Protein A-Specific Monoclonal Antibodies and Prevention ofStaphylococcus aureus Disease in Mice

Hwan Keun Kim,a Carla Emolo,a Andrea C. DeDent,a Fabiana Falugi,a,b Dominique M. Missiakas,a and Olaf Schneewinda

Department of Microbiology, University of Chicago, Chicago, Illinois, USA,a and Novartis Vaccines and Diagnostics, Siena, Italyb

Staphylococcus aureus is a leading cause of human soft tissue infections and bacterial sepsis. The emergence of antibiotic-resis-tant strains (methicillin-resistant S. aureus [MRSA]) has prompted research into staphylococcal vaccines and preventive mea-sures. The envelope of S. aureus is decorated with staphylococcal protein A (SpA), which captures the Fc� portion of immuno-globulins to prevent opsonophagocytosis and associates with the Fab portion of VH3-type B cell receptors to trigger B cellsuperantigen activity. Nontoxigenic protein A (SpAKKAA), when used as an immunogen in mice, stimulates humoral immuneresponses that neutralize the Fc� and the VH3� Fab binding activities of SpA and provide protection from staphylococcal abscessformation in mice. Here, we isolated monoclonal antibodies (MAbs) against SpAKKAA that, by binding to the triple-helical bun-dle fold of its immunoglobulin binding domains (IgBDs), neutralize the Fc� and Fab binding activities of SpA. SpAKKAA MAbspromoted opsonophagocytic killing of MRSA in mouse and human blood, provided protection from abscess formation, andstimulated pathogen-specific immune responses in a mouse model of staphylococcal disease. Thus, SpAKKAA MAbs may be use-ful for the prevention and therapy of staphylococcal disease in humans.

Staphylococcus aureus, a Gram-positive pathogen that colonizesthe skin and nares of humans, causes purulent abscess lesions

in soft tissues and invasive disease with bacteremia, sepsis, or en-docarditis (31). Because of the emergence of multidrug-resistantstrains, methicillin-resistant S. aureus (MRSA), the therapy ofboth hospital- and community-acquired infections is challenging(26). S. aureus infection in humans does not lead to protectiveimmunity, and up to 30% of skin and soft tissue infections reoccureven with antibiotic or surgical therapies (31). Staphylococcalprotein A (SpA) is a key virulence factor that enables S. aureus toevade innate and adaptive immune responses (8, 24). SpA binds tothe Fc� portion of human and animal immunoglobulins, a de-fense mechanism that provides S. aureus with protection fromopsonophagocytic killing (12). Further, SpA associates with theFab portion of VH3-type IgM; the cross-linking of B cell receptorsleads to the activation and clonal expansion of B cells and theirsubsequent apoptotic collapse, a mechanism that suppressesadaptive immune responses during staphylococcal infection (14).

In a mouse intravenous challenge model, S. aureus mutantslacking spa are phagocytosed and killed more rapidly in bloodthan wild-type staphylococci and are also defective in the forma-tion of abscess lesions (24). In contrast to wild-type strains, miceinfected with spa mutants elicit humoral immune responsesagainst several different pathogen-specific products, suggestingthat both antiphagocytic and B cell superantigen activities of SpAplay important roles during the pathogenesis of S. aureus infec-tions in mice (35). SpA is secreted as a precursor with an N-ter-minal signal peptide and a C-terminal LPXTG sorting signal (39).Following signal peptide cleavage and sortase-mediated anchor-ing of the C terminus in the cell wall, five 56- to 61-residue immu-noglobulin binding domains (IgBD) of SpA are displayed on thebacterial surface (32, 42). Each IgBD folds into a triple-helicalbundle with discrete binding sites for Fc� and VH3 Fab (16). Glu-tamine (Q) residues 9 and 10 in helix I of each IgBD are critical forFc� binding, while aspartic acid (D) 36 and 37 in the linker be-tween helix II and III are essential for IgBD association with VH3Fab (22). SpAKKAA is a nontoxigenic protein A molecule with

lysine (K) substitutions at Q9 and Q10 and alanine (A) substitu-tions at D36 and D37 in each of the five IgBDs (22). In contrast towild-type SpA, immunization of mice with purified recombinantSpAKKAA elicits polyclonal antibody responses that neutralize theFc� and VH3 Fab binding activities of staphylococci, promote theopsonophagocytic killing of staphylococci in mouse blood, pro-vide for protection against staphylococcal abscess formation, andenable the development of humoral immune responses againstmultiple secreted antigens of S. aureus strains (22).

Previous work suggested that SpAKKAA may be useful as a vac-cine to prevent S. aureus disease in humans (22). If so, the molec-ular mechanisms of SpAKKAA-mediated staphylococcal diseaseprotection may be defined through the development of monoclo-nal antibodies (MAbs). For example, SpAKKAA-specific MAbscould be used to develop in vitro assays that represent correlates ofdisease protection through the neutralization of SpA and theproof of principle that specific antibody molecules are sufficientto mediate protection from staphylococcal disease. Further,SpAKKAA-specific MAbs could also be exploited as a preventiveregimen or for therapeutic immunization to develop S. aureus-specific adaptive responses in infected individuals whose immunesystems are being diverted via SpA. To pursue these goals, weisolated and characterized SpA-specific MAbs.

Received 3 March 2012 Returned for modification 21 March 2012Accepted 18 July 2012

Published ahead of print 23 July 2012

Editor: J. N. Weiser

Address correspondence to Olaf Schneewind, oschnee@bsd.uchicago.edu.

Supplemental material for this article may be found at http://iai.asm.org/.

Copyright © 2012, American Society for Microbiology. All Rights Reserved.

doi:10.1128/IAI.00230-12

The authors have paid a fee to allow immediate free access to this article.

3460 iai.asm.org Infection and Immunity p. 3460–3470 October 2012 Volume 80 Number 10

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

MATERIALS AND METHODSBacterial strains and growth conditions. S. aureus strains Newman,USA300 (LAC), and USA400 (MW2) were grown in tryptic soy broth(TSB) at 37°C. Escherichia coli strains DH5� and BL21(DE3) were grownin Luria-Bertani (LB) broth with 100 �g · ml�1 ampicillin at 37°C.

Monoclonal antibodies. Mouse monoclonal antibodies were gener-ated by the conventional method (27). Briefly, BALB/c mice (8 weeks old;female; Jackson Laboratory) were immunized by intraperitoneal injectionwith 100 �g purified SpAKKAA emulsified 1:1 with complete Freund’sadjuvant (CFA; Difco). On days 21 and 42, mice were boosted by intra-peritoneal injection with 100 �g of the same antigen emulsified 1:1 withincomplete Freund’s adjuvant (IFA; Difco). On days 31 and 52, mice werebled and serum samples screened by enzyme-linked immunosorbent as-say (ELISA) for specific antibodies. Seventy-nine days following initialimmunization, mice that demonstrated strong antigen immunoreactivityby ELISA were boosted with 25 �g of the same antigen. Three days later,splenocytes were harvested and fused with the mouse myeloma cell lineSP2/mIL-6, an interleukin 6-secreting derivative of the SP2/0 myelomacell line. Supernatants from resulting hybridomas were screened byELISA, and antigen-specific clones were further subcloned by limitingdilution to yield monoclonal antibody-secreting hybridomas arising fromsingle cells. Antibodies were purified from the spent culture supernatantof cell lines. Spa27 monoclonal antibody was purchased from Sigma.

Purification of recombinant proteins. Polypeptides derived from theamino acid sequence of the SpA-EKKAA domain were synthesized by CPCScientific Inc. (Sunnyvale, CA). Lyophilized peptide samples were solubi-lized using either distilled water or dimethyl sulfoxide (DMSO) and thenaliquoted and frozen at �80°C. The use of plasmids for wild-type SpA andSpAKKAA has been previously described (22). Oligonucleotides for thesynthesis of SpAKK (Q9K and Q10K substitutions in each of the fiveIgBDs), SpAAA (D36A and D37A substitutions in each of the five IgBDs),and individual IgBDs (E, D, A, B, and C) of SpAKKAA and Sbi1-4/KKAA

(amino acids ranging from 33 to 255 with substitutions at Q51K, Q52K,Q103K, Q104K, R231A, N238A) were synthesized by Integrated DNA Tech-nologies, Inc. PCR products of SpAKKAA variants were cloned into thepET15b vector generating N-terminal His6-tagged recombinant proteins.The coding sequence of Sbi1-4 was PCR amplified with two primers, 5=-AAAAAAGCTAGCTGGTCTCATCCTCAATTTGAGAAGACGCAACAAACTTCAACTAAG-3= and 5=-AAAAAACTCGAGTTTCCAGAATGATAATAAATTAC-3=, from S. aureus Newman chromosomal DNA with en-gineered N-terminal Strep tag (WSHPQFEK). PCR products of Sbi1-4 andSbi1-4/KKAA were cloned into pET24b vector generating C-terminal His6-tagged recombinant protein with engineered N-terminal Strep tag (WSHPQFEK). All plasmids were transformed into BL21(DE3) for affinity pu-rification. Overnight cultures of recombinant E. coli strains were diluted1:100 into fresh medium and grown at 37°C to an A600 of 0.5, at whichpoint cultures were induced with 1 mM isopropyl �-D-1-thiogalatopyra-noside (IPTG) and grown for an additional 3 h. Bacterial cells were sedi-mented by centrifugation, suspended in column buffer (50 mM Tris-HCl[pH 7.5], 150 mM NaCl), and disrupted with a French pressure cell at14,000 lb/in2. Lysates were cleared of membrane and insoluble compo-nents by ultracentrifugation at 40,000 � g. Proteins in the cleared lysatewere subjected to nickel-nitrilotriacetic acid (Ni-NTA) affinity chroma-tography. Proteins were eluted in column buffer containing successivelyhigher concentrations of imidazole (100 to 500 mM). Protein concentra-tions were determined via the bicinchoninic acid (BCA) assay (ThermoScientific).

Enzyme-linked immunosorbent assay. To determine SpA-specificserum IgG, affinity-purified SpAKKAA was used to coat ELISA plates(Nunc Maxisorp) at 1 �g · ml�1 in 0.1 M carbonate buffer (pH 9.5 at 4°C)overnight. The following day, plates were blocked and incubated withdilutions of hyperimmune sera and developed using OptEIA reagent (BDBiosciences). For the determination of binding affinity of SpA-specificMAbs, ELISA plates were coated with affinity-purified individual immu-noglobulin binding domains or synthetic peptides (H1, H2, H3, H1 and

H3, and H2 and H3) whose sequences were derived from the sequence ofSpA-EKKAA. Peptides were used for plate coating at a concentration of 100nM in 0.1 M carbonate buffer (pH 9.5 at 4°C) overnight. On the followingday, plates were blocked and incubated with variable concentrations ofSpA-specific MAbs. To determine the avidity of specific MAbs, antibody-antigen interactions were perturbed with increasing concentration (0 to 4M) of ammonium thiocyanate. For SpA and Sbi binding assays, affinity-purified SpA and Sbi variants were coated onto ELISA plates at 1 �g · ml�1

in 0.1 M carbonate buffer (pH 9.5 at 4°C) overnight. The following day,plates were blocked and incubated with dilutions of peroxidase-conju-gated human IgG, Fc, and F(ab)2 (Jackson Laboratory) or dilutions ofisotype control antibodies and SpAKKAA-specific MAbs; assays were de-veloped using OptEIA reagent. To measure the inhibition of nonimmuneassociation between human IgG and SpA, plates were incubated witheither 20 �g · ml�1 isotype control antibodies or SpAKKAA-specific MAbsprior to ligand binding. For competition assay, plates were coated with 10ng · ml�1 SpAKKAA in 0.1 M carbonate buffer (pH 9.5) at 4°C overnight.The following day, plates were blocked and incubated with 30 �g · ml�1 ofisotype control antibodies or SpAKKAA-specific MAbs prior to the incu-bation with horseradish peroxidase (HRP)-conjugated SpA-specificMAbs (Innova Biosciences) at a final concentration of 100 ng · ml�1.

Mouse renal abscess model. Affinity-purified antibodies in phos-phate-buffered saline (PBS) were injected at a concentration of 5, 15, 20,or 50 mg · kg�1 of experimental animal weight into the peritoneal cavity ofBALB/c mice (6 weeks old; female; Charles River Laboratories) 4 to 24 hprior to challenge with S. aureus. Overnight cultures of S. aureus strainswere diluted 1:100 into fresh TSB and grown for 2 h at 37°C. Staphylococciwere sedimented, washed, and suspended in PBS to the desired bacterialconcentration. Inocula were quantified by spreading sample aliquots ontryptic soy agar (TSA) and enumerating the colonies that formed uponincubation. BALB/c mice were anesthetized via intraperitoneal injectionwith 100 mg · ml�1 ketamine and 20 mg · ml�1 xylazine per kilogram ofbody weight. Mice were infected by injection with 1 � 107 CFU of S.aureus Newman or 5 � 106 CFU of S. aureus USA300 (LAC) or USA400(MW2) into the periorbital venous sinus of the right eye. On day 4 or 15following challenge, mice were killed by CO2 inhalation. Both kidneyswere removed, and the staphylococcal load in one organ was analyzed byhomogenizing renal tissue with PBS, 0.1% Triton X-100. Serial dilutionsof homogenate were spread on TSA and incubated for colony formation.The remaining organ was examined by histopathology. Briefly, kidneyswere fixed in 10% formalin for 24 h at room temperature. Tissues wereembedded in paraffin, thin sectioned, stained with hematoxylin-eosin,and inspected by light microscopy to enumerate abscess lesions. Immuneserum samples collected at 15 days postinfection were examined by im-munoblotting against 14 affinity-purified staphylococcal antigens immo-bilized onto a nitrocellulose membrane at 2 �g. Signal intensities werequantified as previously described (23). All mouse experiments were per-formed in accordance with the institutional guidelines following experi-mental protocol review and approval by the Institutional Biosafety Com-mittee (IBC) and the Institutional Animal Care and Use Committee(IACUC) at the University of Chicago.

Staphylococcal survival in blood. Whole blood was collected fromBALB/c mice by cardiac puncture, and coagulation was inhibited with 10�g · ml�1 lepirudin. Fifty microliters of a suspension with 5 � 105 CFU S.aureus Newman was mixed with 950 �l of mouse blood in the presence of2 �g · ml�1 of MAbs. Samples were incubated at 37°C with slow rotationfor 30 min and then incubated on ice with a final concentration of 0.5%saponin-PBS. For human blood studies, 50 �l of 5 � 106 CFU S. aureusUSA400 (MW2) was mixed with 950 �l of freshly drawn human blood inthe presence of 10 �g · ml�1 of MAbs. The tubes were incubated at 37°Cwith slow rotation for 120 min. Aliquots were incubated on ice with a finalconcentration of 0.5% saponin-PBS to lyse blood cells. Dilutions of staph-ylococci were plated on agar for colony formation. Experiments withblood from human volunteers were performed with protocols that had

Protein A-Specific Monoclonal Antibodies

October 2012 Volume 80 Number 10 iai.asm.org 3461

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

been reviewed, approved, and supervised by the University of Chicago’sInstitutional Review Board (IRB).

SpA-specific serum IgG. BALB/c mice were injected into the perito-neum with 20 �g affinity-purified SpA variants in the presence of 85 �gMAb 3F6 or its isotype control at days 0 and 11. At day 21, whole bloodwas collected from BALB/c mice to obtain hyperimmune sera.

Measuring the abundance of SpA in circulation. Passively immu-nized BALB/c mice were injected into the peritoneum with 200 �g affin-ity-purified wild-type SpA. At the indicated time intervals, whole bloodwas collected from BALB/c mice with 10 �g · ml�1of lepirudin anticoag-ulant. All samples were kept on ice with a final concentration of 0.5%saponin-PBS. Lysed samples were then diluted in 1:10 PBS and mixedwith SDS-PAGE sample buffer at a 1:1 dilution. Samples were boiled for 5min at 90°C prior to SDS-PAGE gel electrophoresis. Samples were trans-ferred to polyvinylidene difluoride (PVDF) and analyzed by immunoblot-ting with affinity-purified rabbit �-SpAKKAA antibody.

Sbi consumption assay. Overnight cultures of S. aureus Newmanwere diluted 1:100 into fresh TSB, grown for 2 h, and A600 adjusted to 0.4(1 � 108 CFU · ml�1) with prechilled TSB. Cells were washed and incu-bated with either 100 �l of isotype control or MAb 3F6 at a final concen-tration of 100 �g · ml�1 for an hour at 4°C. Following incubation, staph-ylococci were washed with prechilled TSB and incubated with 2 �g ofaffinity-purified wild-type Sbi1-4 for 1 h at 4°C. Staphylococci were sedi-mented by centrifugation at 13,000 � g for 1 min, and supernatants wereremoved and mixed with sample buffer (1:1). Samples were boiled for 5min at 90°C prior to SDS-PAGE gel electrophoresis. Samples were elec-trotransferred to a PVDF membrane and analyzed by immunoblottingwith affinity-purified rabbit �-SpAKKAA antibody.

Sequencing of monoclonal antibodies. Total RNA samples from hy-bridoma cells were isolated using a standardized protocol. Briefly, 1.4 �107 hybridoma cells cultured in DMEM-10 medium with 10% fetal bo-vine serum (FBS) were washed with PBS, sedimented by centrifugation,and lysed in TRIzol (Invitrogen). Samples were mixed with 20% chloro-form and incubated at room temperature for 3 min and centrifuged at10,000 � g for 15 min at 4°C. RNAs in the aqueous layer were removedand washed with 70% isopropanol. RNA was sedimented by centrifuga-tion and washed with 75% diethylpyrocarbonate (DEPC)-ethanol. Pelletswere dried and RNA dissolved in DEPC. cDNA was synthesized with thecDNA synthesis kit (Novagen) and PCR amplified using the PCR reagentsystem (Stratagene), independent primers (5 pmol each), and a mousevariable heavy and light chain-specific primer set (Novagen). PCR prod-ucts were sequenced and analyzed using IMGT/V-QUEST (http://www.imgt.org/IMGT_vquest/share/textes/).

Statistical analysis. Bacterial loads and number of abscesses in theexperimental animal model for S. aureus infection were analyzed with thetwo-tailed Mann-Whitney test to measure statistical significance. Un-paired two-tailed Student’s t tests were performed to analyze the statisticalsignificance of ELISA data, immunoblotting signals, and ex vivo bloodsurvival data. All data were analyzed by Prism (GraphPad Software, Inc.),and P values less than 0.05 were deemed significant.

RESULTSSpAKKAA MAbs protect mice against staphylococcal disease.BALB/c mice were immunized with purified SpAKKAA using aprime-booster regimen, and antigen-specific IgG responses werequantified by ELISA. Animals were euthanized, and their spleno-cytes were fused with myeloma cells. The resulting hybridomaswere screened for the production of antigen-specific MAbs. Ini-tially, protein A-specific MAbs were screened using the functionalassays as well as the murine infection model (see Tables S1 and S2in the supplemental material). After the initial screen, we selectedthree MAbs (5A10, 3F6, and 3D11) for further characterization, asthese antibodies displayed the best immune protection in eachisotype group (Table 1; see also Table S1). BALB/c mice were

immunized with affinity-purified MAbs (5 mg · kg�1 bodyweight) and challenged by injecting 1 � 107 CFU S. aureus New-man, a methicillin-sensitive clinical isolate (MSSA) (2), into theperiorbital venous sinus of the right eye. The ability of staphylo-cocci to seed abscesses in renal tissues was examined by histopa-thology 4 days after challenge (Table 1). In homogenized renaltissues of control mice (immunized with 5 mg · kg�1 isotype con-trol MAbs), average staphylococcal loads of 5.02 log10 CFU · g�1

(IgG1), 4.64 log10 CFU · g�1 (IgG2a), and 5.24 log10 CFU · g�1

(IgG2b) were recovered (Table 1). Compared to isotype MAb-treated controls, animals that received protein A-specific MAbsdisplayed a reduction in staphylococcal load (2.80 log10 CFU · g�1

[5A10], 2.28 log10 CFU · g�1 [3F6], and 2.72 log10 CFU · g�1

[3D11]) as well as abscess formation (Table 1). Of note, not allSpAKKAA MAbs generated protection against staphylococcal dis-ease (see Table S1 in the supplemental material) even though theseantibodies bound with appreciable affinity to their antigen (see,for example, 3A6 and 6D11 in Table S2 in the supplemental ma-terial).

SpAKKAA MAbs protect mice against MRSA challenge. Co-horts of BALB/c mice were immunized with MAbs 5A10, 3F6, and3D11 (5 mg · kg�1) or a combination of all three MAbs (15 mg ·kg�1) and challenged with strain USA400 (MW2), a highly viru-lent community-acquired MRSA isolate (3). Compared to isotypeMAb-treated controls, animals that received any one of the threeMAbs (5A10, 3F6, 3D11) harbored a reduced bacterial load andfewer staphylococcal abscesses in renal tissues (Table 2). Animalsthat had been immunized with a mixture of all three MAbs (15mg · kg�1) displayed an even greater reduction in staphylococcalload (2.03 log10 CFU · g�1 reduction; P � 0.0002) and in abscessformation (vaccine versus mock, P � 0.0004). It is likely that

TABLE 1 Immunization with SpAKKAA MAbs protects mice against S.aureus disease

MAba

Staphylococcal load and abscess formation in renal tissue

log10 CFUg�1b

Pvaluec Reductiond

Number ofabscessese

Pvaluec

IgG1Mock 5.02 � 0.66 2.00 � 0.945A10 2.22 � 0.22 0.0019 2.80 0.00 � 0.00 0.0350

IgG2a

Mock 4.64 � 0.49 3.70 � 1.403F6 2.36 � 0.36 0.0010 2.28 0.60 � 0.50 0.0239

IgG2b

Mock 5.24 � 0.51 3.00 � 0.673D11 2.52 � 0.40 0.0010 2.72 0.56 � 0.28 0.0068

a Affinity-purified antibodies were injected into the peritoneal cavity of BALB/c mice ata concentration of 5 mg · kg�1 4 h prior to intravenous challenge with 1 � 107 CFU S.aureus Newman.b Means (�SEM) of staphylococcal load calculated as log10 CFU · g�1 in homogenizedrenal tissues 4 days following infection in cohorts of 10 BALB/c mice per immunizationwith limit of detection at 1.99 log10 CFU · g�1. A representative of three independentand reproducible animal experiments is shown.c Statistical significance was calculated with the two-tailed Mann-Whitney test, and Pvalues were recorded.d Reduction in bacterial load calculated as log10 CFU · g�1.e Histopathology of hematoxylin-eosin-stained, thin-sectioned kidneys from 10animals; the number of abscesses per kidney was recorded and averaged for the finalmean (�SEM).

Kim et al.

3462 iai.asm.org Infection and Immunity

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

enhanced protection is due to administration of increased con-centration of MAbs (15 mg · kg�1 versus 5 mg · kg�1). We arrivedat this hypothesis because the three antibodies, although recogniz-ing similar structural features, do not appear to occupy identicalbinding sites on SpA (see below). Further, increasing the concen-tration of only one of the three MAbs (3F6) caused the same effect:increased protection against staphylococcal disease (see below).

In addition to providing immediate protection against staph-ylococcal challenge, SpAKKAA-specific MAbs may also neutralizethe B-cell superantigen activity of SpA (14), thereby enabling in-fected hosts to generate antibody responses against many differentstaphylococcal antigens (22). To examine this possibility, BALB/cmice were passively immunized with MAb 3F6 or its IgG2a isotypecontrol (20 mg · kg�1) prior to intravenous challenge with S. au-reus USA400 (MW2). Fifteen days after challenge, animals wereeuthanized and staphylococcal load in organ tissue was examined(Fig. 1A). Mice that had been immunized with MAb 3F6 harboreda reduced staphylococcal load (4.77 log10 CFU · g�1 reduction,P 0.0013) as well as a reduced number of abscesses (from 10.14[�2.08] [IgG2a] to 3.00 [�1.00] [3F6], P 0.0065; Fig. 1A).Blood samples withdrawn 15 days postchallenge were examinedfor serum IgG reactive against 14 staphylococcal antigens underconsideration as protective antigens for vaccine development:Coa, ClfA, ClfB, EsxA, EsxB, FnBPA, FnBPB, Hla, IsdA, IsdB,LukD, SdrD, SpAKKAA, and vWbp (9). As observed previouslywith animals that had been actively vaccinated with SpAKKAA,mice that had been passively immunized with MAb 3F6 developedhigher serum IgG titers against several different staphylococcal

antigens (Fig. 1B) (22). In particular, IgG levels against Coa, ClfA,EsxA, EsxB, FnBPB, Hla, IsdA, LukD, SdrD, and vWbp were in-creased in serum samples of MAb 3F6-immunized animals com-pared to those in the control cohort. Nevertheless, serum IgGagainst IsdB, the staphylococcal hemoglobin hemophore (33),was not increased (Fig. 1B). Of note, the IgG titer against SpAKKAA

was sustained over 15 days following passive transfer of MAb 3F6(Fig. 1B).

During staphylococcal infection, recognition of soluble SpA byMAb 3F6 is expected to form immune complexes (IC) that arethen phagocytosed by immune cells. Phagocytosed SpA is thenprocessed by proteolytic enzymes in the phagolysosome, and pep-tide fragments are presented to T and B cells to produce polyclonalantibodies. As a confirmatory test, cohorts of animals received amixture of affinity-purified recombinant protein A variants (SpA,SpAKK, SpAAA, SpAKKAA, and mock [PBS]) in the presence ofMAb 3F6 or its isotype control at days 0 and 11. At day 21, animalswere euthanized and their ability to elicit different classes of SpA-specific antibody was measured by ELISA. All animals failed togenerate SpA-specific antibody responses without MAb treatment(see Fig. S1 in the supplemental material). In addition, animalsthat received B cell superantigens (SpA and SpAKK) failed to gen-erate SpA-specific IgG1 and IgG2a antibodies even in the presenceof MAb 3F6 (see Fig. S1). However, mice treated with SpA variantslacking B cell superantigen activity (SpAAA and SpAKKAA) wereable to generate a significant amount of IgG1 (see Fig. S1). Al-though the estimated amount of soluble protein A during infec-tion (5 to 10 ng per 107 CFU) is well below the dose of affinity-purified protein A in our animal experiments, the data in Fig. S1suggest a potential role of SpA-specific T/B cells in neutralizing Bcell superantigen activity. Taken together, we presume that active

FIG 1 SpAKKAA-specific monoclonal antibodies (MAbs) protect mice againstMRSA infection. Cohorts of animals (n 10) were passively immunized byintraperitoneal injection with either isotype control (IgG2a) or SpAKKAA MAb(3F6) at 20 mg · kg�1. After 24 h, animals were challenged with 5 � 106 CFU ofS. aureus MW2. (A) At 15 days postchallenge, animals were euthanized toenumerate the staphylococcal load in kidneys. (B) Serum samples of miceinfected for 15 days were analyzed for antibodies against the staphylococcalantigen matrix. ClfA, clumping factor A; ClfB, clumping factor B; FnBPA,fibronectin binding protein A; FnBPB, fibronectin binding protein B; IsdA,iron surface determinant A; IsdB, iron surface determinant B; SdrD, serine-aspartic acid repeat protein D; SpAKKAA, nontoxigenic staphylococcal proteinA; Coa, coagulase; EsxA, Ess (early secreted antigen target 6 kDa [ESAT-6]secretion system) extracellular A; EsxB, Ess (ESAT-6 secretion system) extra-cellular B; Hla, alpha-hemolysin; LukD, leukocidin D; vWbp, von Willebrandbinding protein. The values represent the fold increases of samples from MAb3F6-treated animals over the isotype control animal serum samples (n 7 forIgG2a, n 8 for 3F6). Data are the means, and error bars represent � standarderrors of the means (SEM). Results in panels A and B are representative of twoindependent analyses.

TABLE 2 Immunization with SpAKKAA MAbs protects mice againstMRSA challenge

Antibodya

Staphylococcal load and abscess formation in renaltissue

log10 CFUg�1b

Pvaluec Reductiond

No. ofabscessese

Pvaluec

IgG1Mock 7.42 � 0.20 22.3 � 6.35A10 6.00 � 0.21 0.0009 1.42 10.2 � 2.5 0.0482

IgG2a

Mock 7.15 � 0.18 11.8 � 2.03F6 5.80 � 0.21 0.0009 1.35 6.4 � 0.7 0.0323

IgG2b

Mock 7.13 � 0.11 14.0 � 1.83D11 5.81 � 0.25 0.0006 1.32 7.7 � 1.9 0.0489

IgG1�IgG2a�IgG2b

Mock 7.75 � 0.06 17.4 � 1.75A10/3F6/3D11 5.72 � 0.12 0.0002 2.03 6.7 � 0.6 0.0004

a Affinity-purified antibodies were injected into the peritoneal cavity of BALB/c mice ata concentration of individual antibody at 5 mg · kg�1 or combinations of threemonoclonal antibodies at 15 mg · kg�1 24 h prior to intravenous challenge with 1 � 107

CFU S. aureus MW2.b Means (�SEM) of staphylococcal load calculated as log10 CFU · g�1 in homogenizedrenal tissues 4 days following infection in cohorts of 10 BALB/c mice per immunizationwith limit of detection at 1.99 log10 CFU · g�1. A representative of two independent andreproducible animal experiments is shown.c Statistical significance was calculated with the two-tailed Mann-Whitney test, and Pvalues were recorded.d Reduction in bacterial load calculated as log10 CFU · g�1.e Histopathology of hematoxylin-eosin-stained, thin-sectioned kidneys from 10animals; the number of abscesses per kidney was recorded and averaged for the finalmean (�SEM).

Protein A-Specific Monoclonal Antibodies

October 2012 Volume 80 Number 10 iai.asm.org 3463

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

vaccination with SpAKKAA (22), but not passive immunization ofS. aureus-infected mice with neutralizing MAbs, can raise a signif-icant level of protein A-specific antibodies.

MAb Spa27 does not recognize SpAKKAA and fails to elicitprotective immunity in mice. Spa27 is a commercially availableprotein A-specific monoclonal antibody (Sigma) that has beenused over the past 2 decades for the detection of staphylococcalprotein A (36). The Spa27 hybridoma was generated from micethat had been immunized with wild-type staphylococcal protein Apurified from S. aureus strain Cowan I (43). Previous work dem-onstrated that wild-type protein A triggers the clonal expansionand collapse of B cell populations (13, 14), thereby ablating pro-tein A-specific immune response in mice (15), and that wild-typeprotein A encompasses binding sites for both Fc� and Fab VH3(16, 46). We therefore wondered whether Spa27 recognizes wild-type protein A as an antigen. To address this question, we usedELISA with purified recombinant protein A (SpA) or its variantsthat lack either the ability to specifically bind Fc� (SpAKK), the Fabdomain of VH3 (SpAAA), or both (SpAKKAA) (see Fig. S2 in thesupplemental material). The data revealed strong binding ofSpa27 to wild-type SpA and SpAKK but not to SpAAA or SpAKKAA

(Fig. 2B). Spa27 is a mouse IgG1 isotype antibody, which explainsits inability to bind protein A via Fc� (28). The weak associationbetween Spa27 and SpAAA or SpAKKAA could be due to the seem-ingly remote possibility that SpA27 requires residues D36/D37 ineach of the five IgBDs for antigen recognition or, more likely to us,that Spa27 binds SpA via its Fab domain, assuming the antibodybelongs to the VH3 or a related class of antibody.

We examined the biological function of Spa27 by injecting forpairwise comparison MAbs 3F6 or Spa27 (5 mg · kg�1) into theperitoneal cavity of BALB/c mice. These animals were then chal-lenged with S. aureus USA300 (LAC), the highly virulent commu-nity-acquired MRSA strain epidemic in the United States (10). At4 days postchallenge, animals were euthanized, and the bacterialloads in the kidneys of infected animals were determined (Fig. 2B).Compared to mock (PBS) treatment, animals that received MAb3F6 harbored a reduced bacterial load (1.38 log10 CFU · g�1 re-duction, P 0.0011). In contrast to the protective immunity elic-ited by 3F6, MAb Spa27 failed to reduce the bacterial load in kid-

neys of infected animals (0.20 log10 CFU · g�1 increase, P 0.2111). These data reveal that MAb Spa27 does not provide pro-tection against staphylococcal disease. Further, the experimentswith Spa27 illustrate that immunization of mice with wild-typeprotein A may not elicit monoclonal antibodies that can neutral-ize the immune-modulatory attributes of protein A by bindingthis molecule as an antigen.

Recognition of SpAKKAA by MAbs. Microtiter dishes werecoated with SpAKKAA, and ELISA was used to determine the affin-ity constant (Ka [MAb · Ag]/[MAb] � [Ag]) of purified MAbs.MAb 3F6 displayed the highest affinity (Ka of 22.97 � 109 M�1)followed by MAb 5A10 (Ka of 8.47 � 109 M�1) and MAb 3D11 (Ka

of 3.93 � 109 M�1) (Table 3). Each of the five IgBDs alone (EKKAA,DKKAA, AKKAA, BKKAA, and CKKAA) or peptides encompassing he-lix 1, 2, or 3 as well as helices 1�2 and 2�3 of the IgBD EKKAA

domain were examined for antibody binding (Table 3). MAbs5A10 and 3F6 bound all five IgBDs with the same affinity asSpAKKAA. MAb 5A10 did not bind to the helical peptides,whereas MAb 3F6 displayed weak affinity for the helix 1�2peptide. MAb 3D11 bound to BKKAA and CKKAA and weakly toAKKAA but not to EKKAA and DKKAA. In sum, SpAKKAA MAbsthat afforded the highest levels of protection against staphylo-coccal disease in mice bound some or all of the five IgBDs butnot the peptides encompassing only one or two of three helicesof IgBDs. These data suggest that protective MAbs recognizeconformational epitopes of the triple-helical bundle for eachIgBD.

To examine whether the avidities of MAbs play a significantrole in immune protection, ELISA was performed in the pres-ence of increasing concentrations of the chaotropic reagentammonium thiocyanate (Fig. 3). The measured avidity of MAb3F6 was significantly higher than that of MAb 5A10 and 3D11(Fig. 3). Of note, 3D11 displayed relatively low avidity, whichmay be due to its specific interaction with only two of the fiveIgBDs (Fig. 3 and Table 3). From this, we conclude that theavidities of MAbs may not be a major determinant of theirimmune protection in mice.

MAb 3F6 binds Sbi. Sbi, a secreted protein of S. aureus, iscomprised of five distinct domains (49). Two N-terminal domains

FIG 2 SpA monoclonal antibody (Spa27) fails to elicit protective immunity in mice. (A) ELISA examining the association of SpA MAb (Spa27) and SpAKKAA

MAb (3F6) with immobilized wild-type protein A (SpA) and variants lacking immunoglobulin binding via Fc� (SpAKK), Fab (SpAAA), or Fc� and Fab (SpAKKAA)(n 3). (B) Cohorts of animals (n 9 to 15) were passively immunized by intraperitoneal injection with either mock (PBS), Spa27 at 5 mg · kg�1, or 3F6 at 5or 50 mg · kg�1. Twenty-four hours postimmunization, animals were challenged with 5 � 106 CFU of S. aureus USA300. Four days postchallenge, animals wereeuthanized to enumerate the staphylococcal load in kidneys.

Kim et al.

3464 iai.asm.org Infection and Immunity

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

(1 and 2) are homologous to the IgBDs of SpA (48). Domains 3and 4 associate with complement components C3 and factor H,and the C-terminal domain has been proposed to retain somesecreted Sbi molecules in the staphylococcal envelope by bindingto lipoteichoic acids (6, 44). Domains 1 and 2 bind to the Fc�portion of immunoglobulins (1); this activity, in concert with theC3 and factor H binding attributes of domains 3 and 4, promotesthe futile consumption of fluid complement components (17). Sbidoes not seem to exert B cell superantigen activity, as its two IgBDs(domains 1 and 2) lack the canonical two aspartic acid residues atpositions 36 and 37 (16, 24). His-Sbi1-4, a recombinant proteinencompassing both IgBDs and the complement binding domains,retained human IgG in an affinity chromatography experiment(Fig. 4A). His-Sbi1-4/KKAA is a variant with lysine (K) substitutionsof conserved glutamine residues (Q51,52 and Q103,104) in domains 1and 2, i.e., the predicted Fc� binding sites of the Sbi IgBDs, andalanine (A) substitutions of arginine (R231) and aspartic acid(D238) residues of the complement binding domain (17). His-Sbi1-4/KKAA did not retain human IgG during affinity chromatog-raphy (Fig. 4A). When examined by ELISA, His-Sbi1-4 bound tomouse as well as human IgG and to both the Fc and Fab domainsof human IgG, whereas His-Sbi1-4/KKAA did not (see Fig. S3 in thesupplemental material). MAbs 5A10 and 3D11 did not bind toHis-Sbi1-4/KKAA; however, 3F6 bound to the protein (Fig. 4B).Thus, MAb 3F6 may neutralize Sbi or remove secreted Sbi fromcirculation, thereby preventing the consumption of complementfactor C3 by staphylococci.

Binding site competition experiments with SpAKKAA MAbs.ELISA studies revealed that the three MAbs, 5A10, 3F6, and 3D11,bound with similar affinities to wild-type SpA (Fig. 5A). Com-pared to 5A10, the IgG1 control antibodies displayed little affinityfor SpA. Further, the affinity of the IgG2b control antibody wasreduced compared to that of MAb 3D11. Compared to 3F6, theIgG2a control antibody bound SpA with slightly reduced affinity.In a competitive ELISA with horseradish peroxidase-conjugatedMAbs (5A10-HRP, 3F6-HRP, and 3D11-HRP), isotype controlantibodies did not interfere with the binding of HRP-conjugatedMAbs to SpA (Fig. 5B). The addition of equimolar amounts ofeach MAb reduced the binding of the corresponding HRP conju-gate (Fig. 5B). MAb 3D11 did not prevent the association of HRP-5A10 or HRP-3F6 with SpA; however, MAbs 5A10 and 3F6 inter-fered with HRP-3D11 binding to SpA. MAb 3F6 caused somereduction in the binding of HRP-5A10 to SpA (Fig. 5B). Finally,MAb 5A10 was a weak competitor for the binding of 3F6-HRP toSpA (Fig. 5B). These data suggest that the binding sites for thethree MAbs on the surface of the triple-helical bundles of SpA maybe in close proximity to one another or even partially overlap(Table 3).

TABLE 3 Association constants for the binding of MAbs 5A10, 3F6, and 3D11 to SpAKKAA and its fragmentsa

MAb

Association constant (�109 M�1) for antigen or antigen fragment

IgG binding domains of protein A Segments of the EKKAA triple-helical bundle

SpAKKAA EKKAA DKKAA AKKAA BKKAA CKKAA H1 H2 H3 H1�2 H2�3

IgG1 5A10 8.47 9.40 8.19 8.08 7.03 10.12 � � � � �IgG2a 3F6 22.97 17.69 12.41 20.15 27.46 26.46 � 0.01 � 0.41 0.01IgG2b 3D11 3.93 � � 0.87 3.92 3.60 0.02 � � � �a Affinity-purified antibodies (1 mg · ml�1) were serially diluted across ELISA plates coated with antigens (100 nM) to calculate the association constant using Prism (GraphPadSoftware, Inc.). To study the binding of antibodies to protein A antigen, we used the SpAKKAA variant (residues 1 to 291 of mature SpA harboring six N-terminal histidyl residues)with four amino acid substitutions in each of the five immunoglobulin binding domains (IgBD) of protein A (E [residues 1 to 56], D [residues 57 to 117], A [residues 118 to 175], B[residues 176 to 233], and C [residues 234 to 291]). In each IgBD, the glutamines at positions 9 and 10 (amino acid residues from IgBD-E) were replaced with lysine (Q9K, Q10K),and aspartic acids 36 and 37 were substituted with alanine (D36A, D37A). The same substitutions were introduced into proteins spanning individual IgBDs: EKKAA, DKKAA, AKKAA,BKKAA, and CKKAA (all expressed and purified with an N-terminal six histidyl tag). SpAKKAA and individual IgBDs were purified by affinity chromatography from E. coli extracts.Peptides H1, H2, H3, H1�2, and H2�3 were synthesized on a peptide synthesizer and purified via high-performance liquid chromatography (HPLC). The peptides encompasshelices 1 (H1: NH2-AQHDEAKKNAFYQVLNMPNLNA-COOH), 2 (H2: NH2-NMPNLNADQRNGFIQSLKAAPSQ-COOH), 3 (H3: NH2-AAPSQSANVLGEAQKLNDSQAPK-COOH), 1�2 (H1�2; NH2-AQHDEAKKNAFYQVLNMPNLNADQRNGFIQSLKAAPSQ-COOH) or 2�3 (H2�3; NH2-NMPNLNADQRNGFIQSLKAAPSQSANVLGEAQKLNDSQAPK-COOH) of the triple helical bundle of the EKKAA IgBD (residues 1 to 56 of SpAKKAA; NH2-AQHDEAKKNAFYQVLNMPNLNADQRNGFIQSLKAAPSQSANVLGEAQKLNDSQAPK-COOH). The � symbol signifies measurements that were too low to permit the determination of the association constant.

FIG 3 Avidity of protein A-specific monoclonal antibodies. Monoclonal an-tibodies 5A10, 3F6, and 3D11 were incubated with increasing concentrations(0 to 4 M) of ammonium thiocyanate to perturb antigen-antibody-specificinteractions. Data are the means, and error bars represent � SEM. Results arerepresentative of three independent analyses.

FIG 4 SpAKKAA MAb 3F6 binds to Sbi (staphylococcal binder of immuno-globulin). (A) Coomassie blue-stained SDS-PAGE gel revealing the electro-phoretic mobility of human immunoglobulin (hIgG, lane 1), purified recom-binant Sbi encompassing its immunoglobulin binding domains (Sbi1-4, lane2), and variant Sbi, which cannot bind hIgG (Sbi1-4/KKAA, lane 3). Using affin-ity chromatography, His-tagged Sbi1-4 and Sbi1-4/KKAA were eluted fromnickel-nitriloacetic acid Sepharose without (�hIgG; lanes 1 to 3) or with(�hIgG; lanes 4 and 5) incubation with human immunoglobulin. (B) ELISAexamining the association of immobilized Sbi1-4/KKAA with protein A-specificMAbs (n 3). Data are the means, and error bars represent � SEM. Results inpanels A and B are representative of three independent analyses.

Protein A-Specific Monoclonal Antibodies

October 2012 Volume 80 Number 10 iai.asm.org 3465

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

SpAKKAA-MAbs prevent the association of immunoglobulinwith protein A. Mouse antibodies of clan VH3-related families(e.g., 7183, J606, and S107) bind SpA via their Fab portion,whereas those of other VH families (J558, Q52, Sm7, VH10,VH11, and VH12) do not (7). The amino acid sequence of thecomplementarity determining region (CDR) of SpAKKAA-specificMAbs was determined by sequencing cDNA derived from hybrid-oma transcripts. The data showed that MAb 5A10 belongs to theclan VH3 7183 family; its Fab domain likely displays affinity forSpA (Table 4). MAbs 3F6 and 3D11 are members of the VH10 andJ558 families, respectively (Table 4). Fab domains of these anti-body families are not known to associate with SpA.

Wild-type SpA and its variants SpAKKAA, SpAKK, and SpAAA

were purified and used for ELISA binding studies with humanIgG. As expected, SpA bound to IgG or its Fc� and F(ab)2 frag-ments, whereas SpAKKAA did not (see Fig. S2 in the supplementalmaterial). The SpAKK variant (harboring lysine substitutions at all10 glutamine residues) was impaired in its ability to bind Fc� butnot F(ab)2 fragments, whereas the SpAAA variant (harboring ala-nine substitutions at all 10 aspartic acid residues) bound to Fc�but not F(ab)2 (see Fig. S2). The binding of human IgG to wild-type SpA was blocked by all three MAbs (5A10, 3F6, and 3D11) ina manner that exceeded the competition of isotype control MAbs(Fig. 6A). All three MAbs interfered with the binding of humanIgG to SpAKK (Fab binding) or to SpAAA (Fc� binding) (Fig. 6A).Thus, SpAKKAA-specific MAbs prevent the nonimmune associa-tion of SpA with immunoglobulin. Based on these data, we pre-sume that protein A-specific MAbs interact with conformationalepitopes involving helix 2 of IgBDs, a structural element involvedin the Fc� and Fab interactions of SpA.

If MAb 3F6 binds wild-type SpA as an antigen on the staphy-lococcal surface, its Fc� domain should be available for recogni-tion by complement or Fc receptors on the surface of immunecells. To test this prediction, S. aureus was incubated with 3F6, itsisotype control, and affinity-purified Sbi1-4. Antibody-mediatedcosedimentation led to the depletion of soluble Sbi1-4 from thesupernatant, which was analyzed as a measure for the availabilityof Fc� sites on the bacterial surface. Incubation of staphylococci

FIG 5 SpAKKAA-specific MAbs bind wild-type protein A. (A) ELISA examining the binding of immobilized wild-type protein A (SpA) to isotype controlantibodies (IgG1, IgG2a, or IgG2b) or SpAKKAA-specific MAbs (5A10, 3F6, and 3D11). (B) Association of horseradish peroxidase (HRP)-conjugated SpAKKAA-specific MAbs (5A10-HRP, 3F6-HRP, and 3D11-HRP) to immobilized SpAKKAA was examined in a plate reader experiment where SpAKKAA was first incubatedwith isotype control antibodies (IgG1, IgG2a, or IgG2b) or three different SpAKKAA-specific MAbs (5A10, 3F6, and 3D11) to assess the possibility of competitiveinhibition for antibodies that bind the same or closely related sites (n 3). The values at an optical density at 405 nm (OD405) were measured and normalizedto the interaction of SpAKKAA and HRP-conjugated SpA-specific MAbs. Data are the means, and error bars represent � SEM. Data in panels A and B arerepresentative of three independent analyses. Asterisks denote statistical significance (P � 0.05), which was calculated using the two-tailed Student’s t test:5A10-HRP versus 5A10, P 0.0017; 5A10-HRP versus 3F6, P 0.0343; 5A10-HRP versus 3D11, P 0.001; 3F6-HRP versus 5A10, P 0.0279; 3F6-HRP versus3F6, P 0.0001; 3F6-HRP versus 3D11, P 0.0584; 3D11-HRP versus 5A10, P 0.0001; 3D11-HRP versus 3F6, P 0.0005; 3D11-HRP versus 3D11, P 0.0053.

TABLE 4 Amino acid sequences of CDRs of monoclonal antibodies

MAba

Amino acid sequencing data of protein A-specific monoclonalantibodies

Mouse VHfamily CDR1 CDR2 CDR3

5A10 7183 SSVSY DTS QQWSSYPPT3F6 VH10 ESVEYSGASL AAS QQSRKVPST3D11 J558 SSVSY EIS QQWSYPFTa Amplified PCR products from cDNA which was synthesized from total RNA extractedfrom hybridoma cells were sequenced and analyzed using IMGT Vquest.

Kim et al.

3466 iai.asm.org Infection and Immunity

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

with the control MAb, which can associate with SpA only in anonimmune fashion, caused a modest reduction of soluble Sbi1-4

(Fig. 6B). In contrast, incubation of staphylococci with MAb 3F6depleted soluble Sbi1-4, suggesting that MAb 3F6 bound SpA onthe bacterial surface while presenting its Fc� domain for associa-tion with Sbi1-4 (Fig. 6B).

To test whether the binding of MAb 3F6 to SpA occurs in vivo,BALB/c mice were passively immunized with MAb 3F6 or theisotype control antibody. Following the injection of purified SpAinto the peritoneal cavity, the abundance of protein A in circula-tion was assessed by sampling blood over the next 30 min. Com-pared to animals treated with control MAb, injection of MAb

3F6-treated animals caused accelerated clearance of SpA from thebloodstream (Fig. 6C). We presume that immune recognition ofSpA by MAb 3F6 provides for its Fc� domain to mediate Fc recep-tor-mediated removal of antigen-antibody complexes from thebloodstream.

SpAKKAA MAbs promote opsonophagocytic killing of staph-ylococci in human and mouse blood. Eliciting adaptive immuneresponses that promote opsonophagocytic killing of pathogens isa universal goal for vaccine development and licensure (38). Thishas not been achieved for S. aureus, as this pathogen is armedagainst opsonic antibodies via its surface-exposed and secretedSpA and Sbi molecules (24). To test whether SpAKKAA MAbs canpromote opsonophagocytosis, we employed Rebecca Lancefield’sassay of bacterial killing in fresh blood (30). Anticoagulated bloodfrom naïve 6-week-old BALB/c mice was incubated with MSSAstrain Newman in the presence or absence of 2 �g · ml�1 of MAbs5A10, 3F6, and 3D11 or their isotype controls. Blood samples werelysed and plated on agar medium, and staphylococcal load wasenumerated (Fig. 7A). All three MAbs triggered opsonophago-cytic killing of staphylococci, which ranged from 37% of the inoc-

FIG 6 SpAKKAA MAbs prevent the association of staphylococcal protein Awith immunoglobulin. (A) Isotype control antibodies or SpAKKAA MAbs wereused to perturb the binding of human IgG to wild-type protein A (SpA) orvariants that lack the ability to bind Fc� (SpAKK) or Fab (SpAAA) immobilizedon ELISA plates. The values were normalized to the protein A interaction withhuman IgG without antibodies (n 4). (B) Staphylococci were grown tomid-log phase and incubated with either isotype control antibody or MAb 3F6and followed by the addition of 2 �g wild-type Sbi1-4. Upon incubation, Sbi1-4

consumption was measured by immunoblot using affinity-purified �-SpAKKAA rabbit antibody. The values were normalized to Sbi1-4 sedimentationwithout antibody (No Ab). (C) Affinity-purified SpA (200 �g) was injectedinto the peritoneal cavity of mice pretreated with 85 �g (5 mg · kg�1) of eitherisotype control antibody or MAb 3F6. Animals were euthanized at indicatedtime points to measure the amount of SpA in circulating blood by immuno-blotting with affinity-purified �-SpAKKAA rabbit antibody (n 3 per timepoint). The values were normalized to the total amount of SpA injected at 0min. Data are the means, and error bars represent �SEM. Results in panels Ato C are representative of two independent analyses. The asterisks denotesstatistical significance (P � 0.05).

FIG 7 SpAKKAA MAbs promote opsonophagocytic killing of S. aureus inmouse and human blood. (A) Anticoagulated mouse blood was incubatedwith 5 � 105 CFU S. aureus Newman in the presence of isotype mouse anti-body controls or SpAKKAA MAbs (2 �g · ml�1) for 30 min, and survival wasmeasured (n 3). (B) Anticoagulated human whole blood was incubated with5 � 106 CFU S. aureus MW2 in the presence of isotype mouse antibody con-trols or SpAKKAA MAbs (10 �g · ml�1) for 120 min, and survival was measured(n 3). (C to H) At 60 min of incubation of staphylococci in anticoagulatedhuman blood, clusters of extracellular staphylococci were detected in samplesincubated with mouse isotype antibody controls (red arrowheads), whereasstaphylococci were found in close proximity to neutrophils (blue arrowheads)in samples with SpAKKAA MAbs. Data are the means, and error bars represent�SEM. Results in panels A to H are representative of three independent anal-yses. The asterisks denotes statistical significance (P � 0.05).

Protein A-Specific Monoclonal Antibodies

October 2012 Volume 80 Number 10 iai.asm.org 3467

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

ulum (3D11, P 0.0025) to 33% (3F6, P 0.0478) and 16%(5A10, P 0.0280). As a test for opsonophagocytic killing ofstaphylococci in human blood, we recruited healthy human vol-unteers and examined their serum for antibodies specific forSpAKKAA. As reported before, none of the volunteers harboredserum antibodies directed against protein A (data not shown)(22). Anticoagulated fresh human blood samples were incubatedwith MRSA strain USA400 (MW2) in the presence or absence of10 �g · ml�1 MAbs 5A10, 3F6, and 3D11 or their isotype controls(Fig. 7B). All three MAbs triggered opsonophagocytic killing ofstaphylococci, which ranged from 52% of the inoculum (3D11,P 0.0002) to 44% (3F6, P 0.0001) and 34% (5A10, P 0.0035). Blood samples were spread on glass slides, stained withGiemsa, and analyzed by microscopy. Blood samples incubated inthe presence of MAbs 5A10, 3F6, and 3D11 harbored staphylo-cocci that were associated with neutrophils, i.e., they may be asso-ciated with these leukocytes or located within cells (Fig. 7C to E).Blood samples incubated with isotype control MAbs harboredclusters of extracellular staphylococci (red arrowheads) as well asstaphylococci that were associated with leukocytes (blue arrow-heads, Fig. 7F to H).

DISCUSSION

Monoclonal antibodies offer unique opportunities to investigatethe biological attributes of humoral adaptive immune responsesto microbial surface products, revealing both the molecular na-ture of microbial immune evasion and of protective immunity(11). For example, group A streptococcal M protein, a key viru-lence factor and �-helical coiled-coil surface protein (37), confersresistance to opsonophagocytic clearance, which may be over-come by humoral adaptive immune responses during infection(29, 40). MAbs that bind to the �-helical coiled-coil of M proteincannot induce opsonophagocytic killing of group A streptococci,which is, however, achieved by MAbs directed against the N-ter-minal, random coil domain (20, 21). The N-terminal domain ofM proteins is highly variable between clinical isolates, which rep-resents the molecular basis for type-specific immunity (18, 29).

Similar to streptococcal M protein, protein A also functions asthe protective antigen of S. aureus (47). Virtually all clinical iso-lates of S. aureus express protein A; however, the amino acid se-quence of its IgBDs is highly conserved (34). Staphylococcal in-fections in mice or humans do not elicit protein A-specifichumoral immune responses (22), which is explained by the B cellsuperantigen activity of this molecule (41). Immunization withprotein A variants, in particular the SpAKKAA molecule, elicits hu-moral immune responses in mice and rabbits; these antibodiescross-react with wild-type protein A and provide protectionagainst staphylococcal disease in mice (22). Affinity-purified poly-clonal rabbit antibodies can block the B cell superantigen activityof wild-type protein A in mice and enhance the opsonophagocyticcapacity of mouse neutrophils when incubated in anticoagulatedmouse blood. In addition, S. aureus mutants lacking the structuralgene for protein A (spa) display significant defects in virulence andalso permit the development of humoral immune responsesagainst many different staphylococcal antigens as well as the de-velopment of protective immunity (8, 24). Thus, antibodies thatneutralize the immune-modulatory attributes of SpA may notonly provide protection against acute staphylococcal infection butmay also enable the development of protective immune responses

against other staphylococcal antigens and prevent recurrent S. au-reus infections.

We tested this prediction by raising monoclonal antibodiesagainst SpAKKAA. All monoclonal antibodies that elicited protec-tive immunity in mice recognized conformational epitopes ofSpAKKAA and interacted with the triple-helical fold of its IgBDs.Importantly, these monoclonal antibodies with strong affinity andcross-reactivity for multiple or all IgBDs of SpAKKAA also recog-nized wild-type protein A. When tested in vitro, MAbs, in partic-ular 5A10, 3F6, and 3D11, prevented protein A association withthe Fc and the Fab domains of immunoglobulins and triggeredopsonophagocytic killing of S. aureus by phagocytes in mouse andhuman blood. When injected into the peritoneal cavity of mice,MAbs elicited significant immune protection against both MSSAand MRSA isolates. Further, SpAKKAA MAb-mediated neutraliza-tion of SpA in vivo stimulated humoral immune responses againstseveral different S. aureus antigens, supporting the hypothesis thatSpAKKAA antibodies inhibit the B cell superantigen activities ofstaphylococci. Of note, the magnitude of antibody responses to-ward staphylococcal antigens in passively immunized mice wasmuch lower than the magnitude of immune responses elicited inmice actively immunized with SpAKKAA. The main difference be-tween active and passive immunization strategies lies in the devel-opment of antigen-specific T/B cell populations governing hostimmune responses, which are the consequence of active immuni-zation strategies. Future studies are warranted to determine if theprotein A-specific T/B cells are critical in raising appropriate sys-temic immune responses such as TH1/17-mediated recruitment offunctional phagocytes against S. aureus infections (45).

Previous work demonstrated superantigen activity of proteinA toward VH3-type B cell receptors in mice (14). Of note, only 5 to10% of mouse B cells are VH3 clonal and susceptible to protein Asuperantigen (41). Nevertheless, protein A mutant staphylococcidisplay a profound defect in the pathogenesis of abscess formationin mouse models for this disease (8, 24). In contrast to mice, hu-man VH3 clonal B cells comprise up to 50% of the total B cellpopulation (4, 19), suggesting that the impact of protein A super-antigen activity during staphylococcal infection is likely greaterfor human B cell populations (41). If so, protein A-mediated B cellactivation may trigger biased use of VH3 B cell clones and thedevelopment of nonphysiological B cell populations. Ultimately,staphylococcal protein A is expected to diminish VH3-positive Bcells and VH3-type antibodies in the human host. A similar sce-nario is encountered with the HIV envelope glycoprotein gp120,which also interacts with VH3 clonal B cells, causing a clonal deficitof VH3 B cells (antibody genes) in AIDS patients (4, 5). Theseevents are likely key factors in the prevention of neutralizing an-tibody responses during HIV and S. aureus infection (25).

Data presented herein provide corroborating evidence for thegeneral hypothesis that the neutralization of the Fc� and Fabbinding activities of SpA represent a correlate for protective im-munity against S. aureus: such antibodies are expected to triggerthe opsonophagocytic killing of the pathogen in blood and to elicitantibodies that neutralize the secreted virulence factors of staph-ylococci (25).

ACKNOWLEDGMENTS

We thank Carol McShan for the generation of mouse hybridomas andmonoclonal antibody isolation and members of our laboratory for discus-sion and comments on the manuscript.

Kim et al.

3468 iai.asm.org Infection and Immunity

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

This work was supported by grants from the National Institute ofAllergy and Infectious Diseases (NIAID), Infectious Diseases Branch(AI52747 and AI92711 to O.S.). H.K.K., C.E., A.C.D., D.M.M., andO.S. acknowledge membership within and support from the Region VGreat Lakes Regional Center of Excellence in Biodefense and EmergingInfectious Diseases Research Consortium (NIH award 1-U54-AI-057153).

Hwan Keun Kim, Andrea C. DeDent, Dominique M. Missiakas, andOlaf Schneewind declare a conflict of interest as inventors of patent ap-plications that are related to the development of Staphylococcus aureusvaccines and are currently under commercial license.

REFERENCES1. Atkins KL, et al. 2008. S. aureus IgG-binding proteins SpA and Sbi: host

specificity and mechanisms of immune complex formation. Mol. Immu-nol. 45:1600 –1611.

2. Baba T, Bae T, Schneewind O, Takeuchi F, Hiramatsu K. 2008. Genomesequence of Staphylococcus aureus strain Newman and comparative anal-ysis of staphylococcal genomes. J. Bacteriol. 190:300 –310.

3. Baba T, et al. 2002. Genome and virulence determinants of high-virulence community acquired MRSA. Lancet 359:1819 –1827.

4. Berberian L, Goodglick L, Kipps TJ, Braun J. 1993. ImmunoglobulinVH3 gene products: natural ligands for HIV gp120. Science 261:1588 –1591.

5. Berberian L, Valles-Ayoub Y, Sun N, Martinez-Maza O, Braun J. 1991.A VH clonal deficit in human immunodeficiency virus-positive individu-als reflects a B-cell maturational arrest. Blood 78:175–179.

6. Burman JD, et al. 2008. Interaction of human complement with Sbi, astaphylococcal immunoglobulin-binding protein: indications of a novelmechanism of complement evasion by Staphylococcus aureus. J. Biol.Chem. 283:17579 –17593.

7. Cary S, Krishnan M, Marion TN, Silverman GJ. 1999. The murine clanV(H) III related 7183, J606 and S107 and DNA4 families commonly en-code for binding to a bacterial B cell superantigen. Mol. Immunol. 36:769 –776.

8. Cheng AG, et al. 2009. Genetic requirements for Staphylococcus aureusabscess formation and persistence in host tissues. FASEB J. 23:3393–3404.

9. DeDent A, Kim HK, Missiakas DM, Schneewind O. 2012. ExploringStaphylococcus aureus pathways to disease for vaccine development. Se-min. Immunopathol. 34:317–333.

10. Diep BA, et al. 2006. Complete genome sequence of USA300, an epidemicclone of community-acquired meticillin-resistant Staphylococcus aureus.Lancet 367:731–739.

11. Fischetti VA. 1989. Streptococcal M protein: molecular design and bio-logical behavior. Clin. Microbiol. Rev. 2:285–314.

12. Forsgren A. 1970. Significance of protein A production by staphylococci.Infect. Immun. 2:672– 673.

13. Forsgren A, Svedjelund A, Wigzell H. 1976. Lymphocyte stimulation byprotein A of Staphylococcus aureus. Eur. J. Immunol. 6:207–213.

14. Goodyear CS, Silverman GJ. 2003. Death by a B cell superantigen: in vivoVH-targeted apoptotic supraclonal B cell deletion by a staphylococcaltoxin. J. Exp. Med. 197:1125–1139.

15. Goodyear CS, Silverman GJ. 2004. Staphylococcal toxin induced prefer-ential and prolonged in vivo deletion of innate-like B lymphocytes. Proc.Natl. Acad. Sci. U. S. A. 101:11392–11397.

16. Graille M, et al. 2000. Crystal structure of a Staphylococcus aureus proteinA domain complexed with the Fab fragment of a human IgM antibody:structural basis for recognition of B-cell receptors and superantigen activ-ity. Proc. Nat. Acad. Sci. U. S. A. 97:5399 –5404.

17. Haupt K, et al. 2008. The Staphylococcus aureus protein Sbi acts as acomplement inhibitor and forms a tripartite complex with host comple-ment factor H and C3b. PLoS Pathog. 4:e1000250. doi:10.1371/journal.ppat.1000250.

18. Hollingshead SK, Fischetti VA, Scott JR. 1987. A highly conserved regionpresent in transcripts encoding heterologous M proteins of group A strep-tococci. Infect. Immun. 55:3237–3239.

19. Huang C, Stewart AK, Schwartz RS, Stollar BD. 1992. Immunoglobulinheavy chain gene expression in peripheral blood B lymphocytes. J. Clin.Invest. 89:1331–1343.

20. Jones KF, Fischetti VA. 1988. The importance of the location of antibody

binding on the M6 protein for opsonization and phagocytosis of group AM6 streptococci. J. Exp. Med. 167:1114 –1123.

21. Jones KF, et al. 1986. Immunochemical localization and amino acidsequences of crossreactive epitopes within the group A streptococcal M6protein. J. Exp. Med. 164:1226 –1238.

22. Kim HK, Cheng AG, Kim H-Y, Missiakas DM, Schneewind O. 2010.Nontoxigenic protein A vaccine for methicillin-resistant Staphylococcusaureus infections. J. Exp. Med. 207:1863–1870.

23. Kim HK, et al. 2010. IsdA and IsdB antibodies protect mice againstStaphylococcus aureus abscess formation and lethal challenge. Vaccine 28:6382– 6392.

24. Kim HK, Kim HY, Schneewind O, Missiakas DM. 2011. Identifyingprotective antigens of Staphylococcus aureus, a pathogen that suppresseshost immune responses. FASEB J. 25:3605–3612.

25. Kim HK, Thammavongsa V, Schneewind O, Missiakas D. 2012. Recur-rent infections and immune evasion strategies of Staphylococcus aureus.Curr. Opin. Microbiol. 15:92–99.

26. Klevens RM, Edwards JR, Gaynes RP, System NNIS. 2008. The impactof antimicrobial-resistant, health care-associated infections on mortalityin the United States. Clin. Infect. Dis. 47:927–930.

27. Köhler G, Milstein C. 1975. Continuous cultures of fused cells secretingantibodies of predefined specificity. Nature 256:495– 497.

28. Kronvall G, Grey HM, Williams RC, Jr. 1970. Protein A reactivity withmouse immunoglobulins: structural relationship between some mouseand human immunoglobulins. J. Immunol. 105:1115–1123.

29. Lancefield R. 1962. Current knowledge of type-specific M antigens ofgroup A streptococci. J. Immunol. 89:307–313.

30. Lancefield RC. 1928. The antigenic complex of Streptococcus hemolyticus.I. Demonstration of a type-specific substance in extracts of Streptococcushemolyticus. J. Exp. Med. 47:91–103.

31. Lowy FD. 1998. Staphylococcus aureus infections. N. Engl. J. Med. 339:520 –532.

32. Mazmanian SK, Liu G, Ton-That H, Schneewind O. 1999. Staphylococ-cus aureus sortase, an enzyme that anchors surface proteins to the cell wall.Science 285:760 –763.

33. Mazmanian SK, et al. 2003. Passage of heme-iron across the envelope ofStaphylococcus aureus. Science 299:906 –909.

34. McCarthy AJ, Lindsay JA. 2010. Genetic variation in Staphylococcus au-reus surface and immune evasion genes is lineage associated: implicationsfor vaccine design and host-pathogen interactions. BMC Microbiol. 10:173.

35. Palmqvist N, Silverman GJ, Josefsson E, Tarkowski A. 2005. Bacterialcell wall-expressed protein A triggers supraclonal B-cell responsesupon in vivo infection with Staphylococcus aureus. Microbes Infect.7:1501–1511.

36. Perry AM, Ton-That H, Mazmanian SK, Schneewind O. 2002. Anchor-ing of surface proteins to the cell wall of Staphylococcus aureus. III. Lipid IIis an in vivo peptidoglycan substrate for sortase-catalyzed surface proteinanchoring. J. Biol. Chem. 277:16241–16248.

37. Phillips GNJ, Flicker PFPF, Cohen C, Manjula BN, Fischetti VA.1981. Streptococcal M protein: alpha-helical coiled-coil structure andarrangement on the cell surface. Proc. Natl. Acad. Sci. U. S. A. 78:4689 – 4693.

38. Robbins JB, Schneerson R, Szu SC. 1996. Hypothesis: how licensedvaccines confer protective immunity. Adv. Exp. Med. Biol. 397:169 –182.

39. Schneewind O, Model P, Fischetti VA. 1992. Sorting of protein A to thestaphylococcal cell wall. Cell 70:267–281.

40. Scott JR, Guenther PC, Malone LM, Fischetti VA. 1986. Conversion ofan M� group A streptococcus to M� by transfer of a plasmid containingan M6 gene. J. Exp. Med. 164:1641–1651.

41. Silverman GJ, Goodyear CS. 2006. Confounding B-cell defences: lessonsfrom a staphylococcal superantigen. Nat. Rev. Immunol. 6:465– 475.

42. Sjödahl J. 1977. Repetitive sequences in protein A from Staphylococcusaureus. Arrangement of five regions within the protein, four being highlyhomologous and Fc-binding. Eur. J. Biochem. 73:343–351.

43. Sjöquist J, Meloun B, Hjelm H. 1972. Protein A isolated from Staphylo-coccus aureus after digestion with lysostaphin. Eur. J. Biochem. 29:572–578.

44. Smith EJ, et al. 2012. The immune evasion protein Sbi of Staphylococcusaureus occurs both extracellularly and anchored to the cell envelope bybinding lipoteichoic acid. Mol. Microbiol. 83:789 – 804.

45. Spellberg B, Daum R. 2012. Development of a vaccine against Staphylo-coccus aureus. Semin. Immunopathol. 34:335–348.

Protein A-Specific Monoclonal Antibodies

October 2012 Volume 80 Number 10 iai.asm.org 3469

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

46. Stahlenheim G, Sjöquist J. 1970. Protein A from Staphylococcus aureus.IX. Complement fixing activity of protein A-IgG complexes. J. Immunol.103:467– 473.

47. Stranger-Jones YK, Bae T, Schneewind O. 2006. Vaccine assembly fromsurface proteins of Staphylococcus aureus. Proc. Nat. Acad. Sci. U. S. A.103:16942–16947.

48. Zhang L, Jacobsson K, Strom K, Lindberg M, Frykberg L. 1999. Staph-ylococcus aureus expresses a cell surface protein that binds both IgG andbeta2-glycoprotein I. Microbiology 145:177–183.

49. Zhang L, Jacobsson K, Vasi J, Lindberg M, Frykberg L. 1998. Asecond IgG-binding protein in Staphylococcus aureus. Microbiology144:985–991.

Kim et al.

3470 iai.asm.org Infection and Immunity

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from