBS). GP⌬muc protein was purified using Q-Sepharose Fast Flow resin, and eluted fractions containing the proteins were pooled and dialyzed against PBS. For insect cell expression of GP⌬TM for EBOV, SUDV, RESTV, and BDBV, the coding regions for amino acids 1 to 605 (with the exception that for BDBV, residues 1 to 609 were used) followed by a C-terminal 6⫻His tag were cloned into the baculovirus transfer vector pFastBac (Invitrogen). Bacmid DNA was produced as described previously (23) and used to transfect Sf9 insect cells to generate recombinant baculoviruses containing GP⌬TM, which was further amplified in Sf9 cells. The final virus was used to infect Sf9 cells for purification of the proteins from the supernatants 3 days postinfection. After separation of cell debris, the supernatants were concentrated 10 times, mixed with 2 mM CaCl2, 0.25 mM Ni2⫹, 20% glycerol, 10 mM imidazole, 0.5% Triton X-100, and 1 M NaCl, and the pH was adjusted to 7.2. Ni beads (GE Healthcare Life Sciences) were added at 1 ml per liter of concentrated supernatant and mixed overnight at 4°C. The beads were separated by centrifugation, resuspended in PBS plus 0.2% Tween 20, and packed into a Bio-Rad Econo-column. The column was washed with the following buffers: PBS (pH 7.1) supplemented with 20% glycerol and 0.2% Tween 20, followed by PBS supplemented with 20% glycerol, 0.2% Tween 20, and 10 mM imidazole. Protein was eluted with 500 mM imidazole in PBS. Eluted proteins were dialyzed against PBS supplemented with 10% glycerol containing arginine and glutamic acid (pH 7.4). For production of soluble GP (sGP), the full coding sequence of EBOV sGP including delta peptide followed by a C-terminal tag was expressed in 293T cells. The supernatants were passed through a Ni column to separate the delta peptide. The flowthrough was concentrated and used as a source of sGP. Proteolytically cleaved GP ectodomains were produced in S2 cells as described previously (24). Immunization of mice and generation of monoclonal antibodies. Female BALB/c mice (6 to 8 weeks old) were immunized with a combination of GP⌬muc proteins for EBOV (Mayinga), SUDV (Boniface), and MARV (Musoke) (25 ␮g each) on study days 0, 14, and 28 (group 1) or study days 0, 14, 28, 42, and 56 along with 20 ␮g Sigma adjuvant system by the intramuscular (i.m.) route. An intravenous (i.v.) dose of the same antigens without adjuvant was administered on day 35 (group 1) or day 63 (group 2). Mice were euthanized on day 42 (group 1) or 64 (group 2), and their spleens and lymph nodes were harvested. Lymphocytes (1 ⫻ 108), which had been collected from spleens and lymph nodes, were fused with the 2 ⫻ 107 SP2/0-Ag14 myeloma cells (ATCC), using polyethylene glycol (PEG) according to method A in the ClonaCell-HY hybridoma cloning manual (Stemcell Technologies) (25). Fused cells were incubated overnight in medium C to allow the fused cells to go through one cell cycle to express the hypoxanthine-guanine phosphoribosyltransferase (HRPT) enzyme that will allow them to survive in selection medium D. The day after the fusion, cells were plated in methylcellulose-containing medium D in 96-well plates and left undisturbed for 10 to 14 days at 37°C and 5% CO2. The cells on the plates were fed with medium E every few days. Supernatants were screened for antibodies reactive to EBOV, SUDV, and MARV glycoproteins. Cells from wells identified as positive against two or all three glycoproteins in the screening enzyme-linked immunosorbent assay (ELISA), were expanded and recloned by limiting dilutions using medium E. Animal immunization studies and all associated experimental procedures were preapproved and performed according to guidelines set by the Noble Life Sciences (Gaithersburg, MD) Animal Care and Use Committee (protocol 12-05-017-IBT). This animal facility used is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care and adheres to the principles stated in the Guide for the Care and Use of Laboratory Animals (26). Animals were euthanized in accordance with the 2013 American Veterinary Medical Association (AVMA) Guidelines on Euthanasia using carbon dioxide exposure, followed by cervical dislocation. The choice of numbers of mice per group (n ⫽ 3) is based on the minimum number of animals that is required to yield a consistent anti-

Journal of Virology

jvi.asm.org

267

Holtsberg et al.

body response, generation of sufficient B cells for analysis, and generation of sufficient serum volumes for analysis. Production and purification of MAbs. Hybridoma cells were grown in bioreactor flasks (CELLine CL350 [catalog no. 90010; Integra]). Supernatants were collected every 3 to 5 days and combined for purification using an Akta fast protein liquid chromatograph (FPLC). Produced antibodies were captured by protein A (catalog no. 17-0403-01; GE Healthcare) and washed and eluted with 0.1 M glycine buffer (pH 2.4). Fractions, containing the Ab peak were collected, neutralized with Ab buffer (20 mM L-histidine [pH 6.0], 150 mM NaCl, and 4% sucrose) and dialyzed against the same buffer overnight at 40°C. ELISA. Purified glycoproteins were immobilized overnight at 100 ng/ well in PBS at 4°C on 96-well Nunc MaxiSorp plates (Thermo Fisher Scientific) and incubated with hybridoma supernatants or serial dilutions of purified MAbs. Bound antibodies were detected using a horseradish peroxidase (HRP)-conjugated anti-mouse secondary antibody and 3,3=,5,5=-tetramethylbenzidine (TMB) substrate. Absorbance values determined at 650 nm were transformed using Softmax 4 parameter curvefit (Molecular Devices). The half-maximal effective concentration (EC50) value at the inflection point of the curve is reported. Competition ELISAs. His-tagged EBOV GP⌬TM was immobilized at 100 ng/well on 96-well nickel-coated plates (Promega) at 4°C overnight. The next day, the plates were washed and blocked for 1 h, and a set of two MAbs was then added and allowed to bind for 1 h at room temperature. To detect competition between mouse and humanized MAbs, 1 ␮g/ml mouse MAb (4 ␮g/ml for m8C4) was mixed with a 10-fold excess of humanized MAb and detected using goat anti-mouse antibody conjugated to HRP. To detect competition between two mouse MAbs, m8C4 was biotinylated with EZ-Link NHS-biotin according to the manufacturer’s protocol (Life Technologies, Carlsbad, CA) and detected with antistreptavidin antibody conjugated to HRP. KZ52 competition with humanized MAbs also required biotinylation and detection with antistreptavidin antibody conjugated to HRP. After 1 h, TMB substrate was added, and absorbance values were determined at 650 nm on a VersaMax plate reader. Percent competition was calculated from MAb binding in the presence of an irrelevant MAb control. For mapping of linear epitopes, the competition assay was set up as described above and performed in the presence or absence of 100-fold molar excess of a set of overlapping peptides spanning the sequences of filovirus GPs. Neutralization assays with VSV-GPSUDV and VSV-GPEBOV. Neutralization assays were performed using vesicular stomatitis virus pseudotyped to display the GP from either SUDV or EBOV in place of its native G glycoprotein (VSV-GPSUDV or VSV-GPEBOV, respectively). The viral genome encodes an enhanced green fluorescent protein (eGFP). Infection is scored by counting fluorescent cells. The protocol for VSV-GP production has been described elsewhere (27). Neutralization of live EBOV and SUDV. Antibodies were diluted in Vero growth medium (Eagle minimum essential medium with Earle’s salts and L-glutamine, 5% fetal bovine serum [FBS], 1% penicillin-streptomycin) at two times the desired final concentrations, mixed with equal volume of live virus (EBOV or SUDV), and incubated at 37°C for 1 h before adding to Vero cells in 96-well black plates. The cells were incubated with MAb/virus inoculum (at a multiplicity of infection [MOI] of ⬃0.5) for 1 h at 37°C and washed with PBS, and growth medium alone without antibody was added to all wells. The cells were fixed at 48 h postinfection, and infected cells were determined by an indirect immunofluorescence assay (IFA) using virus-specific MAbs (EBOV-specific KZ52, SUDV-specific 17F6) and fluorescently labeled secondary antibodies (goat anti-human IgG conjugated to Alexa Fluor 488; Molecular Probes). The percentages of infected cells were determined using an Operetta and Harmony software. Data are expressed as the percentage of inhibition relative to control cells treated with vehicle for both EBOV and SUDV. Antibody-mediated cellular phagocytosis. (i) Cells. Murine monocytes (RAW cells) and human monocytes (THP-1 cells) were cultured in

268

jvi.asm.org

RPMI 1640 containing 10% FBS. White blood cells were isolated from peripheral blood of BALB/c mice using ACK lysis buffer (Sigma), and neutrophils were identified by flow cytometry as SSC-Ahigh (SSC stands for side scatter) GR-1⫹, CD14⫺, CD3⫺. Monocyte-derived dendritic cells (moDCs) were generated by differentiating CD14⫹ monocytes from peripheral blood and culturing for 6 days in the presence of interleukin 4 (IL-4) or granulocyte-macrophage colony-stimulating factor (GM-CSF) (Miltenyi). (ii) Phagocytic assay. EBOV-GP (IBT Bioservices) was biotinylated and conjugated to streptavidin-coated Alexa Fluor 488 beads (Life Technologies). Antibodies (5 ␮g/ml to 0.1 ␮g/ml) were incubated with beads for 2 h at 37°C. RAW cells, THP-1 cells, freshly isolated mouse neutrophils, or moDCs were added at a concentration of 1.0 ⫻ 105 cells/well and incubated for approximately 16 to 18 h (THP-1 cells and moDCs) or 1 h (RAW cells and neutrophils). Cells were analyzed by flow cytometry, and a phagocytic score was determined using the percentage of fluorescein isothiocyanate (FITC)-positive (FITC⫹) cells and the median fluorescent intensity (MFI) of the FITC⫹ cells. Analysis of glycan content of antibodies. The relative abundance of antibody glycan structures were quantified by capillary electrophoresis as previously described (28). Briefly, antibodies were digested with IDES (FabRICATOR; Genovis) to separate Fc and F(ab=)2, and N-linked glycans were removed using peptide-N-glycosidase F (PNGase F) (New England BioLabs). Free glycans were labeled with 50 mM 9-aminopyrene1,4,6-trisulfonic acid (APTS) in 1.2 M citric acid and 1 M sodium cyanoborohydride in tetrahydrofuran (THF), and excess dye was removed using a size exclusion resin. Labeled glycans were loaded onto a 3130 XL ABI DNA Sequencer using a POP7 polymer in a 36-cm capillary. Peaks of 19 substructures were identified, and the relative abundance of a given structure was determined by calculating the area under the curve of each peak divided by the total area of all peaks. Filovirus challenge studies in mice. (i) Murine EBOV model. The lethal mouse-adapted EBOV mouse model was developed at the U.S. Army Medical Research Institute of Infectious Diseases (USAMRIID), using adult mice by serial passages of EBOV (Zaire) in progressively older suckling mice (29, 30). This model has been thoroughly validated (31). Female BALB/c mice, aged 6 to 8 weeks, were purchased from Charles River Laboratory. Upon arrival, mice were housed in microisolator cages and provided chow and water ad libitum. On day 0, mice were transferred to a biosafety level 4 containment area and challenged intraperitoneally with a target of 1,000 PFU of mouse-adapted EBOV. In two independent experiments, groups of mice (5 or 10 mice per group) were treated intraperitoneally with a range of antibody doses as indicated in the legend to Fig. 6. Control mice were simultaneously challenged but given only PBS at the corresponding treatment intervals. Mice were weighed as groups and monitored daily for 28 days postinfection. When experimental conditions in independent experiments were the same, the data for those groups were pooled for graphic presentation in figures and statistical analysis. (ii) Murine SUDV model. The SUDV mouse model was developed at the USAMRIID utilizing the IFN-␣/␤R⫺/⫺ mouse model (mice lacking receptors for alpha interferon [IFN-␣] and IFN-␤) (32). IFN-␣/␤R⫺/⫺ mice (B6.129S2-Ifnar1tm1Agt/Mmjax), aged 8 to 10 weeks, on the C57BL/6 background were purchased from Jackson Laboratories (Bar Harbor, ME) and used for all SUDV challenge experiments. Upon arrival, mice were housed in microisolator cages and provided chow and water ad libitum. On day 0, mice were transferred to a biosafety level 4 containment area and challenged intraperitoneally with a target of 1,000 PFU SUDVBoniface virus. Groups of mice (10 mice per group [5 male, 5 female]) were treated intraperitoneally with a range of doses of each antibody as indicated. Control mice were simultaneously challenged but given only PBS at the corresponding treatment intervals. Mice were weighed as groups and monitored daily for 28 days postinfection. Animal research was conducted under a protocol approved by the U.S. Army Medical Research Institute of Infectious Diseases Institutional Animal Care and Use Committee (IACUC) in compliance with the Animal

Journal of Virology

January 2016 Volume 90 Number 1

Cross-Reactive Mouse-Derived Filovirus Antibodies

FIG 1 Immunization design for generation of pan-filovirus antibodies and serology results. (A) Two groups (three mice in each group) were immunized intramuscularly (IM) with the indicated antigens on the indicated days. Animals received an intravenous (IV) boost 3 days (group 1) or 1 day (group 2) before euthanasia and harvest of splenocytes, followed by fusion and hybridoma development. The number of total and positive clones are indicated. S, reactive to SUDV; E, reactive to EBOV; M, reactive to MARV. (B) Sera from terminal bleeds of each group were pooled and tested for antibody response to GP⌬muc of EBOV, SUDV, and MARV using ELISA. Data show binding at 1:100 dilution of serum. OD650, optical density at 650 nm. (C) Percent neutralization of VSV-GP pseudotyped viruses for EBOV, SUDV, and MARV by 1:10 dilution of terminal sera from group 1.

Welfare Act and other federal statutes and regulations relating to animals and experiments involving animals. The USAMRIID facility is fully accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care International and adheres to the principles stated in the Guide for the Care and Use of Laboratory Animals (26). Challenge studies were conducted under maximum containment in an animal biosafety level 4 facility. Animal studies were performed in a blind manner (personnel assessing results were unaware of which animals received which treatment). Animals were not specifically allocated or randomized into groups. The number of mice to be used in these studies was selected to measure and determine differences in levels of protection elicited by the different MAb treatments. Experience with the use of various analyses for determining the probability of differences between control and experimental mouse groups indicates the need for 5 to 10 mice per group. For lethality studies, a statistical difference between 0% in the control and ⱖ30% in the treated groups can be demonstrated using 10 mice per group with ⬎90% power using a two-sided alpha level of 0.05. In the cases where large numbers of antibodies were tested in preliminary studies, five mice per group were tested, and then the results were confirmed with the larger group of 5 or 10 mice. In these cases, the results of the combination of both studies are shown. To demonstrate statistical differences in the treatment groups, data were analyzed using GraphPad software, version 6 (La Jolla, CA) using the Mantel-Cox log rank test and confirmed using the GehanBreslow-Wilcoxon test.

RESULTS

Antibody generation. To drive the generation of broadly reactive antibodies, BALB/c mice were immunized with a mixture of pu-

January 2016 Volume 90 Number 1

rified, engineered glycoprotein (GP) ectodomains lacking the highly divergent mucin-like domain (MLD) (GP⌬muc) for EBOV (Mayinga), SUDV (Boniface), and MARV (Musoke) using two different immunization strategies (Fig. 1A). Mice were also boosted intravenously with the same protein cocktail (without adjuvant) 1 to 3 days before harvesting spleens for fusion. Total IgG and neutralizing responses against the three filovirus GPs were determined in terminal bleeds (Fig. 1B and C). Splenocytes were harvested and fused to generate hybridomas. More than 1,300 clones were initially screened for binding to the GP ectodomain (GP⌬TM) for EBOV, SUDV, and MARV. As expected, the majority of the reactive clones were specific for a single GP species and were excluded from further analysis. A single MAb (m21D10) reactive to all three GP molecules and six clones reactive to both EBOV and SUDV but nonreactive to MARV were identified and grown for further analysis (Fig. 1A). Upon secondary screening, four cross-reactive but ebolavirusspecific clones (m16G8, m8C4, m17C6, and m4B8 [the “m” prefix indicates the murine origin of the antibody]) exhibited greater strength and breadth of binding. These four clones along with m21D10 were further subcloned to ensure clonal purity, and the MAbs were produced, purified, and analyzed for binding to GP⌬TM from EBOV, SUDV, BDBV, RESTV, and MARV. The MAbs displayed various degrees of binding to most species of ebolavirus, but none, other than m21D10, bound to MARV, con-

Journal of Virology

jvi.asm.org

269

Holtsberg et al.

FIG 2 Binding profile of pan-ebolavirus MAbs. ELISA plates were coated with GP⌬TM from the indicated filovirus species. The binding of the MAbs was tested over a wide range of concentrations as described in Materials and Methods. Data are shown as the optical density at 650 nm (OD650) for m16G8 (A), m8C4 (B), m17C6 (C), m4B8 (D), and m21D10 (E) graphed against the concentration of MAbs (in micrograms/milliliter).

sistent with the primary screening results (Fig. 2). The monoclonal antibody m16G8 bound moderately to BDBV but with low affinity to EBOV, SUDV, and RESTV glycoproteins (Fig. 2A). m8C4 showed preferential binding to SUDV and EBOV and to a lesser extent to BDBV but failed to bind to RESTV (Fig. 2B). m17C6 displayed strong binding to EBOV but exhibited lower levels of binding to SUDV and RESTV and very poor binding to BDBV (Fig. 2C). The strongest and most balanced binding was displayed by m4B8, which bound to the four ebolavirus species at low nanogram/milliliter concentrations (Fig. 2D). The only MARV-reactive clone, m21D10, exhibited the highest binding for BDBV and RESTV, followed by MARV and EBOV, and displayed a low level of binding to SUDV (Fig. 2E), thus representing a pan-filovirus antibody. Subtype-specific ELISA showed that m16G8 is an IgG2a, while the other MAbs are IgG1 (data not shown). Characterization of the antibodies. None of the four pan-

270

jvi.asm.org

ebolavirus MAbs detected GP under denaturing conditions (data not shown), suggesting that they react with discontinuous epitopes. To determine the general binding regions of the conformational antibodies, we compared the binding of the MAbs to several EBOV GP constructs containing various functional domains of the protein. EBOV GP consists of the disulfide-bound GP1 and GP2 subunits that are responsible for receptor binding and membrane fusion, respectively (Fig. 3A). GP1 forms a chalice consisting of the receptor binding region (RBR) and the glycan cap positioned at the rim of the chalice as well as a C-terminal, highly glycosylated and disordered mucin-like domain (Fig. 3A) (20). GP2 wraps around GP1, and along with the N terminus of GP1, forms the base of the chalice. The unedited EBOV GP gene encodes soluble GP (sGP), which consists of amino acids 31 to 295, followed by a unique C-terminal tail (33, 34). sGP includes the glycan cap but lacks the MLD and GP2. During viral entry, GP undergoes cleavage by cathepsins at the N terminus of the glycan

Journal of Virology

January 2016 Volume 90 Number 1

Cross-Reactive Mouse-Derived Filovirus Antibodies

FIG 3 Binding pattern of the pan-ebolavirus conformational MAbs. (A) Schematic of the domain organization of EBOV GP and modeled structures of the various truncated proteins used for domain mapping (based on PDB accession no. 3CSY). The colors in the structures shown match the linear diagram. The disordered MLD is shown as gray spheres. SP, signal peptide; TM, transmembrane. (B to D) Dose-dependent binding of each MAb to GP⌬TM (B), GP⌬muc (C), GPcl (D), and sGP (E) is shown as determined by ELISA. The EC50s for each MAb are shown in each panel. NR, nonreactive.

cap to generate cleaved GP (GPcl), representing truncated GP1 lacking both the glycan cap and MLD, which remains associated with GP2. We used GP⌬TM, GP⌬muc, GPcl, and sGP to systematically explore the binding region of the four pan-ebolavirus MAbs. As shown in Fig. 3B to E, m4B8 and m17C6 bound to all four GP forms, indicating that the respective conformational epitopes are in GP1 between amino acids 31 to 200. In contrast, m16G8 failed to bind to sGP (Fig. 3E), and since its binding was not dependent on MLD (Fig. 3B and C), it is likely to have contact points in GP2. Its relatively poor binding to GPcl suggests that its epitope could be affected by enzymatic cleavage. The m8C4 MAb bound to all GP forms except for GPcl (Fig. 3D), suggesting that its epitope lies within the glycan cap or the cathepsin cleavage site itself. In contrast to the pan-ebolavirus MAbs, m21D10 reacted with chemically denatured antigen (data not shown) and recognized GP from EBOV, SUDV, BDBV, RESTV, and MARV in Western blots (Fig. 4A), indicating that it binds to a continuous epitope. To define the linear epitope of m21D10, we used a library of 15residue overlapping peptides spanning the entire GP from EBOV, SUDV, and MARV to search for peptides that specifically compete with the binding of m21D10 to GP. This competition assay (Fig. 4B and C) revealed that m21D10 binds to residues 81 to 90 of EBOV and SUDV, as well as positions 65 to 74 of MARV GP. This stretch of amino acids is located within the RBR, corresponds to ␣1␤4 of EBOV GP (20), and is highly conserved across filoviruses (Fig. 4C and D). This epitope appears to be heavily concealed on the surface of GP, as removal of the MLD significantly enhanced

January 2016 Volume 90 Number 1

the binding of m21D10 to EBOV, SUDV, and MARV GP (Fig. 4E and F). The binding of m21D10 to GPcl was exponentially higher than its binding to GP⌬muc (EC50 reduced ⬎1 log unit), indicating that the glycan cap also restricts access to this site (Fig. 4F). Using competition ELISA, the pan-ebolavirus MAbs were tested for possible epitope overlaps among each other or previously described anti-EBOV MAbs KZ52 (21) and 13C6 (35, 36) (Table 1). KZ52 binds at the base of the GP chalice (20) to an epitope shared by ZMapp components 2G4 and 4G7 (19), while 13C6, another component of ZMapp (14), is a glycan cap binder (19). m16G8 did not show significant competition with any of the four pan-ebolavirus MAbs, KZ52, or 13C6, suggesting that it recognizes a unique epitope with little steric interference by the other MAbs tested. The m8C4 MAb that likely binds the glycan cap or the cathepsin site showed no competition with the other three pan-ebolavirus MAbs but partially competed with 13C6 for binding to EBOV GP. However, 13C6 did not compete with m8C4 for binding to SUDV GP, suggesting that m8C4 and 13C6 share contact sites on the EBOV glycan cap, but not on SUDV, and therefore, the epitopes for these two MAbs are related but not identical. 13C6 did not compete with any other antibody tested. Partial competition was observed between m17C6 and h4B8, suggesting overlapping conformational epitopes. KZ52 also partially displaced m8C4, but m8C4 was unable to displace KZ52. Partial displacement of m8C4 with both KZ52 and 13C6 clearly indicates that m8C4 binds to a novel site. KZ52 did not show competition with m16G8, m17C6, or m4B8. Virus neutralization activity. The MAbs were tested for neu-

Journal of Virology

jvi.asm.org

271

Holtsberg et al.

FIG 4 Reactivity in Western blot and epitope mapping of m21D10. (A) Reactivity of m21D10 to GP⌬TM from SUDV, EBOV, MARV, BDBV, and RESTV by Western blotting. The positions of molecular markers (in kilodaltons) are shown to the left of the blot. (B) m21D10 was preincubated with or without a 100-fold molar excess of specific SUDV peptide (S7), EBOV peptides (Z16 and Z17), MARV peptides (M16 and M17), or a control (Cntrl) irrelevant peptide before being added to ELISA plate wells coated with GP⌬muc from SUDV, EBOV, or MARV, and the specific binding was detected using HRP-conjugated anti-mouse antibodies. Prior to this experiment using the same approach, the approximate regions of binding were estimated using pools of five overlapping peptides. (C) Sequence alignment of the competing peptides from EBOV, SUDV, and MARV, as well as corresponding sequences from BDBV, RESTV, and TAFV. The box indicates the core binding epitope of 21D10. (D) Partial view of EBOV GP crystal structure (PDB accession no. C3SY) with the m21D10 contact residues, corresponding to ␣1 and ␤4, highlighted in yellow in one of the GP monomers. The glycan cap for this monomer is shown in blue, and the rest of GP1 is shown in pale green. (E) Binding of m21D10 to GP⌬TM and GP⌬muc of SUDV and MARV. (F) Dose-dependent binding of m21D10 to GP⌬TM, GP⌬muc, and GPcl determined by ELISA.

tralization of EBOV and SUDV using GFP-expressing VSV pseudotyped with EBOV or SUDV GP (27). Out of the MAbs tested, only m8C4 showed appreciable neutralizing activity in this system. As shown in Fig. 5A, m8C4 effectively inhibited both EBOV GP-VSV and SUDV GP-VSV with 50% inhibitory concentration (IC50) values of about 1.5 and 0.75 ␮g/ml, respectively. The neutralizing activity of the MAbs was then tested on live virus using a high-content imaging-based assay. Consistent with the VSV pseu-

272

jvi.asm.org

dotype assay, only m8C4 effectively neutralized both SUDV (Fig. 5B) and EBOV (Fig. 5C). The neutralization potential of m21D10 was also tested against EBOV, SUDV, and MARV using the VSV pseudotype system, but this antibody exhibited no neutralizing activity at up to 100 ␮g/ml (data not shown). Efficacy in the mouse model of EBOV infection. The efficacy of the MAbs was tested in BALB/c mice using mouse-adapted EBOV (MA-EBOV) (29). Groups of mice were challenged with a

Journal of Virology

January 2016 Volume 90 Number 1

Cross-Reactive Mouse-Derived Filovirus Antibodies

TABLE 1 Binding competition between pan-ebolavirus MAbs, KZ52, and 13C6a % binding competition between pan-ebolavirus MAbs Antigen

Detecting antibody

EBOV GP

m8C4-biotin m16G8 m17C6 m4B8 KZ52-biotin c13C6

SUDV GP

m8C4-biotin

m8C4 1 11 8 5 59

m16G8

h17C6

h4B8

KZ52

c13C6

1

11 10

8 13 66

58 3 6 2

59 11 14 6 3

10 13 6 11

66 1 14

6 6

3

6

5

3

8

0

a

Competition ELISA was performed as described in Materials and Methods. To enable detection of the antibodies in the presence of each other, some antibodies were biotinylated (m8C4 and KZ52), while some were used as humanized (h4B8 and h17C6 [h stands for humanized]) or mouse-human chimeric forms (c13C6 [c stands for chimeric]).

target dose of 1,000 PFU of MA-EBOV, followed by two intraperitoneal injections of antibody (25 mg/kg of body weight) at 2 h and 3 days postinfection. As shown in Fig. 6A, using this regimen, all mice treated with m4B8 and 7 out of 15 mice receiving m8C4 survived the lethal challenge, while m17C6 and m16G8 provided only 20% and 13% protection, respectively. In contrast, all 15 control mice as well as mice treated with m21D10 succumbed to infection within 5 to 9 days. Control mice and animals treated with m21D10 lost more than 25% weight before dying, while all groups of mice treated with pan-ebolavirus MAbs showed considerably less weight loss (Fig. 6B). Given the strong protection afforded by two injections of m4B8, we examined whether delayed treatment with this antibody would provide protection. When treatment with m4B8 was delayed until 3 days after infection, 80% of the mice survived the challenge (Fig. 6C) with less than 18% average weight loss (Fig. 6D). The success of the EBOV-specific cocktails ZMab (18), MB003 (11), and ZMapp (14) has suggested that antibodies targeting multiple distinct epitopes may act synergistically and have greater therapeutic efficacy. To this end, we examined whether the efficacy of m8C4 (binding near or on the glycan cap) can be enhanced by m16G8 (an antibody likely to bind GP2). When mice were treated with a cocktail of the two antibodies at 15 mg/kg, 80% of the animals survived the challenge (Fig. 6E). In contrast, m8C4 and m16G8 provided only 47% and 13% survival, respectively, when administered individually at a higher dose of 25 mg/kg (Fig. 6A). Delayed treatment (single dose of 15 mg/kg on day 3) with the combination of m8C4 and m16G8 also led to 30% survival (Fig. 6E). Mice treated twice with m8C4 plus m16G8 lost less than

10% weight, while the average weight loss in controls was more than 25% (Fig. 6F). Efficacy in a SUDV infection model. A lethal mouse model for SUDV has been recently developed using mice lacking receptors for IFN-␣ and IFN-␤ (IFN-␣␤R⫺/⫺) (32). Since m8C4 showed strong neutralization toward SUDV, as a proof of concept, we sought to test this antibody in the SUDV mouse model. Groups of 10 mice were infected with 1,000 PFU of wild-type SUDV. One group received m8C4 (10 mg/kg) 24 h before and 24 h and 72 h after infection, while a second group received m8C4 (5 mg/kg) only on day 1 postinfection. While all control mice and mice receiving a single 5-mg/kg dose of m8C4 succumbed to infection, treatment with three doses of m8C4 led to 70% protection and lower and delayed weight loss (Fig. 7A and B). Thus, m8C4 represents the prototype of an antibody with cross neutralization and protective efficacy against two widely divergent ebolavirus species. Effector functions mediated by pan-ebolavirus antibodies. Antibody-mediated activation of innate immune cell effector functions, including phagocytosis and cytotoxic activity mediated by neutrophils, monocytes, and NK cells, can play a critical role in protection from viral and bacterial infections (37). Thus, to explore the role of nonneutralizing functions in antibody-mediated protection, antibodies were evaluated for their ability to rapidly recruit phagocytic clearance of EBOV-specific immune complexes by distinct innate immune cells. Using the murine RAW phagocytic cell line, potent phagocytosis was observed for all MAbs with the exception of m16G8 (Fig. 8A). Notably, only m4B8, and to a lesser extent m8C4, induced appreciable phagocytosis using the human monocytic THP-1 cells (Fig. 8B). Similar to

FIG 5 Neutralizing activity of the pan-ebolavirus MAbs against EBOV and SUDV. (A) Percent neutralization by m8C4 of vesicular stomatitis virus (VSV) pseudotyped with EBOV or SUDV GP. (B and C) Neutralizing activity of the pan-ebolavirus antibodies against wild-type SUDV (B) and EBOV (C) determined using a high-content imaging-based neutralization assay as described in Materials and Methods.

January 2016 Volume 90 Number 1

Journal of Virology

jvi.asm.org

273

Holtsberg et al.

FIG 6 Protection of mice from lethal challenge by pan-ebolavirus antibodies. (A and B) Survival (A) and weight loss (B) (group averages) in groups of mice that were infected with 1,000 PFU of MA-EBOV and treated intraperitoneally with 25 mg/kg of the indicated MAb 2 h or 3 days postinfection. Statistical differences were assessed for each treatment group compared to the negative-control group using the Mantel-Cox log rank test (P values of ⬍0.0001 for 4B8, 0.7424 for 17C6, 0.5458 for 16G8, 0.0005 for 8C4, and 0.8579 for 21D10). (C and D) Efficacy of m4B8 (30 mg/kg) treatment on day 3 (P ⫽ 0.0024 compared to the negative-control value). (E and F) Efficacy of combination of m8C4 and m16G8 (15 mg/kg) treatment on days 0 and 3 (P values of 0.0004 compared to the negative-control value and 0.0226 compared to the value for day 3 only) or day 3 only (P value of 0.2460 compared to the negative-control value). The number of mice per group and doses are shown in each panel.

the RAW cells, m4B8, m17C6, and m8C4 induced robust neutrophil-mediated phagocytosis; however, m16G8 mediated negligible activity (Fig. 8C). Finally, m4B8 and m8C4 induced higher phagocytosis by human monocyte-derived dendritic cells (moDCs) than m17C6 and m16G8 did (Fig. 8D). Interestingly, phagocytic activity was associated with lower galactosylation, reduced fucosylation, and enhanced addition of bisecting N-acetylglucosamine (GlcNAc) glycan structures (Fig. 8E), which have previously been associated with Fc effector functionality (38, 39). These data demonstrate that different Fc effector functional profiles exist among the MAbs, including highly phagocytic antibodies (m4B8 and m8C4) that induce robust phagocytosis by all innate immune cells tested, antibodies that induce selective

phagocytosis in particular cell types (m17C6), and antibodies that do not induce phagocytosis at all (m16G8), and suggest that unique Fc effector profiles may be important for understanding the underlying mechanisms of antibody-mediated protection from EBOV infection. DISCUSSION

The 2014-2015 outbreak of EVD in West Africa has highlighted the threat of filoviruses to global health (40). While significant progress has been made toward vaccines and immunotherapeutics for EBOV, the development of therapeutics against other filovirus species has been lagging behind. Broadly protective treat-

FIG 7 Efficacy of m8C4 against Sudan virus infection. Groups of 10 IFN-␣␤⫺/⫺ mice were infected with wild-type SUDV. Two groups of mice received m8C4 either once on day 1 at 5 mg/kg or on days ⫺1, ⫹1, and ⫹3 relative to infection at a dose of 10 mg/kg. Control mice received PBS only. (A and B) Survival (A) and weight change (B) were monitored and recorded. Statistical differences were assessed for each treatment group compared to the negative-control group (P ⫽ 0.3137 for day 1 only or P ⫽ 0.0001 for days ⫺1, ⫹1, and ⫹3) or when the treatments were compared to each other (P ⫽ 0.0001) using the Mantel-Cox log rank test.

274

jvi.asm.org

Journal of Virology

January 2016 Volume 90 Number 1

Cross-Reactive Mouse-Derived Filovirus Antibodies

FIG 8 Antibody Fc-mediated innate immune effector function. (A to D) Antibodies were assayed for the ability to induce phagocytosis of EBOV GP-coated beads by RAW cells (A), THP-1 cells (B), murine neutrophils (C), or moDCs (D). The concentration of antibody and the phagocytic score are shown. The dotted lines show the values for the no-antibody control. (E) The Fc glycan composition of each antibody was determined by capillary electrophoresis. The relative abundance of bisecting GlcNAc, and afucosylated or galactosylated glycan structures are shown for each antibody.

ment modalities are urgently needed, as the nature of future outbreaks cannot be predicted. Here we report novel murine monoclonal antibodies crossreactive to four species of ebolavirus and a pan-filovirus antibody. The only ebolavirus species that was not tested in this study is TAFV, which has caused disease in only one nonlethal human case (41). However, the TAFV GP shows 90% sequence identity with BDBV in regions excluding the MLD, and thus, these antibodies are very likely to bind to TAFV. We applied an immunization strategy in mice using a cocktail of engineered glycoproteins from three filovirus species lacking the highly divergent MLD to drive immune responses to conserved epitopes. More than 1,300 clones had to be screened to select four MAbs with broad ebolavirus reactivity and a single pan-filovirus MAb, indicating that such antibodies are rare. Indeed, most of the antibodies identified by Wilson et al. in mice vaccinated with an EBOV GP antigen were EBOV specific and recognize the MLD (36). Epitopes conserved across filovirus species may not be immunodominant, and the B cell response is likely

January 2016 Volume 90 Number 1

dominated by species-restricted epitopes during natural infection, which always involves a single species. This is reminiscent of antibodies to subdominant, broadly neutralizing, epitopes within the stem region of influenza virus hemagglutinin that are rarely induced during seasonal infections and that require special immunogen design or prime-boost regimens to be effectively elicited (42–45). This is the first report of a single antibody (m8C4) to exhibit neutralization and protective efficacy against two different filovirus species. This finding is particularly significant because EBOV and SUDV are the most divergent ebolaviruses (56% amino acid sequence identity within GP). Currently, the only other available SUDV MAb immunotherapeutic candidates include 16F6 and its synthetic, human analogs E10 and F4, which confer protection and memory immunity in the SUDV murine model (22, 46). However, these SUDV MAbs are strictly species specific and bind at a distinct epitope at the base of prefusion GP (22), and therefore, m8C4 reveals a novel protective epitope for SUDV GP. Notably, protection in NHPs from lethal SUDV challenge by any

Journal of Virology

jvi.asm.org

275

Holtsberg et al.

MAb therapy has yet to be demonstrated. Nonetheless, our results suggest that effective pan-ebolavirus therapeutic antibodies can be developed. Another MAb described here, m4B8, binds with high affinity to EBOV, SUDV, BDBV, and RESTV and provides 100% efficacy after two injections in a mouse model of EBOV infection. Single delivery of this MAb on day 3 (peak of viremia in mice [29]) also provided significant protection. While m16G8 and m17C6 provided low levels of protection in mice, consistent with the current paradigm emerging from studies with ZMapp (14), a cocktail of m8C4 and m16G8 displayed synergy when combined at lower doses. Previous results indicated that the most effective EBOV antibodies target the glycan cap (13C6) and the base of the trimeric GP chalice (4G7, 2G4, and KZ52) (14, 19). Since the KZ52 epitope appears to be strictly species restricted, it is significant that none of the pan-ebolavirus MAbs reported here bind the KZ52-type base epitope, suggesting that there are additional epitopes that can be exploited for development of broadly protective antibodies. On the other hand, there may be additional epitopes on or surrounding the glycan cap that can be exploited. For example, m8C4 binds to a conformational epitope that appears to involve the glycan cap and possibly neighboring residues within the core GP1 and cathepsin cleavage site. Such an epitope would be positioned between 13C6 and KZ52 binding sites, consistent with partial competition of m8C4 by 13C6 and KZ52. It is important to note that in contrast to 13C6, m8C4 neutralized both EBOV and SUDV in the absence of complement, further indicating that m8C4 recognizes a novel neutralizing epitope. Gross epitope mapping suggested that m16G8 may bind to GP2, since this MAb bound only to forms of GP that retained GP2 but failed to bind to sGP. Additionally, m16G8 binding to GPcl was severely reduced (Fig. 3), despite retention of GP2 in GPcl. This suggested that m16G8 may be making additional contacts with specific residues that are lost after proteolysis within the cathepsin cleavage loop (residues 190 to 213), a disordered loop that traverses over GP2 in the trimeric structure connecting the base to the glycan cap (20). The binding of m16G8 was not displaced by excess KZ52, suggesting that m16G8 engages a distinct epitope. The observed enhancement of efficacy between m8C4 and m16G8 (Fig. 6) is particularly significant, given that both these MAbs bind GP with low to moderate affinity. Therefore, further affinity maturation of these two MAbs may lead to a very potent pan-ebolavirus cocktail. This observation further supports the notion that effective treatment can be achieved by targeting multiple epitopes within these critical domains. The most protective cross-reactive MAb, m4B8, binds to a conformational epitope within the first ⬃170 amino acids of mature GP1. While the details of this epitope remain to be defined by crystallographic analysis, it appears to be a novel epitope, as m4B8 did not compete with any known antibodies tested in our study. Interestingly, m4B8 did not show neutralizing activity. It is possible that in vitro neutralization assays may not be sensitive enough to detect modest neutralizing activities, as in plaque reduction assays, we did observe low levels of neutralization by m4B8 (data not shown). Alternatively, the primary mechanism of action of m4B8 may involve Fc-dependent effector mechanisms. Indeed, our studies show that m4B8 was highly effective in mediating phagocytosis by neutrophils, monocytes, and dendritic cells. The RBR shows high degree of homology among filoviruses.

276

jvi.asm.org

FIG 9 Crystal structure of MARV GP. The crystal structure of MARV GP (PDB accession no. 3X2D) is shown with residues contacted by MR78 highlighted in magenta, residues contacted by both MR78 and m21D10 in orange, and amino acids contained only in the peptides reacting with m21D10 in yellow. GP1 is shown in dark gray, and GP2 is shown in light gray. The GP2 of the adjacent molecule is shown in cyan with F535 within the fusion loop highlighted in green. F535 packs into a cavity surrounded by F72 and R73. The R73 side chain undergoes polar interaction with the F535 backbone (yellow dots).

Contrary to our expectation, only a single MAb (m21D10) emerged from our screen that bound to RBR. This antibody bound poorly to EBOV GP ectodomain or virus-like particles (data not shown), but the binding was enhanced exponentially upon proteolysis at the cathepsin cleavage site, suggesting that access is restricted by MLD and glycan cap. Recently, Flyak et al. isolated several related MAbs from a human survivor of MARV infection (47) with various degrees of reactivity to EBOV GP. Cocrystal structure of one of these MAbs (MR78) showed that it binds to the putative receptor binding site of MARV GP (24). The linear epitope for m21D10 is distinct from the MR78 epitope, but it does share several contact residues (Fig. 9). This epitope includes MARV GP F72 (equivalent to EBOV F89, which appears to be part of the receptor binding site [20, 24]) and R73 (equivalent to EBOV R90), which makes contacts with the fusion loop of the adjacent GP molecule (Fig. 9) in both MARV (24) and EBOV (20) GP structures. Unlike MR78-type antibodies, m21D10 did not neutralize MARV, possibly due to differences in sequence or accessibility of the site. It is possible that an affinity matured version of m21D10 may overcome the restricted access to the RBR. In addition to neutralization, antibodies can provide protection following infection through the recruitment of Fc-dependent innate immune effector functions. Interestingly, three of the four pan-ebolavirus antibodies induced Fc-dependent phagocytosis in multiple innate immune cells, linked to a glycan profile marked by lower galactosylation, lower fucosylation, and increased bisecting GlcNAc addition. Fc effector functionality is also modulated by subclass, with IgG2a and IgG2b being the more functional subclasses of mouse antibodies, based on Fc␥ receptor (Fc␥R) interactions (48). Interestingly, despite m16G8 being IgG2a, whereas the other three antibodies are IgG1, m16G8 still exhibited reduced function compared to the other antibodies. Fc-mediated functionality and neutralization are not mutually exclusive activities, and recent data using HIV-specific neutralizing antibodies suggest that many neutralizing antibodies depend on Fc effector functions to mediate their in vivo protective activity

Journal of Virology

January 2016 Volume 90 Number 1

Cross-Reactive Mouse-Derived Filovirus Antibodies

(49), highlighting the importance of recruiting Fc-mediated innate immune functions, even in the presence of potent neutralization. Thus, while it is likely that while direct virus neutralization by m8C4 significantly contributes to the protection observed in vivo, this antibody may also utilize Fc effector functions to mediate protective immunity. In summary, we have discovered novel pan-ebolavirus conformational epitopes that can be targeted to generate effective antibody cocktails against multiple ebolaviruses. Both novel and cross-reactive antibodies are needed, as current cocktails are reactive only against EBOV, and two of the three components of ZMapp overlap with each other at the base. These results provide a basis for further testing of synergistic pan-ebolavirus cocktails in NHPs. Furthermore, these broadly reactive antibodies represent invaluable tools for development of diagnostics for filoviruses. ACKNOWLEDGMENTS This work was supported by a contract (HDTRA1-13-C-0015) from the U.S. Defense Threat Reduction Agency (DTRA) to M.J.A., as well as R01 AI067927 and U19 AI109762 from National Institutes of Health (NIH) to E.O.S. M.J.A. further acknowledges funding from NIH (R01 AI067927). J.R.L. also acknowledges funding from the NIH (U19AI109762). Research was conducted under an IACUC-approved protocol in compliance with the Animal Welfare Act, PHS policy, and other federal statutes and regulations relating to animals and experiments involving animals. The facility where this research was conducted is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International and adheres to the principles stated in the 2011 Guide for the Care and Use of Laboratory Animals. Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the U.S. Army.

FUNDING INFORMATION HHS | NIH | National Institute of Allergy and Infectious Diseases (NIAID) provided funding to Erica Ollmann Saphire under grant numbers R01 AI067927 and U19 AI109762. DOD | Defense Threat Reduction Agency (DTRA) provided funding to M. Javad Aman under grant number HDTRA1-13-C-0015.

REFERENCES 1. Kuhn JH, Andersen KG, Baize S, Bao Y, Bavari S, Berthet N, Blinkova O, Brister JR, Clawson AN, Fair J, Gabriel M, Garry RF, Gire SK, Goba A, Gonzalez JP, Gunther S, Happi CT, Jahrling PB, Kapetshi J, Kobinger G, Kugelman JR, Leroy EM, Maganga GD, Mbala PK, Moses LM, Muyembe-Tamfum JJ, N=Faly M, Nichol ST, Omilabu SA, Palacios G, Park DJ, Paweska JT, Radoshitzky SR, Rossi CA, Sabeti PC, Schieffelin JS, Schoepp RJ, Sealfon R, Swanepoel R, Towner JS, Wada J, Wauquier N, Yozwiak NL, Formenty P. 2014. Nomenclature- and databasecompatible names for the two Ebola virus variants that emerged in Guinea and the Democratic Republic of the Congo in 2014. Viruses 6:4760 – 4799. http://dx.doi.org/10.3390/v6114760. 2. Kuhn JH, Andersen KG, Bao Y, Bavari S, Becker S, Bennett RS, Bergman NH, Blinkova O, Bradfute S, Brister JR, Bukreyev A, Chandran K, Chepurnov AA, Davey RA, Dietzgen RG, Doggett NA, Dolnik O, Dye JM, Enterlein S, Fenimore PW, Formenty P, Freiberg AN, Garry RF, Garza NL, Gire SK, Gonzalez JP, Griffiths A, Happi CT, Hensley LE, Herbert AS, Hevey MC, Hoenen T, Honko AN, Ignatyev GM, Jahrling PB, Johnson JC, Johnson KM, Kindrachuk J, Klenk HD, Kobinger G, Kochel TJ, Lackemeyer MG, Lackner DF, Leroy EM, Lever MS, Muhlberger E, Netesov SV, Olinger GG, Omilabu SA, Palacios G, Panchal RG, et al. 2014. Filovirus RefSeq entries: evaluation and selection of filovirus type variants, type sequences, and names. Viruses 6:3663– 3682. http://dx.doi.org/10.3390/v6093663. 3. Feldmann H, Kiley MP. 1999. Classification, structure, and replication of filoviruses. Curr Top Microbiol Immunol 235:1–21.

January 2016 Volume 90 Number 1

4. Feldmann H, Klenk HD. 1996. Marburg and Ebola viruses. Adv Virus Res 47:1–52. http://dx.doi.org/10.1016/S0065-3527(08)60733-2. 5. Gire SK, Goba A, Andersen KG, Sealfon RS, Park DJ, Kanneh L, Jalloh S, Momoh M, Fullah M, Dudas G, Wohl S, Moses LM, Yozwiak NL, Winnicki S, Matranga CB, Malboeuf CM, Qu J, Gladden AD, Schaffner SF, Yang X, Jiang PP, Nekoui M, Colubri A, Coomber MR, Fonnie M, Moigboi A, Gbakie M, Kamara FK, Tucker V, Konuwa E, Saffa S, Sellu J, Jalloh AA, Kovoma A, Koninga J, Mustapha I, Kargbo K, Foday M, Yillah M, Kanneh F, Robert W, Massally JL, Chapman SB, Bochicchio J, Murphy C, Nusbaum C, Young S, Birren BW, Grant DS, Scheiffelin JS, Lander ES, Happi C, Gevao SM, Gnirke A, Rambaut A, Garry RF, Khan SH, Sabeti PC. 2014. Genomic surveillance elucidates Ebola virus origin and transmission during the 2014 outbreak. Science 345:1369 – 1372. http://dx.doi.org/10.1126/science.1259657. 6. Chippaux JP. 2014. Outbreaks of Ebola virus disease in Africa: the beginnings of a tragic saga. J Venom Anim Toxins Incl Trop Dis 20:44. http://dx .doi.org/10.1186/1678-9199-20-44. 7. Pigott DM, Golding N, Mylne A, Huang Z, Weiss DJ, Brady OJ, Kraemer MU, Hay SI. 2015. Mapping the zoonotic niche of Marburg virus disease in Africa. Trans R Soc Trop Med Hyg 109:366 –378. http://dx .doi.org/10.1093/trstmh/trv024. 8. Barrette RW, Metwally SA, Rowland JM, Xu L, Zaki SR, Nichol ST, Rollin PE, Towner JS, Shieh WJ, Batten B, Sealy TK, Carrillo C, Moran KE, Bracht AJ, Mayr GA, Sirios-Cruz M, Catbagan DP, Lautner EA, Ksiazek TG, White WR, McIntosh MT. 2009. Discovery of swine as a host for the Reston ebolavirus. Science 325:204 –206. http://dx.doi.org/10 .1126/science.1172705. 9. Dye JM, Herbert AS, Kuehne AI, Barth JF, Muhammad MA, Zak SE, Ortiz RA, Prugar LI, Pratt WD. 2012. Postexposure antibody prophylaxis protects nonhuman primates from filovirus disease. Proc Natl Acad Sci U S A 109:5034 –5039. http://dx.doi.org/10.1073/pnas.1200409109. 10. Marzi A, Yoshida R, Miyamoto H, Ishijima M, Suzuki Y, Higuchi M, Matsuyama Y, Igarashi M, Nakayama E, Kuroda M, Saijo M, Feldmann F, Brining D, Feldmann H, Takada A. 2012. Protective efficacy of neutralizing monoclonal antibodies in a nonhuman primate model of Ebola hemorrhagic fever. PLoS One 7:e36192. http://dx.doi.org/10.1371/journal .pone.0036192. 11. Pettitt J, Zeitlin L, Kim DH, Working C, Johnson JC, Bohorov O, Bratcher B, Hiatt E, Hume SD, Johnson AK, Morton J, Pauly MH, Whaley KJ, Ingram MF, Zovanyi A, Heinrich M, Piper A, Zelko J, Olinger GG. 2013. Therapeutic intervention of Ebola virus infection in rhesus macaques with the MB-003 monoclonal antibody cocktail. Sci Transl Med 5:199ra113. http://dx.doi.org/10.1126/scitranslmed.3006608. 12. Qiu X, Audet J, Wong G, Pillet S, Bello A, Cabral T, Strong JE, Plummer F, Corbett CR, Alimonti JB, Kobinger GP. 2012. Successful treatment of Ebola virus-infected cynomolgus macaques with monoclonal antibodies. Sci Transl Med 4:138ra81. http://dx.doi.org/10.1126 /scitranslmed.3003876. 13. Qiu X, Fernando L, Melito PL, Audet J, Feldmann H, Kobinger G, Alimonti JB, Jones SM. 2012. Ebola GP-specific monoclonal antibodies protect mice and guinea pigs from lethal Ebola virus infection. PLoS Negl Trop Dis 6:e1575. http://dx.doi.org/10.1371/journal.pntd.0001575. 14. Qiu X, Wong G, Audet J, Bello A, Fernando L, Alimonti JB, Fausther-Bovendo H, Wei H, Aviles J, Hiatt E, Johnson A, Morton J, Swope K, Bohorov O, Bohorova N, Goodman C, Kim D, Pauly MH, Velasco J, Pettitt J, Olinger GG, Whaley K, Xu B, Strong JE, Zeitlin L, Kobinger GP. 2014. Reversion of advanced Ebola virus disease in nonhuman primates with ZMapp. Nature 514:47–53. http://dx.doi.org /10.1038/nature13777. 15. Qiu X, Wong G, Fernando L, Audet J, Bello A, Strong J, Alimonti JB, Kobinger GP. 2013. mAbs and Ad-vectored IFN-alpha therapy rescue Ebola-infected nonhuman primates when administered after the detection of viremia and symptoms. Sci Transl Med 5:207ra143. http://dx.doi .org/10.1126/scitranslmed.3006605. 16. Nyakatura EK, Frei JC, Lai JR. 2015. Chemical and structural aspects of Ebola virus entry inhibitors. ACS Infect Dis 1:42–52. http://dx.doi.org/10 .1021/id500025n. 17. Saphire EO. 2013. An update on the use of antibodies against the filoviruses. Immunotherapy 5:1221–1233. http://dx.doi.org/10.2217/imt.13.124. 18. Qiu X, Audet J, Wong G, Fernando L, Bello A, Pillet S, Alimonti JB, Kobinger GP. 2013. Sustained protection against Ebola virus infection following treatment of infected nonhuman primates with ZMAb. Sci Rep 3:3365. http://dx.doi.org/10.1038/srep03365.

Journal of Virology

jvi.asm.org

277

Holtsberg et al.

19. Murin CD, Fusco ML, Bornholdt ZA, Qiu X, Olinger GG, Zeitlin L, Kobinger GP, Ward AB, Saphire EO. 2014. Structures of protective antibodies reveal sites of vulnerability on Ebola virus. Proc Natl Acad Sci U S A 111:17182–17187. http://dx.doi.org/10.1073/pnas.1414164111. 20. Lee JE, Fusco ML, Hessell AJ, Oswald WB, Burton DR, Saphire EO. 2008. Structure of the Ebola virus glycoprotein bound to an antibody from a human survivor. Nature 454:177–182. http://dx.doi.org/10 .1038/nature07082. 21. Maruyama T, Rodriguez LL, Jahrling PB, Sanchez A, Khan AS, Nichol ST, Peters CJ, Parren PW, Burton DR. 1999. Ebola virus can be effectively neutralized by antibody produced in natural human infection. J Virol 73:6024 – 6030. 22. Dias JM, Kuehne AI, Abelson DM, Bale S, Wong AC, Halfmann P, Muhammad MA, Fusco ML, Zak SE, Kang E, Kawaoka Y, Chandran K, Dye JM, Saphire EO. 2011. A shared structural solution for neutralizing ebolaviruses. Nat Struct Mol Biol 18:1424 –1427. http://dx.doi.org/10 .1038/nsmb.2150. 23. Warfield KL, Dye JM, Wells JB, Unfer RC, Holtsberg FW, Shulenin S, Vu H, Swenson DL, Bavari S, Aman MJ. 2015. Homologous and heterologous protection of nonhuman primates by Ebola and Sudan virus-like particles. PLoS One 10:e0118881. http://dx.doi.org/10.1371/journal.pone .0118881. 24. Hashiguchi T, Fusco ML, Bornholdt ZA, Lee JE, Flyak AI, Matsuoka R, Kohda D, Yanagi Y, Hammel M, Crowe JE, Jr, Saphire EO. 2015. Structural basis for Marburg virus neutralization by a cross-reactive human antibody. Cell 160:904 –912. http://dx.doi.org/10.1016/j.cell.2015.01 .041. 25. Stemcell Technologies. 2009. ClonaCell-HY hybridoma cloning kit technical manual version 2.2.0. Stemcell Technologies, Vancouver, Canada. 26. National Research Council. 2011. Guide for the care and use of laboratory animals, 8th ed. National Academies Press, Washington, DC. 27. Chandran K, Sullivan NJ, Felbor U, Whelan SP, Cunningham JM. 2005. Endosomal proteolysis of the Ebola virus glycoprotein is necessary for infection. Science 308:1643–1645. http://dx.doi.org/10.1126 /science.1110656. 28. Mahan AE, Tedesco J, Dionne K, Baruah K, Cheng HD, De Jager PL, Barouch DH, Suscovich T, Ackerman M, Crispin M, Alter G. 2015. A method for high-throughput, sensitive analysis of IgG Fc and Fab glycosylation by capillary electrophoresis. J Immunol Methods 417:34 – 44. http://dx.doi.org/10.1016/j.jim.2014.12.004. 29. Bray M, Davis K, Geisbert T, Schmaljohn C, Huggins J. 1998. A mouse model for evaluation of prophylaxis and therapy of Ebola hemorrhagic fever. J Infect Dis 178:651– 661. http://dx.doi.org/10.1086/515386. 30. Bray M, Davis K, Geisbert T, Schmaljohn C, Huggins J. 1999. A mouse model for evaluation of prophylaxis and therapy of Ebola hemorrhagic fever. J Infect Dis 179(Suppl 1):S248 –S258. http://dx.doi.org/10.1086 /514292. 31. Gibb TR, Bray M, Geisbert TW, Steele KE, Kell WM, Davis KJ, Jaax NK. 2001. Pathogenesis of experimental Ebola Zaire virus infection in BALB/c mice. J Comp Pathol 125:233–242. http://dx.doi.org/10.1053/jcpa .2001.0502. 32. Brannan JM, Froude JW, Prugar LI, Bakken RR, Zak SE, Daye SP, Wilhelmsen CE, Dye JM. 2015. Interferon alpha/beta receptor-deficient mice as a model for Ebola virus disease. J Infect Dis 212(Suppl 2):S282– S294. http://dx.doi.org/10.1093/infdis/jiv215. 33. Sanchez A, Trappier SG, Mahy BW, Peters CJ, Nichol ST. 1996. The virion glycoproteins of Ebola viruses are encoded in two reading frames and are expressed through transcriptional editing. Proc Natl Acad Sci U S A 93:3602–3607. http://dx.doi.org/10.1073/pnas.93.8.3602. 34. Volchkov VE, Feldmann H, Volchkova VA, Klenk HD. 1998. Processing of the Ebola virus glycoprotein by the proprotein convertase furin. Proc Natl Acad Sci U S A 95:5762–5767. http://dx.doi.org/10.1073/pnas.95.10 .5762.

278

jvi.asm.org

35. Olinger GG, Jr, Pettitt J, Kim D, Working C, Bohorov O, Bratcher B, Hiatt E, Hume SD, Johnson AK, Morton J, Pauly M, Whaley KJ, Lear CM, Biggins JE, Scully C, Hensley L, Zeitlin L. 2012. Delayed treatment of Ebola virus infection with plant-derived monoclonal antibodies provides protection in rhesus macaques. Proc Natl Acad Sci U S A 109:18030 – 18035. http://dx.doi.org/10.1073/pnas.1213709109. 36. Wilson JA, Hevey M, Bakken R, Guest S, Bray M, Schmaljohn AL, Hart MK. 2000. Epitopes involved in antibody-mediated protection from Ebola virus. Science 287:1664 –1666. http://dx.doi.org/10.1126/science .287.5458.1664. 37. Nimmerjahn F, Ravetch JV. 2008. Fcgamma receptors as regulators of immune responses. Nat Rev Immunol 8:34 – 47. http://dx.doi.org/10.1038 /nri2206. 38. Shields RL, Lai J, Keck R, O’Connell LY, Hong K, Meng YG, Weikert SH, Presta LG. 2002. Lack of fucose on human IgG1 N-linked oligosaccharide improves binding to human Fcgamma RIII and antibodydependent cellular toxicity. J Biol Chem 277:26733–26740. http://dx.doi .org/10.1074/jbc.M202069200. 39. Davies J, Jiang L, Pan LZ, LaBarre MJ, Anderson D, Reff M. 2001. Expression of GnTIII in a recombinant anti-CD20 CHO production cell line: expression of antibodies with altered glycoforms leads to an increase in ADCC through higher affinity for FC gamma RIII. Biotechnol Bioeng 74:288 –294. 40. Baize S. 2015. Ebola virus in West Africa: new conquered territories and new risks— or how I learned to stop worrying and (not) love Ebola virus. Curr Opin Virol 10:70 –76. http://dx.doi.org/10.1016/j.coviro .2015.01.008. 41. Muyembe-Tamfum JJ, Mulangu S, Masumu J, Kayembe JM, Kemp A, Paweska JT. 2012. Ebola virus outbreaks in Africa: past and present. Onderstepoort J Vet Res 79:451. http://dx.doi.org/10.4102/ojvr.v79i2.451. 42. Fan X, Hashem AM, Chen Z, Li C, Doyle T, Zhang Y, Yi Y, Farnsworth A, Xu K, Li Z, He R, Li X, Wang J. 2015. Targeting the HA2 subunit of influenza A virus hemagglutinin via CD40L provides universal protection against diverse subtypes. Mucosal Immunol 8:211–220. http://dx.doi.org /10.1038/mi.2014.59. 43. Lerner RA. 2011. Rare antibodies from combinatorial libraries suggests an S.O.S. component of the human immunological repertoire. Mol Biosyst 7:1004 –1012. http://dx.doi.org/10.1039/c0mb00310g. 44. Schneemann A, Speir JA, Tan GS, Khayat R, Ekiert DC, Matsuoka Y, Wilson IA. 2012. A virus-like particle that elicits cross-reactive antibodies to the conserved stem of influenza virus hemagglutinin. J Virol 86:11686 – 11697. http://dx.doi.org/10.1128/JVI.01694-12. 45. Wei CJ, Boyington JC, McTamney PM, Kong WP, Pearce MB, Xu L, Andersen H, Rao S, Tumpey TM, Yang ZY, Nabel GJ. 2010. Induction of broadly neutralizing H1N1 influenza antibodies by vaccination. Science 329:1060 –1064. http://dx.doi.org/10.1126/science.1192517. 46. Chen G, Koellhoffer JF, Zak SE, Frei JC, Liu N, Long H, Ye W, Nagar K, Pan G, Chandran K, Dye JM, Sidhu SS, Lai JR. 2014. Synthetic antibodies with a human framework that protect mice from lethal Sudan ebolavirus challenge. ACS Chem Biol 9:2263–2273. http://dx.doi.org/10 .1021/cb5006454. 47. Flyak AI, Ilinykh PA, Murin CD, Garron T, Shen X, Fusco ML, Hashiguchi T, Bornholdt ZA, Slaughter JC, Sapparapu G, Klages C, Ksiazek TG, Ward AB, Saphire EO, Bukreyev A, Crowe JE, Jr. 2015. Mechanism of human antibody-mediated neutralization of Marburg virus. Cell 160:893–903. http://dx.doi.org/10.1016/j.cell.2015.01.031. 48. Bruhns P. 2012. Properties of mouse and human IgG receptors and their contribution to disease models. Blood 119:5640 –5649. http://dx.doi.org /10.1182/blood-2012-01-380121. 49. Bournazos S, Klein F, Pietzsch J, Seaman MS, Nussenzweig MC, Ravetch JV. 2014. Broadly neutralizing anti-HIV-1 antibodies require Fc effector functions for in vivo activity. Cell 158:1243–1253. http://dx.doi .org/10.1016/j.cell.2014.08.023.

Journal of Virology

January 2016 Volume 90 Number 1