8
ORIGINAL PAPER R. E. Biagini D. L. Sammons J. P. Smith B. A. MacKenzie C. A. F. Striley S. A. Robertson J. E. Snawder C. P. Quinn Simultaneous measurement of specific serum IgG responses to five select agents Received: 10 January 2005 / Revised: 24 February 2005 / Accepted: 8 March 2005 / Published online: 2 June 2005 Ó Springer-Verlag 2005 Abstract Select Agents are defined by CDC and the USDA Animal and Plant Health Inspection Service (APHIS) as biological agents or toxins deemed a threat to public, animal, or plant health, or to animal or plant products. They are classified on the basis of their ease of dissemination, mortality/morbidity rate, and potential for social disruption. A subset of these agents includes Bacillus anthracis,Yersinia pestis, Francisella tularensis, ricin toxin (RT), and staphylococcal enterotoxin B (SEB). Infection or intoxication with these agents has been shown to elicit an antigen-specific serum IgG re- sponse. We describe a fluorescent covalent microsphere immunoassay (FCMIA) for measurement of specific IgG antibodies to seven different antigens from five different select agents; B. anthracis [protective antigen (PA) and lethal factor (LF)], Y. pestis (F1 and V anti- gens), F. tularensis, RT and SEB simultaneously in hu- man B. anthracisvaccinee sera (containing anti-PA and anti-LF IgG) which had been spiked with animal specific IgG antibodies to the other select agents. Inter-assay and intra-assay coefficients of variation were 6.5 and 13.4%, respectively (N=4). There were no significant differences (P>0.70) between assay responses when the assays were performed individually or multiplexed. When the observed versus expected interpolated concentrations were compared, highly linear relation- ships were observed (r 2 values from 0.981 to 0.999, P<0.001). Minimum detectable concentrations (MDC) ranged from 0.3 ng mL 1 (Y. pestis F1) to 300 ng mL 1 (RT). Finally, the curves showed responses were linear for most analytes from their MDC to 125 (SEB) to 1,300 (Y. pestis F1)·their MDC. These data indicate that multiplexed FCMIA is a sensitive and accurate method for simultaneous measurement of specific IgG in serum to CDC select agents and may be of value in screening either decontamination workers or the general population for exposure to/infection with these agents. Keywords Anthrax Tularemia Plague Staphylococcal enterotoxin B Ricin Introduction The Department of Health and Human Services/Centers for Disease Control and Prevention (DHHS/CDC) has classified several agents which may be used in a bioter- rorism attack as select agents [1]. They are classified as such on the basis of ease of dissemination, mortality/ morbidity rate, and potential for social disruption. These agents include viruses, bacteria, rickettsiae, fungi, and toxins. A subset of these agents includes Bacillus anthracis, Yersinia pestis, Francisella tularensis, ricin toxin (RT), and staphylococcal enterotoxin B (SEB). Exposure to, infection with, and vaccination with B. anthracis [2, 3], Y. pestis[46], F. tularensis [7], RT [8] or SEB [9] have been shown to cause production of specific IgG antibodies. Comparison of pre-bioterrorism and post-bioterrorism-incident IgG antibody levels has been shown to be a potentially useful method for biological monitoring of decontamination and clean-up workers for potential exposure to bioterrorism agents [2]. The potential exists for exposure to multiple select agents in a single bioterrorism event [3]. In this paper we describe a multiplexed fluorescent covalent microsphere R. E. Biagini (&) D. L. Sammons J. P. Smith B. A. MacKenzie C. A. F. Striley S. A. Robertson J. E. Snawder Biomonitoring and Health Assessment Branch, Division of Applied Research and Technology, Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health, 4676 Columbia Parkway, Cincinnati, OH 45226, USA E-mail: [email protected] Tel.: +1-513-5338196 Fax: +1-513-5338494 C. P. Quinn Microbial Pathogenesis and Immune Response (MPIR) Laboratory, Centers for Disease Control and Prevention, National Center for Infectious Diseases, Atlanta, GA 30333, USA Anal Bioanal Chem (2005) 382: 1027–1034 DOI 10.1007/s00216-005-3204-6

Simultaneous measurement of specific serum IgG responses to five select agents

Embed Size (px)

Citation preview

ORIGINAL PAPER

R. E. Biagini Æ D. L. Sammons Æ J. P. Smith

B. A. MacKenzie Æ C. A. F. Striley Æ S. A. RobertsonJ. E. Snawder Æ C. P. Quinn

Simultaneous measurement of specific serum IgG responsesto five select agents

Received: 10 January 2005 / Revised: 24 February 2005 / Accepted: 8 March 2005 / Published online: 2 June 2005

� Springer-Verlag 2005

Abstract Select Agents are defined by CDC and theUSDA Animal and Plant Health Inspection Service(APHIS) as biological agents or toxins deemed a threatto public, animal, or plant health, or to animal or plantproducts. They are classified on the basis of their ease ofdissemination, mortality/morbidity rate, and potentialfor social disruption. A subset of these agents includesBacillus anthracis,Yersinia pestis, Francisella tularensis,ricin toxin (RT), and staphylococcal enterotoxin B(SEB). Infection or intoxication with these agents hasbeen shown to elicit an antigen-specific serum IgG re-sponse. We describe a fluorescent covalent microsphereimmunoassay (FCMIA) for measurement of specificIgG antibodies to seven different antigens from fivedifferent select agents; B. anthracis [protective antigen(PA) and lethal factor (LF)], Y. pestis (F1 and V anti-gens), F. tularensis, RT and SEB simultaneously in hu-man B. anthracisvaccinee sera (containing anti-PA andanti-LF IgG) which had been spiked with animal specificIgG antibodies to the other select agents. Inter-assayand intra-assay coefficients of variation were 6.5 and13.4%, respectively (N=4). There were no significantdifferences (P>0.70) between assay responses when theassays were performed individually or multiplexed.When the observed versus expected interpolated

concentrations were compared, highly linear relation-ships were observed (r2values from 0.981 to 0.999,P<0.001). Minimum detectable concentrations (MDC)ranged from 0.3 ng mL�1 (Y. pestis F1) to 300 ngmL�1 (RT). Finally, the curves showed responses werelinear for most analytes from their MDC to 125 (SEB) to1,300 (Y. pestis F1)·their MDC. These data indicatethat multiplexed FCMIA is a sensitive and accuratemethod for simultaneous measurement of specific IgG inserum to CDC select agents and may be of value inscreening either decontamination workers or the generalpopulation for exposure to/infection with these agents.

Keywords Anthrax Æ Tularemia Æ Plague ÆStaphylococcal enterotoxin B Æ Ricin

Introduction

The Department of Health and Human Services/Centersfor Disease Control and Prevention (DHHS/CDC) hasclassified several agents which may be used in a bioter-rorism attack as select agents [1]. They are classified assuch on the basis of ease of dissemination, mortality/morbidity rate, and potential for social disruption.These agents include viruses, bacteria, rickettsiae, fungi,and toxins. A subset of these agents includes Bacillusanthracis, Yersinia pestis, Francisella tularensis, ricintoxin (RT), and staphylococcal enterotoxin B (SEB).Exposure to, infection with, and vaccination with B.anthracis [2, 3], Y. pestis[4–6], F. tularensis [7], RT [8] orSEB [9] have been shown to cause production of specificIgG antibodies. Comparison of pre-bioterrorism andpost-bioterrorism-incident IgG antibody levels has beenshown to be a potentially useful method for biologicalmonitoring of decontamination and clean-up workersfor potential exposure to bioterrorism agents [2].The potential exists for exposure to multiple selectagents in a single bioterrorism event [3]. In this paper wedescribe a multiplexed fluorescent covalent microsphere

R. E. Biagini (&) Æ D. L. Sammons Æ J. P. SmithB. A. MacKenzie Æ C. A. F. Striley Æ S. A. RobertsonJ. E. SnawderBiomonitoring and Health Assessment Branch,Division of Applied Research and Technology,Centers for Disease Control and Prevention,National Institute for Occupational Safety and Health,4676 Columbia Parkway, Cincinnati, OH 45226, USAE-mail: [email protected].: +1-513-5338196Fax: +1-513-5338494

C. P. QuinnMicrobial Pathogenesis and ImmuneResponse (MPIR) Laboratory,Centers for Disease Control and Prevention,National Center for Infectious Diseases,Atlanta, GA 30333, USA

Anal Bioanal Chem (2005) 382: 1027–1034DOI 10.1007/s00216-005-3204-6

immunoassay (FCMIA) for measurement of specificIgG antibodies to B. anthracis(protective antigen [PA]and lethal factor [LF]), Y. pestis (F1 and V antigens[YPF and YPV]), F. tularensis (FT), RT and SEB (BTantigens) simultaneously in human B. anthracis vaccineesera which had been fortified with animal select agent-specific IgG antibodies. The FCMIA method uses5.6 lm diameter microspheres (composed of polysty-rene, divinylbenzene, and methacrylic acid), with surfacecarboxylate functionality for covalent attachment of theBT antigens. Internally the microspheres were dyed withproprietary red and infrared-emitting fluorochromes. Byadjusting the concentrations of each fluorochrome,spectrally addressable microsphere sets were obtained.b-Phycoerythrin labeled secondary antibodies were usedfor detection. A 635 nm classification laser excites andclassifies the red and infrared fluorochromes embeddedin the microspheres. A 532 nm reporter laser excites thegreen fluorescent molecules bound to the surface of themicrospheres. Using this method, concentrations ofnumerous analytes can be measured simultaneously byisolating detector responses (green fluorescence) fromnumerous spectrally isolated microspheres simulta-neously (multiplexing).

Anthrax is primarily a zoonotic disease of animals.The toxicity from B. anthracis exposure occurs as a re-sult of the actions of binary toxins (lethal toxin andedema toxin) produced by the organism [2, 10]. Anthraxmay be contracted as a result of gastrointestinal, cuta-neous, or inhalation exposure to B. anthracis spores [11].There are no data suggesting that patient-to-patienttransmission of anthrax occurs [12]. Y. pestis is thecausative agent of plague. Bubonic plague occurs inhumans when bitten by infected rodent fleas. Inhalationexposure to Y. pestis results in pneumonic plague. Y.pestis contains numerous virulence factors including Vand F1 antigens which inhibit phagocytosis. Modernexperience with person-to-person spread of pneumonicplague is limited and it is also the more difficult toprotect against using vaccines [13]. The causative agentof tularemia, F. tularensis, is one of the most infectiouspathogenic bacteria known, requiring inoculation orinhalation of as few as ten organisms to cause disease[14, 15]. There is no evidence of human-to-humantransmission of tularemia [16]. Ricin toxin (RT) is alectin found in the bean of the castor plant, Riciniscommunis, and is one of the most toxic plant toxinsknown. It consists of two polypeptide chains, the enzy-matically active A-chain and the lectin binding B-chain,linked by a disulfide bond. The A-chain is a ribosome-inactivating enzyme, which is specific for the de-puri-nation of a single adenosine in the 28S subunit ofeucaryotic ribosomal ribonucleic acid (rRNA) with aresulting block in protein synthesis [17]. Although ricin’slethal toxicity is approximately 1000th that of botulinumneurotoxin A toxin, ricin may have significance as abiological weapon because of its heat stability andworldwide availability as a by-product of castor oilproduction [18]. Staphylococcal enterotoxin B (SEB) is

one of seven enterotoxins produced by strains ofStaphylococcus aureus. The staphylococcal enterotoxinsare the most frequent cause of common food poisoning.More severe physiological consequences, for example alife-threatening toxic shock-like syndrome, may resultfrom exposure to these toxins through a non-entericroute. Although high-dose exposures may well causefatalities, it is the incapacitating consequences of inha-lation exposure to agents such as SEB that are of mostconcern in the context of bioterrorism [18].

Data from this study provide proof of concept forFCMIA evaluation of antibody responses to multipleagents and demonstrate the potential utility of thisapproach in a bioterrorism response scenario.

Methods and materials

Chemicals/anti-sera

A mixed-antibody standard anti-sera reagent was pre-pared by combining human anti-anthrax standard serumAVR801 (109.1 lg mL�1 anti-PA IgG) [19], and humananti-anthrax standard serum AVR733 (65.5 lg mL�1

anti-LF IgG and 198 lg mL�1 anti-PA IgG; C.P. Quinnunpublished data), mouse monoclonal anti-Y. pestis F1antigen IgG (Advanced ImmunoChemical, Long Beach,CA, USA; product No. G1-YPF19, 5.7 mg mL�1),rabbit anti-F. tularensis IgG (Becton, Dickinson, StLouis, MO, USA; product No. 240939, 5 mg mL�1),rabbit anti-SEB IgG (Toxin Technology, Sarasota, FL,USA; BS 202, 3 mg mL�1) and mouse monoclonalmouse anti-RT IgG (product No. RDI-TRK2R1-CP75,5 mg mL�1), and monoclonal mouse anti-Y. pestis Vantigen IgG (product No. RDI-TRK3YPV-13;Research Diagnostics, Flanders, NJ, USA; 5 mg mL�1).The use of all human anthrax clinical samples wasapproved by the CDC Human Subjects Review Board.

The detection antibody mixture contained pre-ab-sorbed (mouse, rabbit, and rat) phycoerythrin labeled(PE) goat anti-human IgG (Open Biosystems, Hunts-ville, AL, USA; product No. SAB-1294), PE-goat anti-rabbit IgG obtained (Sigma, St Louis, MO, USA;product No. P-9795), and pre-absorbed (human andrabbit) PE-donkey anti-mouse IgG (Research Diagnos-tics, Flanders, NJ, USA; product No. RDI-715116150).

Antigens used were recombinant PA (product No.171B), and LF (product No. 172B; List Biological Labo-ratories, Campbell, CA, USA), RT (Vector Laboratories,Burlingame, CA, USA; product No. L-1090) and Staph-ylococcal enterotoxin type B (SEB, Toxin Technology;product No. BT 202). F. tularensis antigen was providedby Dr Venkat Venkateswaran, Lawrence Livermore Na-tional Labs, Livermore, CA, USA and YPV and YPFantigens were provided byDrDianeWilliamson, DefenceScience and Technology Laboratory, Porton Down, UK.

The EDC [1-ethyl-3-(3-dimethyl-aminopropyl)carbo-diimide hydrochloride] and N-hydroxysulfosuccinimide,sodium salt (sulfo-NHS) were obtained from Pierce

1028

Biotechnology (Rockford, IL, USA). Buffers used werephosphate-buffered saline (PBS, pH 7.4; Sigma), stor-age/blocking buffer (PBS, 5% dried skim milk (Difco,Sparks, MD, USA), 0.05% NaN3(Sigma), pH 7.4),wash buffer (PBS containing 0.05% Tween 20; Sigma),activation buffer (0.1 mol L�1NaH2PO4, pH 6.2; Sig-ma), dilution buffer (PBS, 5% dried skim milk, 0.05%NaN3, 1% Tween 20, pH 7.4) and coupling buffer(0.05 mol L�12-(N morpholinoethanesulfonic acid,MES; Sigma; pH 5.0).

Preparation of microspheres

The seven BT antigens (PA, LF, YPV, YPF, FT, SEBand RT) were coupled to separate unique sets of car-boxylate-modified (Luminex) microspheres [2, 3]. Briefly,after washing twice with 80 lL activation buffer, sevenindividual sets of spectrally differentiable carboxylatedmicrospheres (1.25·107) were pelleted (5,000g for 2 min)in 1.5 mL centrifuge tubes using a microcentrifuge (Ep-pendorf, Hamburg, Germany). The microspheres wereresuspended by sonication (mini sonicator; Cole Parmer,Vernon Hills, IL, USA), and gentle vortex mixing (Vor-tex Genie, VWR, West Chester, PA, USA) in 80 lLactivation buffer to which 10 lL activation buffer con-taining 50 mg mL�1 EDC and 10 lL activation buffercontaining 50 mg mL�1 sulfo-NHS had been added. Themixture was incubated for 20 min at room temperature.The microspheres were then washed twice in 500 lLcoupling buffer and solutions of the seven BT antigens(40 lg mL�1) in 500 lL coupling buffer were added andthe mixtures incubated for 2 h at room temperature, withgentle shaking. The coupled microspheres were thenwashed twice in 1 mL wash buffer and stored in 0.5 mLstorage/blocking buffer. Microsphere concentrationswere determined using a microscope and hemocytometer.

Preparation of anti-serum

The binding characteristics of the human and animalantibodies to the seven select agent antigens were suchthat different amounts of each were necessary to produceanimal antibody-supplemented human sera. The anti-body mix standard anti-sera contained 3 lL AVR801,8 lL AVR733, 3 lL 1:100 (diluted in PBS) anti-Y. pestisV, 20 lL 1:250 (diluted in PBS) anti-Y. pestis F1, 4 lLanti-F. tularensis, 3 lL anti-SEB and 16 lL anti-ricin fora total mixture of 86 lL which was diluted to 800 lLwith dilution buffer before use. The stock surrogateserum contained 5.9, 6.5, 2.0, 0.6, 25, 12 and100 lg mL�1 anti-PA, anti-LF, anti-YPV, anti-YPF1,anti-FT, anti-SEB and anti-RT, respectively.

Multiplexed and monoplex analyses

The Bio-Plex suspension array used to measure antibodybinding (Bio-Rad Laboratories, Hercules, CA, USA)

was systematically calibrated and validated before eachuse using a maintenance, calibration, and validationplate, and standards supplied by the manufacturer. Toperform an assay, suspensions of the seven microspheres(100 microspheres lL�1) were incubated for 1 h at 37�Cin blocking buffer to minimize non-specific bindingduring subsequent steps. An aliquot (50 lL) of thissuspension [approximately 5000 microspheres, eitherindividually (monoplex) or mixed (multiplexed)] wasadded to the wells of a 1.2 lm filter membrane microt-iter plate (Millipore, Bedford, MA, USA; Part# MA-BVN1250) and the liquid aspirated by use of a vacuummanifold filtration system (Millipore, Part#MAVM09601). The microspheres were then washedthree times with 200 lL wash buffer, each wash followedby vacuum aspiration. To run an experiment, 50 lL (induplicate) of diluted antibody mix standard anti-sera(1:2, 1:4, 1:10, 1:20, 1:40, 1:100, 1:200, 1:400, 1:1000,1:2000, 1:4000, 1:10,000, 1:20,000 and 1:40,000 withdilution buffer) or reagent blank were added to themicrospheres in the wells of the filter membrane mi-crotiter plate and incubated for 30 min at 37�C withshaking. The liquid was vacuum aspirated and the wellswere again washed three times with 200 lL wash buffer.An aliquot (50 lL) of detector mix in dilution buffer(4 lg mL�1 protein for all antibodies) was added to thewells of the plate and incubated for 15 min at 37�C. Thewells were again washed three times with 200 lL washbuffer and resuspended in 100 lL wash buffer. The platewas shaken vigorously for approximately 1 min to dis-perse the microspheres and was placed into the auto-sampler platform of the Bio-Plex suspension arrayinstrument coupled with a 96 well plate autosampler, asdescribed previously [2, 3, 20, 21] using software andsheath fluid supplied by the manufacturer. The instru-ment was programmed to collect data from 100 micro-spheres (classified by their internal fluorescence ratio)and acquire the median fluorescence intensity (MFI) ofthe microsphere-select agent antigen-antisera-anti-IgG-R-PE complex (Fig. 1).

Data analysis

Standard curves were constructed from four-parameterlogistic log fits (4-PL) of MFI against concentration(ng mL�1) of each anti-select agent antigen IgG [2, 3,20, 21] (SigmaPlot; Systat Software, Point Richmond,CA, USA). Duplicate results from individual dilutionswere averaged and concentrations interpolated fromthe standard curves. The FCMIA was run four sepa-rate times. Minimum detectable concentrations (MDC)were calculated from the intersection of the asymptoteof the 4-PL fit’s 95% confidence interval with theregression line [3, 22]. Assessment of the ‘‘goodness offit’’ and the dynamic ranges of the assays were inves-tigated by evaluating the fit of the standards data tothe 4-PL model by ‘‘standards recovery’’, calculated byevaluating interpolated results from each 4-PL fit and

1029

comparing it to the actual dilutions of select agentantigens added to the system [23] by use of the rela-tionship:

Percent recovery ¼ 100

� observed dilution from 4PL fit

expected dilution from 4PL fit

� �

Fig. 1 Diagram of fluorescencecovalent microbeadimmunosorbent assay(FCMIA)

Fig. 2 Four-parameter logisticregression of multiplexedfluorescence covalentmicrobead immunosorbentassay of B. anthracis protectiveantigen and lethal factor, Y.pestis F1 and V antigen, F.tularensis, ricin toxin, andstaphylococcal enterotoxin Bversus concentration (ngmL�1) of anti-sera mix (see textfor details)

1030

The resulting data were analyzed for linearity byregression analysis [21]. The dynamic ranges for eachspecific anti-BT antigen IgG were determined by sys-tematically removing data pairs from both ends of theregression line in such a way as to maximize the r2 value.Friedman’s repeated measure ANOVA was used toevaluate whether there were any differences between theassay’s responses when the analytes were measuredalone or in a multiplex assay. Wilcoxon’s rank sum testwas used to test if the concentrations returned from the

4-PL fits were different from the concentrations added(SigmaStat, Systat). P £ 0.05 was considered statisticallysignificant.

Results

Four-parameter logistic log fits of ng mL�1 anti-BTantigen versus MFI responses are shown in Fig. 2 foranti-PA and LF, YPF and YPV, FT, RT, and SEB. The

Fig. 3 Comparison ofmonoplexed (black fill) versusmultiplexed (hatched fill)fluorescence covalentmicrobead immunosorbentassay median fluorescenceintensities for B. anthracisprotective antigen and lethalfactor, Y. pestis F1 and Vantigen, F. tularensis, ricin toxinand staphylococcal enterotoxinB versus concentration (ngmL�1) of anti-sera mix (see textfor details)

1031

seven curves had excellent relative fits to the 4-PLmodels with mean r2 values of 0.997±0.002 (± standarddeviation). The mean interassay coefficient of variation(CV) was 13.4% (N=840) and the intra-assay CV was6.5% (N=4 independent assays). These values are wellwithin the <10 and <20% intra-assay and inter-assayguidelines for these types of assay [24]. To determinewhether multiplexing the assays had any effect on theirperformance, comparisons of the assays’ responses per-formed individually versus multiplexed were compared(Fig. 3). No significant differences in MFI (Wilcoxon,P>0.700) were observed when results from multiplexingthe assays were compared with performing them indi-vidually. The mean percentage MFI of the multiplexedassays versus the monoplexed assays was 98.7±7.3(±SD). When the interpolated concentrations of anti-BT agent IgG in the surrogate sera from the 4-PLstandard curves were compared with the actual con-centrations of anti-BT agent IgG added, no significantdifferences were observed (Wilcoxon, P>0.700).Regression analysis of observed versus expected ngmL�1 specific anti-toxin IgG yielded a range of r2 valuesfrom 0.981 to 0.999 (mean=0.991±0.005; Table 1).Mean recoveries for all analytes were 103±4.3% (±SD), values which are in the 70–130% range consideredacceptable for these types of assay [23]. The (MDC)ranged from 0.3 ng mL�1 for anti-YPF to 300 ngmL�1 for anti-RT. Linear dynamic ranges for the mul-tiplexed assays ranged from 0.3–400 ng mL�1for YPFto 300–50,000 ng mL�1 for RT (Table 1).

Discussion

According to the US State Department and the USDepartment of Defense, future deployment of biologicalagents by foreign militants or terrorists is likely [25]. TheCDC and USDA have prioritized some of these agents(a subset of which are B. anthracis,Y. pestis, F. tularensis,RT, and SEB) as select agents, because to their ease ofdissemination, mortality/morbidity rate, and potentialfor social disruption [1].

The multiple select agent FCMIA reported on in thisstudy had excellent intra-assay and inter-assay coeffi-cients (6.5 and 13.5%, respectively), values which are wellwithin the <10 and <20% intra-assay and inter-assayguidelines for these types of assay [24], and suggests ex-tremely good precision [26]. The seven anti-BT agent IgGstandard curves were excellent fits to the 4-PL modelswith mean r2 values of 0.997±0.002 (± standard devi-ation). When these models were used to predict theamount of each of the anti-BT agent IgG added, highlysignificant linear relationships were observed. Meanrecoveries for all analytes were 103±4.3% (± SD),values which are in the 70–130% range consideredacceptable for these types of assay [23]. Dynamic rangesfor measurement of each of the antibodies in the surro-gate sera ranged from 0.3–400 ng mL�1 for YPF to 300–50,000 ng mL�1 for RT. TheMDC ranged from 0.3 ngmL�1 for YPF to 300 ng mL�1 for RT. The MDC anddynamic ranges are in the general range of those reportedin human/animal sera after bioterror attacks [27], naturalinfections, or vaccinations [4, 5, 28, 29], although, occa-sionally, serum titers are reported rather than MDC,which makes direct comparison somewhat difficult.

ELISA, although quantitative, precise, sensitive, andaccurate [26], are designed to measure one IgG (or other)analyte per assay. Numerous ELISA have been de-scribed for measuring anti-BT agent IgG antibodies [2–9, 30–32]. The ELISA methods used are very similar,incorporating coating, blocking, analyte binding, sec-ondary antibody incubation, and substrate addition. Ingeneral these steps take approximately 4 h, althoughoccasionally 24 h incubation is involved [33], yielding 96data points per plate (96-well microtiter plate). The totaltime necessary to perform the FCMIA in this work,from the start of incubation to the end of data collectionis 3 h, but yields 672 data points per 96-well microtiterplate. As the level of multiplexing is increased, thethroughput is increasingly elevated. For example, wehave reported the simultaneous measurement of 24pneumococcal polysaccharide vaccine serotypes usingFCMIA for a total throughput of 2304 analytes per 96-well microtiter plate [20].

Table 1 Minimum detectable concentrations (MDC) and dynamic ranges for a multiplexed fluorescence covalent microsphere immu-noassay (FCMIA) for specific anti-IgG to B. anthracis (protective antigen and lethal factor), Y. pestis (V and F1 antigens), F. tularensis,Ricin toxin, and Staphylococcal enterotoxin B

Antigen MDCa

(ng mL�1 IgG)R2valuesb (observed versusexpected regressions)

Dynamic ranges(ng mL�1 IgG)

B. anthracis protective antigen 3 0.997 3�1,500B. anthracis lethal factor 11 0.998 11�3,250Y. pestis V 1 0.999 1�1,000Y. pestis F1 0.3 0.981 0.3�400Ricin toxin 300 0.981 300�50,000Staphylococcal enterotoxin B 12 0.998 12�1,500F. tularensis 6 0.996 6�5,000aMinimum detectable concentrationbDerived from linear regression of interpolated results from each 4-PL fit and comparing it with the actual dilutions of select agentantigens added to the system

1032

The value of agent-specific IgG antibody levels asretrospective diagnostic or epidemiologic tools foridentification of bioterror agent attacks has been dem-onstrated [2, 27]. Serology was particularly important indiagnosing cases of cutaneous anthrax after the 2001bioterror attacks, because for 3 of 10 cutaneous anthraxcases serology was the only positive laboratory test re-sult obtained [34]. In an outbreak setting, when time-sensitive or treatment-sensitive diagnostic methods (e.g.culture, polymerase chain reaction, and biopsy) may notbe applicable, serological testing may be the onlyconfirmatory diagnostic tool that can be used [27].Multiplex analyses have particular significance in thepost-bioterrorism incident scenario, where multiplechemical and/or biological agents may be used simul-taneously or sequentially. Multiple simultaneous out-breaks of different diseases have been proposed as oneepidemiologic indication of the intentional use ofbiological agents [35]. This scenario will result in a needfor rapid, accurate, and precise detection of numerousbiological markers of infection or exposure specific to arange of biological and chemical agents.

When the number of agents and samples to be mea-sured becomes large, the shortcomings of ELISA indi-vidual agent assay could tax our national logistic andassay capacity for diagnostic serological assays.Although other technologies for serodiagnosis areemerging [36, 37], these are not yet widely available. Inthis report we provide proof of principle that CDC selectagent-specific antibodies against F. tularensis, RT, andSEB and for two virulence factors from each of B. an-thracis and Y. pestis can be measured simultaneously ina multiplexed FCMIA analysis. All of these agents havebeen shown to have the capacity to be immunogenic [2–9]. At present, numerous public health laboratories havethe necessary equipment and knowledge to performELISA testing for exposure to/infection with biologicalagents of terror. Incorporating multiplexed FCMIAtechnology into an emergency response strategy mayimprove our capacity to respond to future, possiblymore complex, bioterrorism events.

Acknowledgements Mention of a product or company name doesnot constitute endorsement by CDC. This work was supported inpart by an interagency agreement between NIOSH and NIEHS(Y02ES10189).

References

1. DHHS (2002) Possession, use, and transfer of select agents andtoxins; interim final rule. Fed Regist 240(67):76866–76905

2. Biagini RE, Sammons DL, Smith JP, Page EH, Snawder JE,Striley CA, MacKenzie BA (2004) Determination of serum IgGantibodies to Bacillus anthracis protective antigen in environ-mental sampling workers using a fluorescent covalent micro-sphere immunoassay. Occup Environ Med 61(8):703–708

3. Biagini RE, Sammons DL, Smith JP, MacKenzie BA, StrileyCA, Semenova V, Steward-Clark E, Stamey K, Freeman AE,Quinn CP et al (2004) Comparison of a multiplexed fluorescentcovalent microsphere immunoassay and an enzyme-linkedimmunosorbent assay for measurement of human immuno-

globulin G antibodies to anthrax toxins. Clin Diagn LabImmunol 11(1):50–55

4. Splettstoesser WD, Grunow R, Rahalison L, Brooks TJ,Chanteau S, Neubauer H (2003) Serodiagnosis of human pla-gue by a combination of immunomagnetic separation and flowcytometry. Cytometry 53A(2):88–96

5. Splettstoesser WD, Rahalison L, Grunow R, Neubauer H,Chanteau S (2004) Evaluation of a standardized F1 capsularantigen capture ELISA test kit for the rapid diagnosis of pla-gue. FEMS Immunol Med Microbiol 41(2):149–155

6. Thullier P, Guglielmo V, Rajerison M, Chanteau S (2003)Short report: Serodiagnosis of plague in humans and rats usinga rapid test. Am J Trop Med Hyg 69(4):450–451

7. Valipour A, Koller H, Kreuzer A, Kossler W, Csokay A,Burghuber OC (2003) A case of primary tularemic pneumoniapresenting with necrotizing mediastinal and hilar lymph nodes.Wien Klin Wochenschr 115(5–6):196–199

8. Maddaloni M, Cooke C, Wilkinson R, Stout AV, Eng L,Pincus SH (2004) Immunological characteristics associatedwith the protective efficacy of antibodies to ricin. J Immunol172(10):6221–6228

9. Coffman JD, Zhu J, Roach JM, Bavari S, Ulrich RG,Giardina SL (2002) Production and purification of a re-combinant Staphylococcal enterotoxin B vaccine candidateexpressed in Escherichia coli. Protein Exp Purif 24(2):302–312

10. Swartz MN (2001) Recognition and management of an-thrax—an update. N Engl J Med 345(22):1621–1626

11. Inglesby TV, Henderson DA, Bartlett JG, Ascher MS, EitzenE, Friedlander AM, Hauer J, McDade J, Osterholm MT,O’Toole T et al (1999) Anthrax as a biological weapon: medicaland public health management. Working Group on CivilianBiodefense. Jama 281(18):1735–1745

12. Pile JC, Malone JD, Eitzen EM, Friedlander AM (1998) An-thrax as a potential biological warfare agent. Arch Intern Med158(5):429–434

13. Inglesby TV, Dennis DT, Henderson DA, Bartlett JG, AscherMS, Eitzen E, Fine AD, Friedlander AM, Hauer J, Koerner JFet al (2000) Plague as a biological weapon: medical and publichealth management. Working Group on Civilian Biodefense.Jama 283(17):2281–2290

14. Saslaw S, Eigelsbach HT, Wilson HE, Prior JA, Carhart S(1961) Tularemia vaccine study. I. Intracutaneous challenge.Arch Intern Med 107:689–701

15. Saslaw S, Eigelsbach HT, Prior JA, Wilson HE, Carhart S(1961) Tularemia vaccine study II. Respiratory challenge. ArchIntern Med 107:702–714

16. Dennis DT, Inglesby TV, Henderson DA, Bartlett JG, AscherMS, Eitzen E, Fine AD, Friedlander AM, Hauer J, Layton Met al (2001) Tularemia as a biological weapon: medical andpublic health management. Jama 285(21):2763–2773

17. Barbieri L, Baltelli M, Stirpe F (1993) Ribosomes-inactivatingproteins from plants. Biochem Biophys Acta 1154:237–282

18. Ulrich R, Sidell S, Taylor T, Wilhelmsen C, Franz D (1997)Staphylococcal Enterotoxin B And Related Pyrogenic Toxins.In: Sidell F, Takafuji E, Franz D (eds) Medical aspects ofchemical and biological warfare. Washington, DC: Office ofThe Surgeon General; pp 621–630

19. Semenova VA, Steward-Clark E, Stamey KL, Taylor TH Jr,Schmidt DS, Martin SK, Marano N, Quinn CP (2004) Massvalue assignment of total and subclass immunoglobulin G in ahuman standard anthrax reference serum. Clin Diagn LabImmunol 11(5):919–923

20. Biagini RE, Schlottmann SA, Sammons DL, Smith JP, Snaw-der JC, Striley CA, MacKenzie BA, Weissman DN (2003)Method for simultaneous measurement of antibodies to 23pneumococcal capsular polysaccharides. Clin Diagn LabImmunol 10(5):744–750

21. Biagini RE, Smith JP, Sammons DL, MacKenzie BA, StrileyCA, Robertson SK, Snawder JE (2004) Development of asensitivity enhanced multiplexed fluorescence covalent micro-bead immunosorbent assay (FCMIA) for the measurement of

1033

glyphosate, atrazine and metolachlor mercapturate in waterand urine. Anal Bioanal Chem 379(3):368–374

22. O’Connell M, Belanger B, Haaland P (1993) Calibration andassay development using the four-parameter logistic model.Chemom Intell Lab Syst 20:97–114

23. Nix B, Wild D (2001) Calibration curve-fitting. In: Wild D (ed)The immunoassay handbook. Nature Publishing Group, NY,pp 198–210

24. Jacobson RH (1998) Validation of serological assays for diag-nosis of infectious diseases. Rev Sci Tech 17(2):469–526

25. Wasserman GM, Grabenstein JD, Pittman PR, RubertoneMV, Gibbs PP, Wang LZ, Golder LG (2003) Analysis of ad-verse events after anthrax immunization in US Army medicalpersonnel. J Occup Environ Med 45(3):222–233

26. Quinn CP, Semenova VA, Elie CM, Romero-Steiner S, GreeneC, Li H, Stamey K, Steward-Clark E, Schmidt DS, MothershedE et al (2002) Specific, sensitive, and quantitative enzyme-linked immunosorbent assay for human immunoglobulin Gantibodies to anthrax toxin protective antigen. Emerg InfectDis 8(10):1103–1110

27. Quinn CP, Dull PM, Semenova V, Li H, Crotty S, Taylor TH,Steward-Clark E, Stamey KL, Schmidt DS, Stinson KW et al(2004) Immune responses to Bacillus anthracis protectiveantigen in patients with bioterrorism-related cutaneous orinhalation anthrax. J Infect Dis 190(7):1228–1236

28. Porsch-Ozcurumez M, Kischel N, Priebe H, Splettstosser W,Finke EJ, Grunow R (2004) Comparison of enzyme-linkedimmunosorbent assay, Western blotting, microagglutination,indirect immunofluorescence assay, and flow cytometry forserological diagnosis of tularemia. Clin Diagn Lab Immunol11(6):1008–1015

29. Campbell DE, Kemp AS (1998) Production of antibodies tostaphylococcal superantigens in atopic dermatitis. Arch DisChild 79(5):400–404

30. Eyles JE, Elvin SJ, Westwood A, Lebutt CS, Alpar HO,Somavarapu S, Williamson ED (2004) Immunisation againstplague by transcutaneous and intradermal application of sub-unit antigens. Vaccine 22(31–32):4365–4373

31. Smallshaw JE, Firan A, Fulmer JR, Ruback SL, Ghetie V,Vitetta ES (2002) A novel recombinant vaccine which protectsmice against ricin intoxication. Vaccine 20(27–28):3422–3427

32. LeClaire RD, Bavari S (2001) Human antibodies to bacterialsuperantigens and their ability to inhibit T-cell activation andlethality. Antimicrob Agents Chemother 45(2):460–463

33. Waag DM, Sandstrom G, England MJ, Williams JC (1996)Immunogenicity of a new lot of Francisella tularensis livevaccine strain in human volunteers. FEMS Immunol MedMicrobiol 13(3):205–209

34. Jernigan DB, Raghunathan PL, Bell BP, Brechner R, BresnitzEA, Butler JC, Cetron M, Cohen M, Doyle T, Fischer M et al(2002) Investigation of bioterrorism-related anthrax. UnitedStates, 2001: epidemiologic findings. Emerg Infect Dis8(10):1019–1028

35. Pavlin JA (1999) Epidemiology of bioterrorism. Emerg InfectDis 5(4):528–530

36. Mezzasoma L, Bacarese-Hamilton T, Di Cristina M, Rossi R,Bistoni F, Crisanti A (2002) Antigen microarrays for serodi-agnosis of infectious diseases. Clin Chem 48(1):121–130

37. Bacarese-Hamilton T, Mezzasoma L, Ardizzoni A, Bistoni F,Crisanti A (2004) Serodiagnosis of infectious diseases withantigen microarrays. J Appl Microbiol 96(1):10–17

1034