ABSTRACT
Diphtheria antitoxin (DAT) has been the cornerstone of the treatment of Corynebacterium diphtheriae infection for more than 100 years. Although the global incidence of diphtheria has declined steadily over the last quarter of the 20th century, the disease remains endemic in many parts of the world, and significant outbreaks still occur. DAT is an equine polyclonal antibody that is not commercially available in the United States and is in short supply globally. A safer, more readily available alternative to DAT would be desirable. In the current study, we obtained human monoclonal antibodies (hMAbs) directly from antibody-secreting cells in the circulation of immunized human volunteers. We isolated a panel of diverse hMAbs that recognized diphtheria toxoid, as well as a variety of recombinant protein fragments of diphtheria toxin. Forty-five unique hMAbs were tested for neutralization of diphtheria toxin in in vitro cytotoxicity assays with a 50% effective concentration of 0.65 ng/ml for the lead candidate hMAb, 315C4. In addition, 25 μg of 315C4 completely protected guinea pigs from intoxication in an in vivo lethality model, yielding an estimated relative potency of 64 IU/mg. In comparison, 1.6 IU of DAT was necessary for full protection from morbidity and mortality in this model. We further established that our lead candidate hMAb binds to the receptor-binding domain of diphtheria toxin and physically blocks the toxin from binding to the putative receptor, heparin-binding epidermal growth factor-like growth factor. The discovery of a specific and potent human neutralizing antibody against diphtheria toxin holds promise as a potential therapeutic.
INTRODUCTION
Corynebacterium diphtheriae is a Gram-positive bacteria that secretes a potent toxin that inhibits protein synthesis in eukaryotic cells by disrupting elongation factor 2 (EF2) function (1). The main function of the toxin is cessation of protein synthesis, causing cell death. The initial symptoms of C. diphtheriae infection include pharyngeal pseudomembrane formation or cutaneous ulcer. If untreated, the diphtheria toxin can enter the circulation leading to cardiac and neurologic sequelae (2). Diphtheria toxin (DT) consists of a single protein with a disulfide bond linking two fragments (3, 4). Fragment A consists of the catalytic domain and fragment B contains the translocation and receptor-binding domains. DT is estimated to be lethal at ∼0.1 μg/kg in humans, and vaccination with formalin-treated culture filtrate containing DT, also known as diphtheria toxoid, has been utilized as an effective prophylactic since the early 1920s (2).
According to the World Health Organization (WHO), in 2011 there were 4,887 cases of diphtheria reported throughout the world (http://www.who.int/immunization_monitoring/diseases/diphteria/en/index.html). Although diphtheria is preventable by vaccination, the disease is thought to persist because of regional variations in vaccine compliance, inadequate booster regimens and immunosenescence (5). Complications associated with diphtheria infection can be prevented by the intravenous administration of 10,000 to 100,000 IU of equine diphtheria antitoxin (DAT), the dose depending on the extent of the infection. Administration of DAT is complicated since it is an equine derivative with a significant risk of acute and delayed hypersensitivity (http://www.cdc.gov/vaccines/vpd-vac/diphtheria/dat/downloads/protocol_032504.pdf). The efficacy of DAT depends on the rapidity by which it can be administered after the identification of clinical disease (6). Antibiotics serve an important adjunctive role to reduce transmission of the highly contagious organism (7). During the recent outbreak of diphtheria in the Newly Independent States, after the fall of the Soviet Union, the lack of rapid access to DAT was thought to contribute to the excessive mortality (8, 9). In the aftermath of this outbreak, many national and regional health authorities have tried to maintain DAT stockpiles to ensure that their citizens have access to DAT in the event of future diphtheria outbreaks. However, the global supply of DAT is increasingly jeopardized due to the limited number of manufacturers. The reasons for the dwindling supply of DAT are probably multifactorial, but the result is outdated stockpiles in some countries and a total lack of product in other parts of the world (9).
Monoclonal antibody (MAb) technology may provide an approach for the development of a safer but comparably efficacious alternative to DAT. Diphtheria toxin is highly conserved among bacterial strains which increases the likelihood that a neutralizing MAb may be a successful therapeutic (10, 11). In addition, a hMAb is likely to be safer than equine DAT since hypersensitivity reactions to human antibodies are much less common.
Here we describe the identification of an anti-diphtheria hMAb isolated from antibody-secreting cells (ASCs) obtained from a Td vaccine-immunized human volunteer. This antibody, 315C4, potently neutralized diphtheria toxin in a cell-based cytotoxicity assay and prevented toxin from binding to the known diphtheria receptor, heparin binding-epidermal growth factor-like growth factor (HB-EGF [12]). 315C4 also completely protected guinea pigs in a lethality model, and a relative potency compared to the DAT standard was estimated.
MATERIALS AND METHODS
PBMC isolation and ASC sorting procedure.Peripheral blood was obtained from five healthy volunteers through a human research protocol approved by the Institutional Review Board of the University of Massachusetts Medical School. Volunteers who had not received a tetanus/diphtheria (Td) booster in the past 2 years were immunized with one dose of Td vaccine (MassBiologics). At 7 days postvaccination, blood was drawn, and B cells were enriched using RosetteSep (StemCell Technologies). Peripheral blood mononuclear cells (PBMCs) were isolated by a Ficoll-Paque gradient (GE Healthcare). PBMCs were washed and resuspended in phosphate-buffered saline (PBS) with 2% fetal calf serum (FCS) at an average density of 2.5 × 107cells/ml. Prepared PBMCs were incubated with a cocktail of antibodies against CD3, CD20, CD27, CD19, and CD38 (Invitrogen, catalog nos. MHCD0301, MHCD2001, MHCD2704, MHCD1922, and MHCD3819) for 30 min at 4°C. After incubation, cells were washed and resuspended in PBS-2% FCS for fluorescence-activated cell sorting (FACS). Using flow cytometry, stained PBMCs were first analyzed for CD3, CD20, and CD19 staining, and CD3− CD20− cells were further analyzed for the expression of CD27 and CD38. Cells that expressed high levels of CD27 and CD38 were sorted, yielding a population of cells with the phenotype CD19+ CD3− CD20−/low CD27hi CD38hi, the predicted phenotype of the ASCs (13). In addition, cells with the phenotype of CD19hi/CD20hi were sorted as the total B cells. All cells that fit the sorting criteria were bulk sorted to a 15-ml conical tube, followed by single-cell sorting to microtiter plates.
Antibody cloning, expression, and purification.Single-cell reverse transcription-PCR (RT-PCR) was performed with a cocktail of primers (0.2 μM final concentration) targeting the variable region of all known human IgG1 heavy-chain genes and kappa light-chain genes (see Table S1 in the supplemental material). The resulting RT-PCR products were further amplified by a second round of PCR with nested primers containing restriction sites for subcloning. The nested light-chain forward primers also contain an overhang corresponding to the BGHrev sequencing primer sequence (modCLONE-VKs in Table S1 in the supplemental material). The RT-PCR product that had both detectable heavy- and light-chain genes were sequenced and cloned into a mammalian expression vector in frame with an osteonectin leader sequence, as well as the human IgG1 heavy-chain constant region or human kappa light chain constant region contained in the vector. Constructs were transfected into 293T cells, and cell supernatants were harvested 5 days posttransfection for antibody purification by protein A-Sepharose. For stable cell line generation, the vectors containing heavy-chain and light-chain antibody genes were combined into a polycistronic vector and electroporated into CHO cells. Stable transfectants were selected and expanded for large-scale antibody purification. Antibody was purified by protein A chromatography, followed by Fractogel (Millipore) chromatography and nanofiltration.
Recombinant protein expression, purification, and Western blotting.Genomic DNA was isolated from a pellet of C. diphtheriae bacteria (Toronto strain, obtained by Massachusetts Public Health Biologic Laboratories in 1936) using a DNeasy kit (Qiagen). PCR was performed to subclone fragment A, fragment B, the T domain, and the R domain of the diphtheria toxin into a pET vector containing N-terminal thioredoxin/histidine (TRX) and C-terminal myc/His6 epitope tags (Novagen). Vectors were transformed into BL21(DE3) cells (Invitrogen) and expression was induced with 0.6 mM IPTG (isopropyl-β-d-thiogalactopyranoside; Sigma). Proteins were purified with Ni-NTA agarose (Invitrogen) and eluted with 250 mM imidazole (Sigma). All proteins were dialyzed against PBS, concentrated, and stored at −20°C. Protein purity and integrity was assessed by SDS-PAGE followed by Sypro Ruby (Invitrogen) staining as described by the manufacturer. R domain truncations and point mutants were created by site-directed mutagenesis or subcloning from the original diphtheria R domain vector, expressed and purified as described. For Western blot analysis, cell pellets were run on a 12% Tris-glycine gel under reducing conditions. Proteins were detected using 315C4 or an anti-His6 MAb (in-house), followed by horseradish peroxidase (HRP)-conjugated anti-human or anti-mouse IgG (Jackson Immunoresearch). Membranes were developed with chemiluminescent detection (GE Healthcare).
Relative affinity determination and epitope competition assays.Fortebio Octet technology was used to determine the affinity of the hMAb 315C4 to diphtheria toxin, fragment B, and fragment R. Anti-human IgG biosensors were used to capture 315C4 in PBS (10 μg/ml), followed by binding to the indicated antigen at three concentrations: 12.5, 25, and 50 μg/ml. The results for association (Kon) and dissociation (Koff) rate constants were calculated to derive the dissociation constant (KD) using Fortebio software. For competitive binding assay, anti-human IgG Biosensors were first loaded with one hMAb, followed by binding to toxoid, fragment B, or fragment R. The biosensors were then reloaded with the first hMAb to ensure saturation of the fragment. These biosensor-hMAb-antigen complexes were further loaded with the second hMAb in PBS (10 μg/ml) to assess binding competition. Each assay was repeated with the reversed order of hMAb loading to confirm the results.
Epitope mapping ELISA.Various dilutions of purified hMAbs were tested in an enzyme-linked immunosorbent assay (ELISA) for reactivity against diphtheria toxoid, TRX-fragment A-His, TRX-fragment B-His, TRX-T domain-His, and TRX-R domain-His. EIA/RIA high binding plates (Costar) were coated with 0.5 limit of flocculation (Lf) of toxoid/ml or 1 μg/ml of purified recombinant toxin fragments in PBS, followed by incubation overnight at 4°C. Primary hMAbs were diluted to 1 μg/ml in enzyme-linked immunosorbent assay (ELISA) blocking buffer (PBS, 1% bovine serum albumin, 0.05% Tween) in the first well, titrated 1:3 across the plate, and incubated for 1 h at room temperature. Antibody binding was detected with goat anti-human IgG-alkaline phosphatase (1:2,000; Jackson Immunoresearch). Plates were developed with p-nitrophenyl phosphate disodium salt (PNPP) at 1 mg/ml in 1 M diethanolamine for 20 min at room temperature. The absorbance at 405 nm was measured and analyzed using a Molecular Devices Emax plate reader with Softmax software.
Receptor-blocking ELISA.ELISA plates were coated with 1 μg/ml of recombinant human HB-EGF (rhHB-EGF; R&D Systems) overnight at 4°C. After 24 h, in a 96-well round-bottom plate, a titration of antibody was incubated at room temperature with a constant amount of 1 μg/ml of the TRX-fragment B-His or 0.5 Lf/ml of diphtheria toxin. After 1 h of incubation, the antibody and antigen mixture was transferred into an rhHB-EGF-coated plate, followed by incubation for one additional hour at room temperature. The binding activity was detected using an anti-His6 MAb for fragment B or a guinea pig polyclonal serum conjugated to streptavidin to detect toxin. Additional controls were also present, including a nonspecific antibody and diphtheria antitoxin (lot F4510; Center for Biologics Evaluation and Research Standard DAT, U.S. Food and Drug Administration [FDA]).
Cell culture.HEK-293T cells and Vero cells were obtained from the American Type Culture Collection. HEK-293T cells were grown in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum (FBS) and 100 IU of penicillin-streptomycin. Vero cells were grown in Eagle minimal essential medium supplemented with 5% FBS and 100 IU of penicillin-streptomycin. All cells were grown at 37°C with 5% CO2.
Vero cell neutralization assay.A titration of purified antibody ranging from 10 ng/ml to 0.04 pg/ml was incubated for 1 h with a constant amount of diphtheria toxin (lot K559; MassBiologics) in a 96-well tissue culture plate. Vero cells were added to each well at a density of 2 × 104 cells/well. Plates were incubated for 72 h at 37°C in air supplemented with 5% CO2. The cytopathic effect was estimated by alamarBlue staining (Invitrogen; assessing metabolic activity) and read at wavelengths of 544/590 nm on a Victor3 multilabel counter (Applied Biosystems). The resulting fluorescence of each well was plotted, and neutralization of toxin was analyzed based on increased fluorescence corresponding to the reduction of resazurin in alamarBlue by metabolic intermediates. In addition, a titration of diphtheria antitoxin (FDA standard DAT) starting at 0.003 IU/ml was tested in the assay as control. The relative fluorescence for each well was determined, and the results plotted using a four-parameter fit.
Guinea pig lethality assays.All animal experiments were performed in accordance with the National Institutes of Health guidelines and approved by the University of Massachusetts Medical School Institutional Animal Care and Use Committee. Female Hartley guinea pigs were purchased from Charles River Laboratories and quarantined for 1 week. Various doses of purified endotoxin-free 315C4 or FDA standard DAT were tested by mixing with ∼2 Lf of diphtheria toxin (exactly 1.968 Lf, which is defined as the minimal dose of internal toxin lot K561 that, when premixed with 1 IU of FDA standard DAT, causes animal death within 96 h). Mixtures were incubated for 1 h at room temperature and injected into guinea pigs subcutaneously. In addition, in the preventative guinea pig model, 315C4 or DAT (obtained from the FDA, the WHO, or the Centers for Disease Control and Prevention) was subcutaneously injected 24 h prior to ∼2 Lf of diphtheria toxin. Two to four guinea pigs weighing 230 to 300 g were tested for each dose. Animals were monitored 30 days for symptoms of diphtheria, including dehydration, hind-limb weakness, and lethargy. Symptoms and time of death or euthanasia were recorded for each animal throughout the experiment. Animals remaining symptom-free at day 30 (720 h) were euthanized.
RESULTS
Isolation of hMAbs from Td-boosted healthy volunteers.Five healthy volunteers were immunized with Td vaccine (MassBiologics), and PBMCs were isolated from these individuals 7 days after vaccination. The PBMCs were analyzed by flow cytometry for the expression of CD3, CD20, and CD19 (see Fig. S1A in the supplemental material). CD3− CD20−/low cells were further analyzed for expression of CD27 and CD38 (see Fig. S1B in the supplemental material) and CD27hi CD38hi cells were sorted to 96-well plates for downstream single-cell analysis. The final sorted ASCs had a CD3− CD20−/low CD27hi CD38hi phenotype (19). In addition, cells with a CD19hi CD20hi phenotype were sorted separately as the total B cell population to isolate additional antibodies. Due to the effectiveness of antibody recovery in the ASC population, only three of the five patient samples were additionally screened using the total B cell isolation method.
Single-cell RT-PCR was performed on isolated ASCs and B cells with a cocktail of primers targeting the variable region of the heavy-chain genes and light-chain genes. For the present study, we chose to focus on antibodies of the IgG1 subclass with kappa light chains. Forward primers were designed to amplify all known VH and Vκ variable regions (identified using The International Immunogenetics Information System [IMGT]), and reverse primers were specific to human IgG1 and kappa constant regions. A total of 2,436 cells were screened, which yielded 437 hMAbs with paired VH and Vκ RT-PCR products (Table 1). Products were cloned and expressed by transient transfection in HEK-293T cells. Antibodies were screened for recognition of diphtheria toxoid by ELISA, and 53 functional hMAbs specific for diphtheria toxoid were identified. Antibodies were grouped based on the VH family and the VH CDR3 (HCDR3) length and amino acid content. Of the 53 reactive antibodies, 12 contained redundant HCDR3 sequences, yielding a total of 45 hMAbs with unique HCDR3 regions (Table 2).
Screening of human ASC and B cell populations for diphtheria toxoid-reactive antibodies
Heavy-chain complementarity-determining region 3 (HCDR3) of anti-diphtheria toxin hMAbsa
Selection of a lead-candidate hMAb.All 45 unique hMAbs were tested for the ability to inhibit cytotoxicity of diphtheria toxin using a cell-based neutralization assay using Vero cells. A total of 20 hMAbs demonstrated significant toxin neutralization activity with half-maximal effective neutralization concentrations (EC50) ranging from 0.003 to 1.75 nM (Table 3). These 20 hMAbs were selected for further analysis.
Characterization and neutralization activity of hMAbs isolated from human ASCs
Diphtheria toxin is composed of multiple functional domains. Fragment A consists of the catalytic domain, and fragment B contains the translocation (T) and receptor-binding (R) domains. Each of these four domains was expressed and purified from E. coli in the context of a fusion with thioredoxin (TRX; N-terminal) and a C-terminal His6 epitope tag. The purity and integrity of all recombinant proteins were assessed by SDS-PAGE, followed by Sypro Ruby staining (see Fig. S2 in the supplemental material). The 20 hMAbs with significant neutralizing activity were tested for reactivity with the diphtheria toxin functional domain fusion proteins (TRX-fragment A-His, TRX-fragment B-His, TRX-T domain-His, and TRX-R domain-His) by ELISA. Various concentrations of anti-diphtheria antibody, as well as an anti-His6 antibody (a control for the presence of antigen), was applied to each toxin fragment to grossly determine which domain was recognized by each hMAb. The panel of lead-candidate hMAbs contained antibodies that recognized all known functional domains of diphtheria toxin, including the cytotoxic domain (fragment A), the translocation domain (T domain), and the receptor-binding domain (R domain) of fragment B. In addition, some hMAbs were unable to recognize any of the recombinant fragments despite recognizing diphtheria toxoid and were therefore classified as conformation-dependent antibodies (Table 3).
All 20 toxin-neutralizing hMAbs were grossly evaluated in a guinea pig model of intoxication. The guinea pig is the accepted model system for the study of diphtheria intoxication due to its sensitivity to the toxin compared to the highly resistant mouse and rat. A guinea pig lethality model has been well established and standardized for the measurement of DAT potency of neutralizing diphtheria toxin and subsequently preventing mortality (14). To quickly assess the ability of the large panel of hMAbs to neutralize diphtheria toxin in vivo, neutralizing hMAbs were preincubated with diphtheria toxin and subcutaneously injected into animals at a single dose of 50 μg. Animals were closely monitored over 30 days for symptoms of diphtheria intoxication which includes symptoms such as dehydration, hind-limb weakness, and lethargy. Although many of the hMAbs in our panel were able to significantly delay symptoms at a dose of 50 μg, only one was able to completely protect the animals for the full 30 days (Table 3). This hMAb, designated 315C4, became the focus of our investigation.
315C4 binds to diphtheria R domain and competes with a panel of R domain antibodies.Gross epitope mapping of 315C4 revealed interaction with fragment B and, more specifically, the 17-kDa R domain of diphtheria toxin with an affinity of 0.3 nM, as determined by Octet analysis. No recognition of fragment A or the T domain, which consists of the N-terminal portion of fragment B, was observed by ELISA (Fig. 1A). An anti-His6 antibody (anti-His) detection was performed as a control for protein expression and coating (data not shown). To further define the 315C4 epitope, multiple R domain truncations were subsequently generated; however, 315C4 did not recognize any smaller fragments of the R domain. In contrast, antibodies with continuous epitopes, such as 3B3A5, recognized smaller truncations of the R domain (Fig. 1B). Therefore, the epitope of 315C4 was determined to recognize a conformational component of the R domain of diphtheria toxin. In addition, 315C4 was tested for binding with R domain point mutants at a number of key residues of the toxin/receptor binding interface based on the solved crystal structure (15). These mutants include F389D, A430D, L433D, I461D, V468D, F470D, G510D, L512D, and V523D (15). Due to the difficulty of expression of stable soluble versions of some mutants, whole-cell pellets were used in Western blots for testing 315C4 recognition. Among all of the mutations, only L512D appeared to significantly affect 315C4 binding (Fig. 1C).
315C4 recognizes a conformational epitope on the R domain of diphtheria toxin. (A) 315C4 binding to different domains of diphtheria toxin was measured by a titration ELISA. (B) R domain truncation proteins are schematically depicted with the residues defined. The full R domain consists of amino acids 379 to 535 of diphtheria toxin. ELISA results of the recognition by 315C4 (or a relevant diphtheria-specific antibody, 3B3A5) on truncated R domains are represented with a “+” or a “−.” (C) Cell pellets of R domain point mutants (mutated residue indicated above band) were run on a reduced 12% Tris-glycine gel. Gels were blotted with either 315C4 or an anti-His6 MAb, followed by HRP-conjugated anti-human or anti-mouse IgG. (D) Competition of various diphtheria-specific hMAbs for binding to diphtheria toxoid by Octet analysis. The epitopes of hMAbs are represented graphically by overlapping circles. Antibodies recognizing the same epitope are grouped in the same circle, whereas antibodies with competing epitopes are represented as overlapping circles. T domain-specific antibody 3B4G1 and A domain-specific antibody 3B2F3 were used as controls.
To further elucidate the epitope of 315C4, we performed toxoid binding competition analysis with all neutralizing antibodies from our original panel. Antigen was captured on an Octet biosensor, followed by saturation with a primary antibody. The biosensor (bound by antigen and the primary antibody) was then exposed to a second antibody. Detection of a mass increase by the biosensor indicates the two antibodies bind to different epitopes on the protein. All neutralizing antibodies were tested in the toxoid binding competition assay, and the results are summarized in a graphic epitope map in Fig. 1D. Also, a representative sensogram generated by the Octet is shown in Fig. S3 in the supplemental material. Surprisingly, the binding of 315C4 prevented the binding of all other R domain antibodies tested, suggesting that some portion of the 315C4 epitope was shared by all R domain-directed antibodies. However, many of the R domain antibodies did not compete with one another, indicating that these antibodies recognize a variety of distinct R domain epitopes (Fig. 1D). The T domain-specific antibody 3B4G1 and the A domain-specific antibody 3B2F3 were used as controls and did not compete with the R domain antibodies.
315C4 blocks diphtheria toxin and fragment B from binding to HB-EGF.Diphtheria toxin binds to the membrane-bound form of HB-EGF and is internalized into the cell by endocytosis (12). Since 315C4 interacts with the receptor-binding domain of diphtheria toxin, we hypothesized that 315C4 had the ability to block the binding of diphtheria toxin (or its fragments and domains) to HB-EGF.
To explore this hypothesis, a receptor-blocking ELISA was developed. 315C4, DAT, or an irrelevant human antibody were preincubated with fragment B fusion protein or diphtheria toxin and then allowed to bind sHB-EGF, a fragment of HB-EGF that lacks the transmembrane domain but still interacts with diphtheria toxin (16). 315C4 and DAT both significantly decreased the binding of fragment B (Fig. 2A) and diphtheria toxin (Fig. 2B) to HB-EGF compared to a non-R domain MAb and an irrelevant human control antibody. These data suggest that 315C4 neutralization of diphtheria toxin is mediated by the blockade of toxin binding to the receptor HB-EGF.
315C4 blocks diphtheria receptor (HB-EGF) binding with diphtheria toxin. ELISA plates coated with 1 μg/ml of recombinant human HB-EGF were incubated with Trx fusion fragment B protein (A) or 0.5 Lf/ml of diphtheria toxin (B) that was premixed with a titration of antibody (315C4 or negative-control MAbs). The binding activity was detected with an anti-His antibody (A) or a guinea pig polyclonal serum conjugated to streptavidin (B) using standard ELISA methods. A titration of FDA standard DAT was tested in parallel as a control (lower panels). An A domain binding MAb and an unrelated MAb were included in the assay as controls.
315C4 neutralization potency in cell-based assays.To assign an “official” potency for 315C4 for clinical studies, we created a stable CHO cell line for 315C4 production. 315C4 was purified from CHO cell supernatant in a manner similar to purification of antibody for ultimate human use. We assessed the ability of this CHO-produced antibody to neutralize toxin in the Vero cell cytotoxicity assay (Fig. 3A). Using a four-parameter fit, the half-maximal effective neutralization concentration (EC50) was calculated to be 0.65 ng/ml. In addition, the EC50 of DAT was calculated to be 0.00032 IU/ml (Fig. 3B). Therefore, by comparing the EC50s in the Vero cytotoxicity assay, 1 IU of DAT was determined to be equivalent to 2 μg of 315C4 or 500 IU/mg. We also compared the amount of 315C4 and DAT needed to achieve full neutralization (EC99), which was accomplished with 5.5 ng/ml of 315C4 and 0.00075 IU/ml of DAT, resulting in a relative potency of 136 IU/mg.
315C4 inhibits diphtheria toxin cytotoxicity on Vero cells. (A) A titration of purified antibody 315C4 ranging from 0.04 pg/ml to 10 ng/ml was mixed with a constant amount of diphtheria toxin, followed by incubation with Vero cells for 72 h. The relative fluorescence of alamarBlue dye was determined using a Multilabel Counter, and the results were plotted using a four-parameter fit. The half-maximum effective neutralization (EC50) was calculated by SoftMax Pro software. (B) A titration of diphtheria antitoxin (FDA standard DAT) was used as a control.
Potency assignment of 315C4 in the guinea pig lethality model.We had previously confirmed that 315C4 could protect guinea pigs from morbidity and mortality associated with diphtheria intoxication at a dose of 50 μg. To determine the minimal dose of 315C4 required to offer full protection, an expanded set of guinea pig assays were initiated with CHO-produced antibody utilizing two different endpoint criteria. In the standard lethality model used to assign potency to DAT, the antitoxin (either immune serum or DAT) was given to animals after a 1-h preincubation with toxin. According to the international standard, antitoxin, when given together with a standard dose of toxin, is assigned a potency of 1 IU when guinea pig mortality is observed within a time window of 40 to 96 h (14). We refer to this standard approach of potency assignment of DAT as the “window dose” for protection.
To determine the window dose of 315C4 using the standard assay for DAT potency, a range of 315C4 was mixed with standard toxin and administered to guinea pigs. Animals given 5 μg of 315C4 in combination with toxin succumbed to intoxication within the “window” at an average time of 48.5 h; however, doses of 6, 7, and 8 μg provided protection from intoxication for >96 h, with mean times ranging from 185 to 308 h (Table 4). The 1-IU DAT control combined with toxin succumbed to intoxication within the expected range (44.5 h). Therefore, the “window dose” of 315C4 was estimated at 5 μg, which is equivalent to 1 IU (0.2 IU/μg or 200 IU/mg).
Protection of guinea pigs from diphtheria toxin lethal challengea
Given the inherent differences between monoclonal and polyclonal antibody preparations, an alternative, more conservative method for potency assignment was implemented in which the minimum protective dose of the antibody was defined as complete protection from disease symptoms and mortality for up to 30 days. Animals that displayed symptoms but recovered were not considered fully protected. We believe that protection for 30 days is equivalent to complete, durable protection from the initial dose of toxin.
To use complete protection for 30 days as a potency assignment metric, we first needed to determine the minimal dose of DAT that met this more stringent criterion. Standard toxin was mixed with a range of DAT from 1.25 to 2.0 IU (Table 4). Animals receiving 1.25 IU of DAT succumbed to intoxication at an average of 288 h, whereas animals receiving 1.5 IU of DAT varied from death at 264 h to full protection. Animals receiving at least 1.6 IU of DAT were fully protected from intoxication. Based on these results, the minimal fully protective dose of DAT was estimated to be between 1.5 and 1.6 IU (Table 4).
To determine the minimal fully protective dose of 315C4, we tested doses ranging from 20 to 40 μg in 5-μg intervals in combination with standard toxin. All animals receiving a dose of at least 25 μg of 315C4 were fully protected from the effects of diphtheria toxin, whereas one of the animals receiving 20 μg of 315C4 showed symptoms of moderate hind limb weakness in the last half of the assay yet recovered before euthanasia was performed (Table 4). Therefore, 25 μg of 315C4 was roughly equivalent to 1.6 IU of DAT (64 IU/mg).
315C4 is an effective prophylactic agent for diphtheria intoxication.315C4 was then tested in a more stringent model for the ability to prevent intoxication. In this model, rather than premixing toxin with antibody as performed in the standard potency assay, various doses of 315C4 and DAT were subcutaneously injected into animals 24 h prior to administration of 2 Lf of diphtheria toxin. Animals were observed for disease symptoms and mortality for up to 30 days. Interestingly, all animals that received 20 μg (equivalent to 0.1 mg/kg) or greater of 315C4 survived the entire assay with no symptoms (Table 5), whereas animals that received ≥5 IU of DAT also survived the entire assay with no symptoms (Table 5). These results demonstrate that it is feasible to protect animals from intoxication by pretreatment with 315C4. In this more stringent protection model, 20 μg of 315C4 is equivalent to 5 IU of DAT, resulting in a relative potency of 250 IU/mg.
315C4 prevents diphtheria intoxication of guinea pigsa
DISCUSSION
Developing an alternative to an established biological therapeutic that has been used for over 100 years poses unique challenges. However, in light of the risk of hypersensitivity associated with DAT and its extremely limited supply, it is regrettable that a licensed alternative does not exist. This may be due to a perception that there is limited demand for an alternative to DAT and that the potential market for such a product continues to decline. However, while the global incidence of diphtheria is relatively small, the disease seems unlikely to be eradicated in the near future. Development of a safe, effective, and readily available alternative to DAT is a worthwhile goal.
There have been past attempts to make alternatives to DAT. These include a polyclonal human immunoglobulin product that was not sufficiently potent to provide a replacement for DAT (17) and, like all products derived from pooled serum, carried the potential risk of contamination with blood-borne pathogens (18). Another approach was to use a soluble receptor analog, similar to HB-EGF, with the heparin-binding domain deleted and the EGF-like domain mutated to block mitogenic signaling (19). Despite this HB-EGF alteration, the endogenous mitogenic pathways signaled through EGFR are still stimulated, albeit 100-fold less than the wild-type protein, but enough to cause concern of significant side effects. In addition, this treatment consisted of a recombinant protein with inherently less stability than hMAbs, a shorter half-life, and a more complicated purification process.
Antibodies targeting diphtheria toxin have been created for research purposes, and yet none to our knowledge have been developed beyond the preclinical stage. Antibodies that target and inhibit the enzymatic activity and bind without inhibitory effect have been helpful in elucidating the structure and function of the toxin (20, 21). In addition, diphtheria antibodies have been used to rapidly detect diphtheria toxin in patient samples (22). Lastly, Danelli et al. developed murine MAbs that neutralized both fragment A and fragment B of diphtheria toxin and protected guinea pigs from intoxication for at least 6 days (23). These MAbs were not, to our knowledge, brought forward for further development.
One hMAb (DTD4) directed against fragment B of the diphtheria toxin has been shown to have excellent neutralization potency (73.6 IU/mg) and was suggested to be an excellent candidate for the replacement of DAT (24). It is unclear whether this antibody recognizes the translocation (T) domain or the receptor (R)-binding domain. The potency of this antibody was assigned using both the rabbit skin test and in vitro cellular cytotoxicity assays. In our experience, the in vivo skin erythema assay and cell-based assay are not good predictors of durable protection from lethal administration of toxin. As described in the present study, many hMAbs showed very potent neutralization activity in the Vero cell assay but did not fully protect in the guinea pig lethality model. In addition, numerous antibodies described here were tested in a guinea pig skin erythema test, and many showed excellent potency (data not shown); however, only 315C4 provided robust, durable protection from intoxication in the guinea pig lethality model. It is unclear whether DTD4 would provide the robust protection of guinea pigs in the lethality model described herein. In addition, the DTD4 antibody may have properties that make manufacture difficult. The HCDR3 of DTD4 contains an N-linked glycosylation motif that would be expected to adversely impact the consistency of the product from batch-to-batch and may also impact MAb stability if the glycosylation is critical to MAb binding and neutralization. In contrast, 315C4 has been thoroughly tested for large-scale expression and purification in 10-liter bioreactors without processing issues (data not shown). Given these factors and the current available data, we believe that 315C4 provides an improved candidate for the development of a DAT replacement.
Diphtheria antitoxin potency has been assigned historically using either the cutaneous erythrogenic assay in rabbits or guinea pigs or the neutralization of toxin in guinea pigs measured by delay of mortality for up to 96 h (25). More recently, an in vitro neutralization assay that uses Vero cells has shown promise as an alternative to these in vivo assays (23, 26). In the present study, we utilized both in vitro and in vivo approaches to assign the potency of 315C4. The potency of 315C4 based on the EC99 in the in vitro Vero assay was similar to the potency determined using the in vivo intoxication assay with an endpoint of asymptomatic survival. In these experiments, the potency of 315C4 relative to DAT was calculated in each assay to obtain a ratio of IU/mg and was estimated to be 136 IU/mg for the Vero assay and 64 IU/mg for the animal assay. Although assignment of potency based on symptom-free survival at 30 days in the in vivo model was felt to be the most stringent determination of potency, the Vero assay may be a feasible way to measure lot-to-lot variability in the potency of manufactured anti-diphtheria antibodies and also to measure the concentration of neutralizing antibodies in human serum. Further standardization of the Vero cell assay relative to an in vivo method would be useful in this regard.
We showed that 315C4 protected susceptible animals from symptoms and death due to diphtheria intoxication for 30 days, which we speculate equates to complete protection from disease. A more conservative assessment of potency based on symptom-free survival to 30 days, was preferred over the standard 40- to 96-h window. A conservative approach is warranted in assigning potency and demonstrating the effectiveness of a therapeutic directed against a lethal toxin such as diphtheria. The apparent potency of 315C4 should readily allow for product formulation comparable to the current lot of DAT available under IND in the United States (Instituto Butantan). The preliminary potency of 315C4 is conservatively estimated to be 64 IU/mg, which should allow formulation at a concentration of <20 mg/ml to achieve a deliverable dose of 10,000 to 100,000 IU, a concentration comparable to other United States-licensed hMAbs.
In summary, we identified an hMAb that potently neutralizes diphtheria toxin in vitro and in vivo. Additional preclinical and clinical development is planned to demonstrate the safety and efficacy of the antibody. Although the global market for this antibody may be small relative to other “neglected” diseases, we would argue that the global persistence of diphtheria despite high vaccination rates and dwindling supply of DAT signifies an ongoing need for a safe and effective alternative, such as 315C4.
ACKNOWLEDGMENTS
We thank Katherine Baptista, Rachel Wollacott, Jennifer Brennan, Elisabeth Boucher, and Colleen Fenn for technical assistance. We also thank Deborah Molrine for thoughtful scientific discussions and manuscript review.
FOOTNOTES
- Received 11 June 2013.
- Returned for modification 1 July 2013.
- Accepted 7 August 2013.
- Accepted manuscript posted online 12 August 2013.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.00462-13.
- Copyright © 2013, American Society for Microbiology. All Rights Reserved.
