INFECTION AND IMMUNITY, July 2009, p. 2712–2718 Vol. 77, No. 70019-9567/09/$08.00�0 doi:10.1128/IAI.00115-09Copyright © 2009, American Society for Microbiology. All Rights Reserved.

Anti-Alpha-Hemolysin Monoclonal Antibodies Mediate Protectionagainst Staphylococcus aureus Pneumonia�

Brook E. Ragle1 and Juliane Bubeck Wardenburg1,2*Departments of Microbiology1 and Pediatrics,2 University of Chicago, 920 E. 58th Street, Chicago, Illinois 60637

Received 30 January 2009/Returned for modification 3 March 2009/Accepted 9 April 2009

Staphylococcus aureus pneumonia is one of the most common invasive diseases caused by this humanpathogen. S. aureus alpha-hemolysin, a pore-forming cytotoxin, is an essential virulence factor in the patho-genesis of pneumonia. Vaccine-based targeting of this toxin provides protection against lethal staphylococcalpneumonia in a murine model system, suggesting that a monoclonal antibody-based therapy may likewiseprove to be efficacious for prevention and treatment of this disease. We report the generation of two distinctanti-alpha-hemolysin monoclonal antibodies that antagonize toxin activity, preventing human lung cell injuryin vitro and protecting experimental animals against lethal S. aureus pneumonia. Each of these two monoclonalantibodies recognized an epitope within the first 50 amino acid residues of the mature toxin and blocked theformation of a stable alpha-hemolysin oligomer on the target cell surface. Active immunization with the first50 amino acids of the toxin also conferred protection against S. aureus pneumonia. Together, these data revealpassive and active immunization strategies for prevention or therapy of staphylococcal pneumonia andhighlight the potential role that a critical epitope may play in defining human susceptibility to this deadlydisease.

Staphylococcus aureus is a gram-positive human pathogenthat causes a myriad of diseases ranging from minor skin in-fections to life-threatening deep tissue infections and toxinoses(19). Pneumonia is among the most prominent S. aureus-me-diated diseases, accounting for approximately 15% of docu-mented invasive staphylococcal infections (14), and there arean estimated 50,000 cases per year in the United States alone(17). In addition to being one of the most common causes ofventilator-associated pneumonia, S. aureus is increasingly rec-ognized as an important cause of community-acquired pneu-monia, affecting previously healthy adults and children (8, 16).This is particularly notable in association with influenza infec-tion, where concomitant staphylococcal pneumonia is often alethal complication (7, 8, 12). Up to one-half of staphylococcalpneumonia isolates are classified as methicillin (meticillin)-resistant S. aureus (MRSA), confounding the delivery of ap-propriate treatment and resulting in reported mortality as highas 56% (1, 17, 24). The combination of an increasing diseaseburden and declining potency of traditional antimicrobials tocombat S. aureus pneumonia heightens the need for novelprophylactic and therapeutic strategies.

We have defined an essential role of alpha-hemolysin (Hla)in S. aureus pneumonia, as strains lacking this pore-formingcytotoxin are avirulent in a murine model of disease (4). Draw-ing on this knowledge, we have demonstrated that vaccine-based approaches targeting Hla provide protection from lethalpneumonia in experimental animals (5). The ability of Hla toinjure the lung and other tissues rests on the ability of the toxinto form a 2-nm heptameric pore in the plasma membrane of

susceptible cells (2, 26). This chromosomally encoded toxin issecreted as a water-soluble monomer by the majority of S.aureus strains (22). Membrane binding of the monomer per-mits a series of well-defined intermolecular interactions be-tween neighboring monomers, resulting in the formation of abarrel-shaped oligomeric pore that penetrates the membrane(9, 13). Residues located at the N terminus of the mature toxinare essential for assembly of the lytic oligomer, as point mu-tations or truncations within this region disrupt the formationof an active toxin (21, 27, 28).

In addition to its role in the lung, Hla is central to patho-genesis in other tissues, as hla mutants are less virulent inanimal models of intraperitoneal (i.p.), intramammary, andcorneal infection (3, 6, 23). Supporting this role for Hla indisease, immune sera generated against a single point mutantwith a mutation that disrupts pore formation, termed HlaH35L,provide a high degree of protection against pneumonia, i.p.infection, and challenge with purified active toxin (5, 20). Wetherefore built upon these observations by generating mousemonoclonal antibodies (MAbs) following immunization withinactive HlaH35L to investigate whether an antibody with asingle specificity could provide protection against S. aureuspneumonia.

MATERIALS AND METHODS

Bacterial strains and culture. For mouse lung infection, S. aureus strainsNewman and LAC/USA300 were grown at 37°C in tryptic soy broth to an opticaldensity at 660 nm of 0.5. Culture aliquots (50 ml) were centrifuged and washedin phosphate-buffered saline (PBS) prior to resuspension. For mortality studies,S. aureus Newman was resuspended in 750 �l (3 � 108 to 4 � 108 CFU per 30�l), while LAC/USA300 was resuspended in 1,250 �l (2 � 108 CFU per 30 �l).For bacterial load and histopathology experiments, S. aureus Newman was re-suspended in 1,250 �l (2 � 108 CFU per 30 �l). For cytotoxicity studies, 5 ml ofa culture prepared as described above was resuspended in 10 ml F12K medium(Invitrogen). A 100-�l suspension was used for each assay well.

* Corresponding author. Mailing address: Departments of Pediat-rics and Microbiology, University of Chicago, 920 E. 58th St., Chicago,IL 60637. Phone: (773) 834-9763. Fax: (773) 834-8150. E-mail:jbubeckw@peds.bsd.uchicago.edu.

� Published ahead of print on 20 April 2009.

2712

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

Plasmid construction. PCR products encoding serial 50-amino-acid fragmentsof Hla, amplified from S. aureus Newman chromosomal DNA, were cloned intopGEX-6P-1 (GE Healthcare) and transformed into Escherichia coli. Eachconstruct was sequenced for verification. A plasmid encoding glutathioneS-transferase (GST)-HlaH35L was generated as previously described (5). Togenerate a polyhistidine-tagged version of full-length, active Hla, a PCRproduct encoding the mature polypeptide was amplified from S. aureus New-man chromosomal DNA, cloned into pET24b (Novagen), and then trans-formed into E. coli BL21/DE3.

MAbs. MAbs to Hla were generated by the Frank W. Fitch MonoclonalAntibody Facility at the University of Chicago. Splenocytes derived from miceimmunized with full-length HlaH35L were utilized to generate hybridomas. Con-trol MAbs of isotypes immunoglobulin G2a (IgG2a) and IgG2b were purifiedand supplied by the Frank W. Fitch Monoclonal Antibody Facility.

Animals and procedures. Animal experiments were reviewed, approved, andsupervised by the Institutional Animal Care and Use Committee at the Univer-sity of Chicago. For lung infection, 7-week-old C57BL/6J mice (The JacksonLaboratory) were anesthetized before inoculation of 30 �l of an S. aureus sus-pension prepared as described above into the left naris. Animals were placed intoa cage in a supine position for recovery and were observed for the time coursesindicated below. Routinely, a small percentage of animals died within the first 6 hafter inoculation, likely from the combined effects of aspiration and anesthesia.These animals were not included in subsequent statistical analyses.

For passive immunization studies, 7-week-old mice received via i.p. injection420 �l of either IgG2a, IgG2b, MAb 7B8, or MAb 1A9 at the concentrationsindicated below 24 h before S. aureus challenge. For active immunization,4-week-old mice received 20 �g of either GST, GST-HlaH35L, or GST-Hla1-50 incomplete Freund’s adjuvant on day 0 via the intramuscular (i.m.) route, followedby a boost with 20 �g of each protein antigen in incomplete Freund’s adjuvant onday 10. Animals were challenged with S. aureus on day 21. Sera were collectedbefore immunization and on day 20 to assess specific antibody production.

To evaluate the pathological correlates of pneumonia, infected animals werekilled via forced CO2 inhalation before removal of both lungs. The right lung washomogenized for enumeration of the lung bacterial load using serial dilution andplating techniques. The left lung was placed in 10% formalin, embedded inparaffin, and sectioned, and thin sections were stained with hematoxylin-eosinand analyzed by microscopy.

Enzyme-linked immunosorbent assay (ELISA). Serum antibody titers weredetermined with immunoplates (MaxiSorp; Thermo Fisher Scientific) coatedwith 1 �g/ml purified HlaH35L or GST-Hla1-50. Dilutions of either mouse orhuman sera prepared in PBS were incubated in the appropriate plates, whichwere developed with horseradish peroxidase-conjugated secondary antibodiesand a 3,3�,5,5�-tetramethylbenzidene substrate kit (Thermo Scientific) and ex-amined with a spectrophotometer (GENios; Tecan). All human sera were col-lected in accordance with a human subject protocol that was reviewed, approved,and supervised by the Institutional Review Board at the University of Chicago.

Live/dead and cytotoxcity assays. A549 cells were washed and plated in F12Kmedium supplemented with 10% fetal bovine serum at a density of 1.5 � 104 cellsper well in a 96-well plate. For both live/dead and cytotoxcity assays, washedA549 cells were cultured with 100 �l of staphylococcal suspension per well inF12K medium with or without antibody in triplicate wells. After 4 h of incubationat 37°C, either cells were treated with a live (green)/dead (red) reagent (Invitro-gen), or lactate dehydrogenase (LDH) activity was determined (Roche) accord-ing to the manufacturer’s recommendations. Microscopic images of stained cellswere obtained using a microscope (Eclipse TE2000U; Nikon); LDH activity wasmeasured with a spectrophotometer and was expressed as the percentage ofmaximal lysis obtained after detergent treatment of the A549 cells. The resultsare representative of a minimum of two independent experiments.

Protein preparation and immunoblot analysis. All GST- and His-tagged fu-sion proteins were prepared and purified using standard protocols. For dot blotanalysis, each antigen was spotted on a nitrocellulose membrane that was thenblocked with 5% milk in Tris-buffered saline containing 0.1% Tween 20. EachMAb of interest was used at a final concentration of 1 �g/ml. Blots were devel-oped using goat anti-mouse horseradish peroxidase-conjugated secondary anti-body and SuperSignal West Pico chemiluminescent substrate (Thermo Scien-tific). For molecular modeling, atomic coordinates were retrieved from theProtein Data Bank (PDB ID 7ahl) based on the data of Song et al. (26). Modelswere generated using PyMol (http://www.pymol.org/).

Oligomerization and binding assay. Radiolabled Hla was synthesized by invitro transcription and translation in an E. coli S30 extract (Promega) supple-mented with T7 RNA polymerase, rifampin (rifampicin), and [35S]methionineaccording to the manufacturer’s instructions. One hundred twenty microliters of12.5% rabbit red blood cells (RRBC) in K-PBSA/�ME (20 mM potassium

phosphate [monobasic], 150 mM NaCl [pH 7.4], 1 mg/ml bovine serum albumin,1 mM �-mercaptoethanol) was incubated with 30 �l of the radiolabeled Hlamixture in the presence of 0.1 to 5 �M MAb 7B8 or 1A9 for 1 h at 20°C. Thecontrols were analyses performed in the absence of added antibody or in thepresence of 5 �M isotype-matched MAbs as indicated below. Following incuba-tion, samples were centrifuged at 13,000 rpm for 5 min and then washed with 500�l K-PBSA/�ME and centrifuged as described above. Samples were then resus-pended in 90 �l 1� Laemmli buffer and incubated at 37° for 10 min before 12-�lportions of the samples were loaded onto 10% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) gels for electrophoresis. The gelswere dried, and then the results were visualized using a phosphorimager.

RRBC hemolysis assay. Purified, active full-length Hla was added to 900 �l of12.5% RRBC in PBS to a final concentration of 100 nM. The MAbs indicatedbelow or PBS was added to the reaction mixture (100 �l), and the cells were thenincubated at 20°C for 1 h. The reaction mixtures were centrifuged at 13,000 rpmfor 5 min, and the absorbance at 475 nm of the supernatants was measured.Percentages of hemolysis were calculated using the supernatant value for anequivalent number of cells that had been lysed in 1% Triton X-100.

Surface plasmon resonance (SPR). The affinity and rates of association anddissociation between 7B8 and Hla and between 1A9 and Hla were measuredusing a BIAcore 3000. The carboxyl groups on the sensor surface of a CM5 chipwere activated with 0.2 M N-ethyl-N-(3-diethylamino-propyl)carbodiimide and0.05 M N-hydroxysuccinimide. Antibody was bound to the chip by passing 1 �Mantibody in HBS-P buffer (20 mM HEPES [pH 7.4], 150 mM NaCl, 0.005%[vol/vol] surfactant P20) over the activated chip. Free amines were then neutral-ized using 1 M ethanolamine hydrochloride. The control surface was preparedsimilarly, except that running buffer was injected instead of Hla. We measuredthe rates of association and dissociation of purified Hla at concentrations rangingfrom 5 to 100 nM. All measurements were performed in triplicate. After eachbinding experiment the sensor chip was regenerated using 10 mM NaOH. De-rived sensorgrams were analyzed using BIAevaluation 4.1. Affinity constantswere estimated by curve fitting using a 1:1 binding model.

Statistical analysis. In mortality studies statistical significance was determinedusing the Fisher exact test; the significance of LDH release assay results and theresults of bacterial recovery from lungs and red cell hemolysis was calculatedusing the two-tailed Student t test.

RESULTS AND DISCUSSION

Anti-Hla MAbs prevent Hla-mediated cell injury. The abilityto neutralize Hla-mediated S. aureus lung injury through thedevelopment of MAbs offers an exciting new strategy for theprevention and treatment of pneumonia. To this end, we gen-erated murine MAbs to the nontoxic HlaH35L mutant. Immu-noreactive MAbs were examined to determine their ability toprevent injury to cultured alveolar epithelial cells (data notshown). Two MAbs, MAbs 7B8 and 1A9, provided a highdegree of protection against Hla-mediated injury. Uninfectedhuman A549 alveolar epithelial cells retain a green fluoro-phore when they are examined with a live (green)/dead (red)staining reagent (Fig. 1A). Infection of cells with S. aureusNewman, a methicillin-sensitive clinical isolate, resulted in celldeath after 3 h of coculture in the presence of PBS, as dem-onstrated by an increased number of red fluorescent cells (Fig.1B). Treatment of A549 cells with IgG2a, an isotype-matchedcontrol for 7B8, was similar to treatment with PBS (Fig. 1C). Incontrast, treatment with 7B8 provided nearly complete protec-tion from S. aureus-induced death (Fig. 1D). NonspecificIgG2b did not confer protection (Fig. 1E), while isotype-matched MAb 1A9 provided robust protection (Fig. 1F). Thepercentages of cell death are shown in Fig. 1.

To quantify the ability of MAbs 7B8 and 1A9 to antagonizethe effects of Hla on A549 cells, we examined concentration-dependent protection using an LDH release assay. Both MAbsconferred a significant degree of protection at concentrationsranging from 0.5 to 5 �g/ml (3 to 33 nM) when they were

VOL. 77, 2009 MAbs PREVENT S. AUREUS PNEUMONIA 2713

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

added to cocultures of A549 cells with live S. aureus (Fig. 1Gand H) (P � 0.02 for all conditions). The calculated 50%inhibitory concentration for 7B8 was 2.83 �g/ml, while that for1A9 was 1.92 �g/ml. These results reveal the in vitro protectiveefficacy of two novel anti-Hla MAbs which protect human lungepithelial cells from S. aureus-mediated injury. The low con-centrations at which these antibodies have protective effects invitro highlight their potential therapeutic relevance.

Passive immunization with anti-Hla MAbs protects micefrom S. aureus pneumonia. To examine the ability of anti-HlaMAbs to prevent S. aureus-induced pneumonia, we passivelyimmunized 7-week-old C57BL/6J mice with 7B8 or 1A9 deliv-ered via i.p. injection 24 h prior to intranasal (i.n.) infectionwith 3 � 108 to 4 � 108 CFU of S. aureus Newman. Groups of15 mice received 10, 5, 1, or 0.1 mg/kg of 7B8 (Fig. 2A) or 1A9(Fig. 2B); control animals received 10 mg/kg of the corre-sponding control MAb. Mice were observed for acute lethaldisease secondary to S. aureus pneumonia. For animals receiv-ing as little as 1 mg/kg of 7B8 there was a significant reductionin mortality (Fig. 2A) (P � 0.015) compared to animals receiv-ing the isotype control. The protective efficacy of 1A9 was lessprominent; a reduction in mortality over 72 h was evident witha minimum MAb dose of 5 mg/kg (Fig. 2B) (P � 0.039).Concentrations of 7B8 that resulted in a significant decrease inmortality resulted in half-maximal serum antibody titersgreater than 1:547 � 218, similar to anti-Hla titers that we havereported to confer protection in prior immunization studies(5). Interestingly, 1 mg/kg of 1A9 resulted in a half-maximaltiter of 1:606 � 143, yet it was not protective, suggesting that1A9 has either a lower affinity for Hla or a diminished capacityfor functional neutralization.

To evaluate the pathological correlates of 7B8 and 1A9protection, we examined S. aureus recovery from lungs 24 hafter infection for mice that received 10 mg/kg of each MAb ora control. Immunization with 7B8 and 1A9 led to markedreductions in the number of CFU recovered from the rightlung compared to controls (Fig. 2C) (P � 0.027 and P � 0.02,respectively). Gross inspection of infected lungs revealed morefocal disease in animals that received 7B8 or 1A9, as indicatedby a reduction in the dense, red appearance of the tissue (Fig.2D). These observations were confirmed by histopathologicanalysis, which showed that the majority of the airspace incontrol mice was obliterated by inflammatory cell infiltrates oraggregates of S. aureus (Fig. 2E, left panels); in contrast, inanimals treated with 7B8 or 1A9 the lung tissue was predom-inately unaffected (Fig. 2E, right panels).

As an increasing number of severe S. aureus pneumonias arecaused by MRSA isolates, we examined the efficacy of 7B8 and1A9 for protecting mice against the highly virulent MRSAstrain LAC/USA300. We have previously shown that LAC/USA300 secretes approximately twice as much Hla as S. aureus

FIG. 1. Anti-Hla MAbs protect human alveolar epithelial cellsfrom S. aureus injury. (A to F) Live (green)/dead (red) staining ofA549 alveolar epithelial cells as imaged by fluorescence microscopy 4 hafter infection. Cells were not infected (A) or cocultured with S. aureusNewman in medium treated with PBS (B) or with 2 �g/ml IgG2a (C),7B8 (D), IgG2b (E), or 1A9 (F). Bars, 20 �m. The percentage of deadcells for each experimental condition was calculated by scoring live anddead cells in four independent fields (each field contained a minimumof 60 cells) and expressing the number of dead cells as a fraction of thetotal number of cells; the resulting values are indicated at the bottomright in each panel. The statistical significance of MAb-induced pro-tection was calculated by using Student’s t test (P � 0.004 for IgG2aversus 7B8; P � 0.002 for IgG2b versus 1A9). No statistically signifi-

cant differences were observed between PBS and control IgG-treatedcells. (G and H) LDH release by A549 cells was observed by using cellscocultured with S. aureus Newman in media treated with the indicatedconcentrations of 7B8 (G) or 1A9 (H), revealing that there was asignificant reduction in LDH release for all MAb concentrations ex-amined (P � 0.02). The error bars indicate the standard deviations.

2714 RAGLE AND BUBECK WARDENBURG INFECT. IMMUN.

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

Newman, which directly correlates with increased virulence inthe lung (5). Treatment of mice with 7B8 or 1A9 24 h prior toi.n. infection with 2 � 108 CFU of LAC/USA300 resulted in amarked reduction in mortality compared to the controls (Fig.3A) (P � 0.013 and P � 0.002, respectively), confirming theability of these MAbs to protect against both methicillin-sen-sitive S. aureus and MRSA isolates upon prophylactic admin-istration.

The clinical utility of MAbs that prevent S. aureus pneumo-nia, especially disease caused by virulent strains such as LAC/USA300, is readily appreciated when surgical and intensivecare patients that are highly susceptible to staphylococcal ven-

tilator-associated pneumonias are considered (15, 25). A morefar-reaching impact, however, would be possible if these MAbsconfer protection following the onset of infection. To examinetherapeutic efficacy, mice were immunized via the i.p. routewith 10 mg/kg control IgG2a or 7B8. Anti-Hla MAb-treatedanimals received doses 24 h prior to infection or at 4, 8, or 12 hafter infection with S. aureus Newman. Treatment up to 8 hfollowing infection, when mice exhibit prominent signs of dis-ease, conferred significant protection from 24-h mortality (P �0.021) (Fig. 3B). This protection was not durable, however, asthe late mortality was similar to that observed for the control.While the absolute 24-h mortality was reduced in mice vacci-

FIG. 2. Passive transfer of anti-Hla MAbs protects against S. aureus pneumonia. Mice were vaccinated via i.p. injection with the indicated dosesof IgG2a or 7B8 (A) or IgG2b or 1A9 (B). They were challenged 24 h later with S. aureus Newman via the i.n. route, and mortality was recordedat 24, 48, and 72 h postinfection (P � 0.039; 15 animals per group). The CFU recovered from the right lung were enumerated (C), demonstratingthat there was a significant decrease in the bacterial burden in mice vaccinated with the 7B8 or 1A9 MAb (P � 0.027 and P � 0.02, respectively;15 animals per group). The horizontal bars indicate the mean bacterial load. (D and E) Decreases in the gross pathological (D) and histopathologic(E) hallmarks of disease in animals immunized with 7B8 and 1A9 were evident when these animals were compared to control animals.

VOL. 77, 2009 MAbs PREVENT S. AUREUS PNEUMONIA 2715

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

nated 12 h after infection, the results were not statisticallysignificant. These data suggest that Hla may be essential for anearly stage of pathogenesis when the toxin may be crucial forinciting lung injury. Once this injury has occurred, antagonismof the toxin may be of little benefit. An extrapolation of thesefindings to human disease implies that treatment with suchMAbs early in the course of S. aureus pneumonia has thepotential to delay disease progression, providing a much-needed window of opportunity to enhance the utility of anti-microbial and supportive therapies.

SPR reveals similar affinities of MAbs for Hla. To quantifythe affinity of 7B8 and 1A9 for Hla, we performed SPR studies.Purified toxin (5 to 100 nM) was passed over each MAb, whichwas amino coupled to a CM5 chip, allowing determination ofthe kinetic rate constants for association and dissociation andthe affinity constant. Both 7B8 and 1A9 have affinities in thelow-nanomolar range (8.54 nM and 6.49 nM, respectively)(Fig. 4A), indicating that these MAbs are able to form a tightassociation with Hla, a feature that likely enhances their ca-pacity to neutralize toxin activity.

MAb recognition of an N-terminal epitope prevents Hlaoligomerization. An appreciation of the role of specific aminoacids and regions of Hla in toxin assembly and function hasbeen obtained through studies of the crystal structure of theheptameric toxin and a series of elegant biochemical studies(21, 26, 27). To examine the mechanism by which 7B8 and 1A9neutralize Hla, we generated a series of GST fusion proteinscomprised of overlapping 50-amino-acid segments of the ma-ture toxin to map the epitope(s) recognized by each MAb (Fig.4B). Dot blot analysis of these fusion proteins revealed that7B8 and 1A9 bind specifically to full-length Hla (Fig. 4B, lane1) and to the first 50 amino acids of the toxin (lane 2); aCoomassie blue-stained SDS-PAGE gel included the corre-sponding controls, where full-length Hla (amino acids 1 to 293)(lane 1) migrated at approximately 62 kDa and the panel of50-amino-acid fusion molecules (lanes 2 to 8) migrated near 30kDa. Neither MAb recognizes the epitope in a denaturedform, whether it is presented as a 50-amino-acid segment or asa panel of synthesized 15-mer peptides (data not shown), sug-gesting that the epitope determinants are conformational in

nature. Interestingly, the first 50 amino acids comprise a regionof the mature toxin that is required for the formation of afunctionally active oligomeric pore. The structure of this re-gion of Hla can be appreciated by examination of these resi-dues (Fig. 4C) in the context of the full-length monomer(based on a predicted structure) and the assembled heptamerderived from the known crystal structure (26).

To investigate whether the interaction of each MAb with Hladisrupts the formation of fully assembled Hla, we utilized aRRBC hemolysis assay in which [35S]methionine-labeled Hla isvisualized by SDS-PAGE in its monomeric and heptameric forms(27). To confirm the functional ability of 7B8 and 1A9 to protectagainst hemolysis, we incubated 3 � 108 cells with 100 nM puri-fied, active Hla and each MAb at concentrations of 0.1 to 5 �M.Dose-dependent prevention of hemolysis was observed for bothMAbs (Fig. 5A); a significant reduction in hemolysis was observedwith concentrations of 7B8 as low as 0.5 �M (P � 0.005) and withconcentrations of 1A9 as low as 0.1 �M (P � 0.033). An exami-nation of the radiolabeled monomeric toxin (Hla) (Fig. 5B) dem-onstrated that the binding of Hla to RRBC was not impaired byaddition of 7B8 or 1A9. Importantly, only Hla that is bound to theRRBC membrane was revealed in the autoradiogram. Thus, thesimilar signal intensities observed for the monomer in lanestreated with the MAb and in lanes without antibody treatment orin the presence of control antibody strongly suggest that the

FIG. 4. Anti-Hla MAbs bind to epitopes within the first 50 aminoacids of Hla. (A) SPR was performed by amino coupling 7B8 and 1A9to a CM5 chip. Binding of purified Hla revealed that the affinities ofeach MAb for Hla were in the low-nanomolar range. ka, kinetic rateconstant for association; kd, kinetic rate constant for dissociation; KD,affinity constant. (B) Overlapping, serial 50-amino-acid segments ofHla were purified as GST fusion proteins and used in a dot blotanalysis to map the epitopes recognized by 7B8 and 1A9, which re-vealed that both MAbs bind within the first 50 amino acids. TheCoomassie blue-stained SDS-PAGE gel shows the integrity of eachfusion protein. (C) Structure of Hla as a monomer (left structure) andas a heptamer (middle and right structures), in which the first 50 aminoacids recognized by the MAbs are indicated by black lines.

FIG. 3. Anti-Hla MAbs protect against highly virulent S. aureusLAC/USA300 and have therapeutic value. (A) Immunization with 7B8and 1A9 conferred protection against 48- and 72-h mortality in miceinfected via the i.n. route with S. aureus LAC/USA300 (P � 0.013 andP � 0.002, respectively; 15 animals per group). (B) Mice were passivelyimmunized with 7B8 24 h prior to i.n. challenge with S. aureus Newmanor 4, 8, or 12 h postchallenge. A significant decrease in 24-h mortalitywas observed for animals treated up to 8 h postinfection (P � 0.021; 15animals per group). Statistical significance (P � 0.05) is indicated by anasterisk.

2716 RAGLE AND BUBECK WARDENBURG INFECT. IMMUN.

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

MAbs do not function merely by preventing toxin binding totarget cells. The monomeric form of Hla produced in vitro mi-grated as two bands, as seen in previous studies (13). Each MAb,however, prevented the assembly of Hla into an SDS-stable oli-gomer (Hla7) in a close-dependent fashion (Fig. 5B). The inhibi-tion of heptamer formation in the presence of 7B8 and 1A9correlates with protection from hemolysis, defining a potent mo-lecular mechanism by which these MAbs exert their protectiveeffects. While the affinity constants of these MAbs for Hla weresimilar, 7B8 prevents the formation of the heptamer more effec-tively than 1A9 does, which correlates with enhanced in vivoprotection.

Active vaccination with Hla1-50 prevents mortality due to S.aureus pneumonia. These observations led to the hypothesisthat active immunization with a purified antigen consisting ofthe first 50 amino acids of Hla may prevent disease. Groups of15 4-week-old mice were vaccinated via the i.m. route on day 0with 20 �g of GST, GST-HlaH35L, or GST-Hla1-50 emulsifiedin complete Freund’s adjuvant and then were boosted on day10 with antigen emulsified in incomplete Freund’s adjuvant.On day 21, animals were infected via the i.n. route with 3 � 108

to 4 � 108 S. aureus Newman CFU and evaluated for protec-tion from lethal pneumonia. In agreement with our previouslypublished observations, immunization with GST-HlaH35L ledto complete protection, in contrast to immunization with GSTalone, after which the majority of mice succumbed to infection(P � 0.001) (Fig. 6) (5). Immunization with Hla1-50 resulted in

significant protection over the entire course of the assay, andfull protection was evident at 24 h (P � 0.022). The half-maximal serum antibody titers for each immunogen at the timeof infection were 1:3,178 � 1,468 (GST-HlaH35L) and 1:692 �669 (GST-Hla1-50). These observations, along with our descrip-tion of protective Hla MAbs, have important implications forthe development of clinically relevant immunologic tools tocombat S. aureus pneumonia. First, these studies highlight thepotential to utilize only a portion of Hla as a component of amultisubunit S. aureus vaccine. This may simplify vaccine de-sign, allowing generation of a chimeric antigen in which asegment of Hla can be fused to a second candidate vaccineantigen. Further, a greater margin of safety can be achieved byavoiding the use of a full-length variant of a potent humantoxin. Second, the humanization of MAbs, notably MAb 7B8,with specificity for the N-terminal region of Hla may providean efficient means by which to translate these observations tothe clinical arena.

Serum antibody titers to Hla1-50 vary within the human pop-ulation. Implementation of an effective Hla-based immuno-therapy for S. aureus pneumonia is predicated on the notionthat some humans do not harbor a preexisting, effective hostresponse to this toxin, rendering them more susceptible todisease. Previous studies have clearly documented that serumantibodies to Hla are commonly found within populations ofindividuals with invasive S. aureus disease; however, large-scaleinvestigations of preformed anti-Hla responses in subjectswithout such disease have not been reported (10, 11, 18). Toevaluate the human antibody response to epitopes within thefirst 50 amino acids of Hla, we examined sera from 25 healthyvolunteers using an ELISA. While all individuals had detect-able antibodies to Hla1-50, the coefficient of variation withinthe population was calculated to be 50.3% when a 1:1 dilutionof serum was analyzed (Fig. 6B). This observation has twoimportant implications. First, it suggests that there is a definedpopulation of humans with relatively lower levels of antibodyto at least one epitope of Hla shown in this work to be pro-

FIG. 6. Active immunization with the first 50 amino acids of Hlaprevents S. aureus pneumonia. C57BL/6J mice were vaccinated by i.m.injection with GST, GST-HlaH35L, or GST-Hla1-50 and challenged withS. aureus Newman via the i.n route. Mortality was recorded 24, 48, and72 h after infection (P � 0.022; 15 animals per group). Statisticalsignificance is indicated by an asterisk. (B) ELISA was performedusing the indicated dilutions of serum from 25 human volunteers todetermine the titer of anti-Hla1-50 antibodies, expressed as raw units ofoptical density at 450 nm (OD450). The bottom and top of each boxindicate the lower and upper quartiles, respectively, and the horizontalbar indicates the median for the group of samples. The whiskersindicate the lowest and highest values within a distance of 1.5 times theinterquartile range measured from the lower and upper quartiles,respectively; circles indicate outliers in each set.

FIG. 5. Anti-Hla MAbs prevent oligomerization of the toxin on hostcells. (A) Hemolysis assays were performed with 100 nM purified Hla and12.5% RRBC in PBS. Addition of 0.5 to 5 �M 7B8 and addition of 0.1 to5 �M 1A9 significantly reduced hemolysis, as measured by the A450 ofassay supernatants (P � 0.005 and P � 0.033, respectively). (B) Active[35S]methionine-labeled toxin was synthesized in vitro and added toRRBC in the presence of IgG2a, IgG2b, or the indicated concentrationsof 7B8 or 1A9. Concentration-dependent inhibition of toxin oligomeriza-tion (designated Hla7) was evident with both 7B8 and 1A9. Neither MAbdisrupted binding of the labeled toxin to RRBC, as indicated by thepresence of the Hla monomer in all lanes.

VOL. 77, 2009 MAbs PREVENT S. AUREUS PNEUMONIA 2717

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

tective. These individuals may benefit from vaccine strategiestargeting Hla, perhaps especially in clinical situations such ascritical illness, major surgical procedures, and influenza infec-tion that confer risk for the development of S. aureus pneu-monia. Second, the data highlight the importance of under-standing the human immune response to S. aureus disease andthe immunologic correlates of protection from invasive dis-ease. In order to fully appreciate the role of preexisting anti-bodies to Hla in conferring protection from pneumonia, it willbe necessary to conduct prospective clinical studies in whichthe anti-Hla antibody profile of “at-risk” subjects can be cor-related with the development of disease. A definitive link be-tween anti-Hla antibody levels in general or levels of an anti-body to a specific protective epitope and risk for pneumoniawould make it possible to stratify risk for the development ofdisease. The capacity to prospectively identify at-risk individ-uals would facilitate highly focused application of developingimmunotherapies.

The clinical complexity of S. aureus infection and the fact thatthere is no single virulence factor central to the progression of alldisease manifestations pose a challenge for development of auniversal vaccine. The essential role of Hla in pneumonia, cou-pled with the clinical burden of this disease, highlights the poten-tial of the observations described here to have a significant impactupon ongoing and future efforts to design novel immunotherapiesto specifically combat lung infection. Further, these observationsshed light on a new strategy with which to conceptualize thetargeting of Hla in a multifaceted universal S. aureus vaccine.

ACKNOWLEDGMENTS

This work was supported by an NCI Cancer Center award (grant5P30CA014599-33) to the Frank W. Fitch Monoclonal Antibody Fa-cility), by the Pediatric Scientist Development Program (grant K12-HD00850 to J.B.W.), and by the Departments of Pediatrics and Mi-crobiology at the University of Chicago. This work was sponsored bythe NIH/NIAID Regional Center of Excellence for Bio-defense andEmerging Infectious Diseases Research (RCE) Program.

We acknowledge membership in and support from the Region VGreat Lakes RCE (NIH award 1-U54-AI-057153).

We have no conflicting financial interests.We thank D. Missiakas, A. DeDent, and O. Schneewind for critical

discussions and comments on the manuscript, the Department of Pa-thology at the University of Chicago for histology support, K. Alex-ander for pharmacologic calculations, S. Bond for microscopy support,M. Davis, C. Mulligan, and C. McShan for MAb support, and T.Karrison for statistical support.

REFERENCES

1. Athanassa, Z., I. I. Siempos, and M. E. Falagas. 2008. Impact of methicillinresistance on mortality in Staphylococcus aureus VAP: a systematic review.Eur. Respir. J. 31:625–632.

2. Bhakdi, S., and J. Tranum-Jensen. 1991. Alpha-toxin of Staphylococcusaureus. Microbiol. Rev. 55:733–751.

3. Bramley, A. J., A. H. Patel, M. O’Reilly, R. Foster, and T. J. Foster. 1989.Roles of alpha-toxin and beta-toxin in virulence of Staphylococcus aureus forthe mouse mammary gland. Infect. Immun. 57:2489–2494.

4. Bubeck Wardenburg, J., R. J. Patel, and O. Schneewind. 2007. Surfaceproteins and exotoxins are required for the pathogenesis of Staphylococcusaureus pneumonia. Infect. Immun. 75:1040–1044.

5. Bubeck Wardenburg, J., and O. Schneewind. 2008. Vaccine protectionagainst Staphylococcus aureus pneumonia. J. Exp. Med. 205:287–294.

6. Callegan, M. C., L. S. Engel, J. M. Hill, and R. J. O’Callaghan. 1994.Corneal virulence of Staphylococcus aureus: roles of alpha-toxin and proteinA in pathogenesis. Infect. Immun. 62:2478–2482.

7. Finelli, L., A. Fiore, R. Dhara, L. Brammer, D. K. Shay, L. Kamimoto, A.Fry, J. Hageman, R. Gorwitz, J. Bresee, and T. Uyeki. 2008. Influenza-associated pediatric mortality in the United States: increase of Staphylococ-cus aureus coinfection. Pediatrics 122:805–811.

8. Frazee, B. W. 2007. Update on emerging infections: news from the Centersfor Disease Control and Prevention. Severe methicillin-resistant Staphylo-coccus aureus community-acquired pneumonia associated with influenza—Louisiana and Georgia, December 2006-January 2007. Ann. Emerg. Med.50:612–616.

9. Gouaux, E., M. Hobaugh, and L. Song. 1997. Alpha-hemolysin, gamma-hemolysin, and leukocidin from Staphylococcus aureus: distant in sequencebut similar in structure. Protein Sci. 6:2631–2635.

10. Granstrom, M., I. Julander, and R. Mollby. 1983. Serological diagnosis ofdeep Staphylococcus aureus infections by enzyme-linked immunosorbentassay (ELISA) for staphylococcal hemolysins and teichoic acid. Scand. J. In-fect. Dis. Suppl. 41:132–139.

11. Julander, I. G., M. Granstrom, S. A. Hedstrom, and R. Mollby. 1983. Therole of antibodies against alpha-toxin and teichoic acid in the diagnosis ofstaphylococcal infections. Infection 11:77–83.

12. Kallen, A. J., J. Brunkard, Z. Moore, P. Budge, K. E. Arnold, G. Fosheim, L.Finelli, S. E. Beekmann, P. M. Polgreen, R. Gorwitz, and J. Hageman. 2008.Staphylococcus aureus community-acquired pneumonia during the 2006 to2007 influenza season. Ann. Emerg. Med. 3:358–365.

13. Kawate, T., and E. Gouaux. 2003. Arresting and releasing staphylococcalalpha-hemolysin at intermediate stages of pore formation by engineereddisulfide bonds. Protein Sci. 12:997–1006.

14. Klevens, R. M., M. A. Morrison, J. Nadle, S. Petit, K. Gershman, S. Ray,L. H. Harrison, R. Lynfield, G. Dumyati, J. M. Townes, A. S. Craig, E. R.Zell, G. E. Fosheim, L. K. McDougal, R. B. Carey, and S. K. Fridkin. 2007.Invasive methicillin-resistant Staphylococcus aureus infections in the UnitedStates. JAMA 298:1763–1771.

15. Kollef, M. H., L. E. Morrow, M. S. Niederman, K. V. Leeper, A. Anzueto, L.Benz-Scott, and F. J. Rodino. 2006. Clinical characteristics and treatmentpatterns among patients with ventilator-associated pneumonia. Chest 129:1210–1218.

16. Kollef, M. H., A. Shorr, Y. P. Tabak, V. Gupta, L. Z. Liu, and R. S. Johannes.2005. Epidemiology and outcomes of health-care-associated pneumonia:results from a large US database of culture-positive pneumonia. Chest 128:3854–3862.

17. Kuehnert, M. J., H. A. Hill, B. A. Kupronis, J. I. Tokars, S. L. Solomon, andD. B. Jernigan. 2005. Methicillin-resistant-Staphylococcus aureus hospital-izations, United States. Emerg. Infect. Dis. 11:868–872.

18. Larinkari, U., and V. V. Valtonen. 1984. Comparison of anti-alpha-haemo-lysin and teichoic acid antibody tests in patients with Staphylococcus aureusendocarditis or bacteraemia. J. Infect. 8:221–226.

19. Lowy, F. D. 1998. Staphylococcus aureus infections. N. Engl. J. Med. 339:520–532.

20. Menzies, B. E., and D. S. Kernodle. 1996. Passive immunization with anti-serum to a nontoxic alpha-toxin mutant from Staphylococcus aureus is pro-tective in a murine model. Infect. Immun. 64:1839–1841.

21. Menzies, B. E., and D. S. Kernodle. 1994. Site-directed mutagenesis of thealpha-toxin gene of Staphylococcus aureus: role of histidines in toxin activityin vitro and in a murine model. Infect. Immun. 62:1843–1847.

22. O’Reilly, M., J. C. de Azavedo, S. Kennedy, and T. J. Foster. 1986. Inacti-vation of the alpha-haemolysin gene of Staphylococcus aureus 8325-4 bysite-directed mutagenesis and studies on the expression of its haemolysins.Microb. Pathog. 1:125–138.

23. Patel, A. H., P. Nowlan, E. D. Weavers, and T. Foster. 1987. Virulence ofprotein A-deficient and alpha-toxin-deficient mutants of Staphylococcus au-reus isolated by allele replacement. Infect. Immun. 55:3103–3110.

24. Rubin, R. J., C. A. Harrington, A. Poon, K. Dietrich, J. A. Greene, and A.Moiduddin. 1999. The economic impact of Staphylococcus aureus infectionin New York City hospitals. Emerg. Infect. Dis. 5:9–17.

25. Shorr, A. F., Y. P. Tabak, V. Gupta, R. S. Johannes, L. Z. Liu, and M. H.Kollef. 2006. Morbidity and cost burden of methicillin-resistant Staphy-lococcus aureus in early onset ventilator-associated pneumonia. Crit.Care 10:R97.

26. Song, L., M. R. Hobaugh, C. Shustak, S. Cheley, H. Bayley, and J. E.Gouaux. 1996. Structure of staphylococcal alpha-hemolysin, a heptamerictransmembrane pore. Science 274:1859–1866.

27. Walker, B., and H. Bayley. 1995. Key residues for membrane binding, oli-gomerization, and pore forming activity of staphylococcal alpha-hemolysinidentified by cysteine scanning mutagenesis and targeted chemical modifi-cation. J. Biol. Chem. 270:23065–23071.

28. Walker, B., M. Krishnasastry, L. Zorn, and H. Bayley. 1992. Assembly of theoligomeric membrane pore formed by staphylococcal alpha-hemolysin ex-amined by truncation mutagenesis. J. Biol. Chem. 267:21782–21786.

Editor: J. N. Weiser

2718 RAGLE AND BUBECK WARDENBURG INFECT. IMMUN.

on March 27, 2021 by guest

http://iai.asm.org/

Dow

nloaded from