8
INFZCTION AND IMMUNITY, Feb. 1976, p. 513-520 Copyright 1976 American Society for Microbiology Vol. 13, No. 2 Printed in USA. Staphylococcal Scalded-Skin Syndrome: Development of a Primary Binding Assay for Human Antibody to the Exfoliative Toxin BILL B. WILEY,* LOWELL A. GLASGOW, AND MARVIN ROGOLSKY Departments of Microbiology and Pediatrics, College of Medicine, University of Utah, Salt Lake City, Utah 84132 Received for publication 21 October 1975 Exfoliative toxin (ET) from a phage group II Staphylococcus aureus strain causing staphylococcal scalded-skin syndrome was purified by electrofocusing. Ampholytes and salts were removed from the final product by column chroma- tography on G-50 Sephadex. Sodium dodecyl sulfate-polyacrylamide gels of the final product yielded a single band upon gel electrophoresis, even when 60 ,ig of protein was placed in the gels. Radiolabeling of the purified toxin with 125I yielded a product that still caused exfoliation of suckling mice, indicating that the toxin was still biologically active. A radioimmunobinding assay was devel- oped and used to test rabbit and human sera for antibodies to exfoliative toxin. Although the maximum percentage of binding was not as high as expected (approximately 40%), it was postulated that either iodination had not been sufficiently vigorous or the toxin had sustained immunological damage. The assay was reproducible and more sensitive than the existing neutralization method and readily applicable to the testing of human sera for exfoliative toxin antibodies. The staphylococcal scalded-skin syndrome is represented by a spectrum of clinical entities that include: a generalized exfoliative dermati- tis (Ritter's disease, toxic epidermal necrolysis), bullous impetigo, and a scarlatiniform rash (15). The majority of clinical reports of the dis- ease have emphasized the occurrence of disease primarily in infants and children (15). Initially this age-related susceptibility also appeared to apply to the mouse model described by Melish and Glasgow (14). Subsequently, a small num- ber of cases has been reported in human adults (3-5, 10, 11, 16, 19). In the majority of these cases in adults, with the exception of those reported by Epstein and co-workers (5), some predisposing factor or underlying disease, such as alcoholism, immunosuppression, or compro- mised cell-mediated immunity was present and may have played a role in the development of the disease. The susceptibility of adult skin to the action of the exfoliative toxin (ET) in both humans (3, 24) and experimental animals (4) has now been demonstrated convincingly. These data suggest that the relatively low inci- dence of dermatological manifestations of phage group II staphylococcal infection may be related to the presence of antibody (antitoxin) that is capable of neutralizing the toxin respon- sible for the exfoliation of the superficial layer of the epidermis. The purpose of this report is to describe the development of a radioimmuno- binding assay for the ET and to present initial data concerning: (i) the frequency of occurrence of antibody to ET in the general population, (ii) the incidence of antibody correlated with age, (iii) the presence of antibody in patients with scalded-skin syndrome, and (iv) the sensitivity of the radioimmunobinding assay compared with the in vitro neutralization assay. MATERIALS AND METHODS Cultures. The ET used in this study was produced in vitro from the UT 0002 strain of Staphylococcus aureus, a previously (17) described phage group II isolate that was grown in heart infusion broth (HIB, Difco) on a gyratory water bath shaker with a gas hood (New Brunswick Scientific Co.). Cultures were grown in 100-ml amounts in 250-ml screw-capped Erlenmeyer flasks. Organisms for seeding were grown on 5% sheep blood agar plates and suspended in HIB to yield an optical density of 0.15 when read at 650 nm against an HIB blank in a Spectronic 20 spectrophotometer. The inoculum per flask was 0.3 ml. The flasks with loosened caps were shaken at 120 rpm at 35 C and flushed thrice daily with 100% CO2 for 10 min. Sera. Specimens of human serum from cases of scalded-skin syndrome were kindly supplied by phy- sicians. The sera from "normal" humans were col- 513 on May 28, 2019 by guest http://iai.asm.org/ Downloaded from

Staphylococcal Scalded-Skin Syndrome: Development of … · causing staphylococcal scalded-skin syndrome was purified by electrofocusing. Ampholytes andsalts were removed fromthe

Embed Size (px)

Citation preview

Page 1: Staphylococcal Scalded-Skin Syndrome: Development of … · causing staphylococcal scalded-skin syndrome was purified by electrofocusing. Ampholytes andsalts were removed fromthe

INFZCTION AND IMMUNITY, Feb. 1976, p. 513-520Copyright C© 1976 American Society for Microbiology

Vol. 13, No. 2Printed in USA.

Staphylococcal Scalded-Skin Syndrome: Development of aPrimary Binding Assay for Human Antibody to the

Exfoliative ToxinBILL B. WILEY,* LOWELL A. GLASGOW, AND MARVIN ROGOLSKY

Departments of Microbiology and Pediatrics, College of Medicine, University of Utah, Salt Lake City, Utah84132

Received for publication 21 October 1975

Exfoliative toxin (ET) from a phage group II Staphylococcus aureus straincausing staphylococcal scalded-skin syndrome was purified by electrofocusing.Ampholytes and salts were removed from the final product by column chroma-tography on G-50 Sephadex. Sodium dodecyl sulfate-polyacrylamide gels of thefinal product yielded a single band upon gel electrophoresis, even when 60 ,ig ofprotein was placed in the gels. Radiolabeling of the purified toxin with 125Iyielded a product that still caused exfoliation of suckling mice, indicating thatthe toxin was still biologically active. A radioimmunobinding assay was devel-oped and used to test rabbit and human sera for antibodies to exfoliative toxin.Although the maximum percentage of binding was not as high as expected(approximately 40%), it was postulated that either iodination had not beensufficiently vigorous or the toxin had sustained immunological damage. Theassay was reproducible and more sensitive than the existing neutralizationmethod and readily applicable to the testing of human sera for exfoliative toxinantibodies.

The staphylococcal scalded-skin syndrome isrepresented by a spectrum of clinical entitiesthat include: a generalized exfoliative dermati-tis (Ritter's disease, toxic epidermal necrolysis),bullous impetigo, and a scarlatiniform rash(15). The majority of clinical reports of the dis-ease have emphasized the occurrence of diseaseprimarily in infants and children (15). Initiallythis age-related susceptibility also appeared toapply to the mouse model described by Melishand Glasgow (14). Subsequently, a small num-ber of cases has been reported in human adults(3-5, 10, 11, 16, 19). In the majority of thesecases in adults, with the exception of thosereported by Epstein and co-workers (5), somepredisposing factor or underlying disease, suchas alcoholism, immunosuppression, or compro-mised cell-mediated immunity was present andmay have played a role in the development ofthe disease. The susceptibility of adult skin tothe action of the exfoliative toxin (ET) in bothhumans (3, 24) and experimental animals (4)has now been demonstrated convincingly.These data suggest that the relatively low inci-dence of dermatological manifestations ofphage group II staphylococcal infection may berelated to the presence of antibody (antitoxin)that is capable of neutralizing the toxin respon-sible for the exfoliation of the superficial layer

of the epidermis. The purpose of this report is todescribe the development of a radioimmuno-binding assay for the ET and to present initialdata concerning: (i) the frequency of occurrenceof antibody to ET in the general population, (ii)the incidence of antibody correlated with age,(iii) the presence of antibody in patients withscalded-skin syndrome, and (iv) the sensitivityof the radioimmunobinding assay comparedwith the in vitro neutralization assay.

MATERIALS AND METHODSCultures. The ET used in this study was produced

in vitro from the UT 0002 strain of Staphylococcusaureus, a previously (17) described phage group IIisolate that was grown in heart infusion broth (HIB,Difco) on a gyratory water bath shaker with a gashood (New Brunswick Scientific Co.). Cultures weregrown in 100-ml amounts in 250-ml screw-cappedErlenmeyer flasks. Organisms for seeding weregrown on 5% sheep blood agar plates and suspendedin HIB to yield an optical density of 0.15 when readat 650 nm against an HIB blank in a Spectronic 20spectrophotometer. The inoculum per flask was 0.3ml. The flasks with loosened caps were shaken at120 rpm at 35 C and flushed thrice daily with 100%CO2 for 10 min.

Sera. Specimens of human serum from cases ofscalded-skin syndrome were kindly supplied by phy-sicians. The sera from "normal" humans were col-

513

on May 28, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

Page 2: Staphylococcal Scalded-Skin Syndrome: Development of … · causing staphylococcal scalded-skin syndrome was purified by electrofocusing. Ampholytes andsalts were removed fromthe

514 WILEY, GLASGOW, AND ROGOLSKY

lected at the Primary Children's Hospital and Uni-versity of Utah Medical Center, Salt Lake City,from patients who had no known history of scalded-skin syndrome.

Isolation and purification of toxin. Crude toxinwas precipitated from the HIB cultures by additionof aqueous saturated (NH4)2SO4 to a final concentra-tion of 80%. The crude toxin was collected by centrif-ugation after standing at 4 C for 18 h, dissolved indistilled water, and dialyzed against 0.0015 M phos-phate buffer, pH 7.3, to free it of (NH4)2S04. Afterlyophilization, the crude toxin was electrofocused ina glycerol density gradient with 1% carrier ampho-lytes (LKB Instruments Inc.), pH 6 to 8, and thefractions between pH 6 and 7 were collected andpooled. Glycerol and carrier ampholytes were re-moved by column chromatography on Sephadex G-50 (coarse) equilibrated with the 0.0015 M phosphatebuffer. The final product from a single run wasrelyophilized and made up in distilled water to yielda solution with 500 to 1,000 ,ug of protein per ml.

Estimation of protein. The method ofLowry et al.(12) employing Folin reagent was employed to esti-mate the protein content of ET prior to titration insuckling mice or radioiodination.

Polyacrylamide gel electrophoresis. The electro-focused toxin was subjected to electrophoresis in7.5% polyacrylamide gels with sodium dodecyl sul-fate at pH 7.2 by the methods of Seehafer and Weil(20) and Weber and Osborn (21). Gel tubes measured0.5 by 80 mm. The protein was dissociated withmercaptoethanol and sodium dodecyl sulfate prior toelectrophoresis, and the following proteins were runas standards: bovine serum albumin, ovalbumin,alpha chymotrypsinogen, and beta lactoglobulin.Gels 60 cm in length were used, and the electropho-resis was carried out in a Buchler gel electrophoresistank using a Buchler power supply. A constant cur-rent of 5 mA/gel was employed.

Staining of gels. A special procedure developed byFairbanks et al. (6) using Coomassie blue in twoconcentrations was used for staining gels. Gels weredecolorized by diffusion destaining in a Hoefer de-stainer (Hoefer Instruments Co.).

Molecular weight estimation. The method of We-ber and Osborn (21) was used to estimate the molec-ular weight of ET. A bromophenol blue dye markerwas employed, and the ratio of the distance of mi-gration of the dye front to that of the protein bandwas computed (Rf value). Rf as a function of molecu-lar weight was plotted on single-cycle semilog pa-

per, and the best straight line was extrapolated.From the Rf of the ET band, the molecular weight ofthe toxin could be estimated using the standardcalibration curve.

Radioiodination. 1251 was obtained from Amer-sham/Searle and diluted to 1 ACi/1l. in 5.5 ,uM KI bythe method of Marchalonis (13). Lactoperoxidasewas obtained from Sigma Chemical Co. and made up

fresh each time. A sample protocol was as follows:purified ET, 600 .Lg in 200 Al of 0.0015 M phosphatebuffer; lactoperoxidase 8 ,ug in 80 Al of phosphate-buffered saline (PBS), pH 7.4; 125I (1 ,uCi/ul), 50Ml in 5.5 MuM KI and H202 (freshly diluted), 0.88 mMin PBS, 80 ,l. The above mixture was allowed to

stand at room temperature for 10 min; then an addi-tional 100 ,ul of the 0.88 mM H2O2 in PBS was addedwith mixing. After an additional 10 min, the addi-tion of H202 was repeated and the mixture wasallowed to stand again for 10 min. The reaction wasfinally stopped by the addition of 4 ml of 5 mMcysteine hydrochloride in PBS. Unbound iodine wasremoved by column chromatography on Bio-Gel P-2(Bio-Rad Laboratories) equilibrated with 0.0015 Mphosphate buffer, pH 7.3."Normal" human serum. It is always difficult to

define what is normal with respect to an organismas ubiquitous as S. aureus. However, a human se-rum (obtained from B.W.) that, when diluted 1:40 inTS-BSA buffer [0.15 M NaCl, 0.1 M tris(hydroxy-methyl)aminomethane), 1 mg of bovine serum albu-min per ml, pH 7.41, bound less than 10% of theradioiodinated ET was used. Although the donorhad no history of scalded-skin syndrome, contactwith toxin-producing strains of staphylococci atsome time in the past is highly probable.

Counting of radioactivity. Gamma disintegra-tions were counted on a Beckman Biogammacounter with an 125I isoset. The radioimmunobindingtests were carried out in glass disposable tubes (10by 60 mm) with parafilm caps. This permitted theiraccommodation into the Beckman Biogamma plasticcounting vials. Since the vials were not in contactwith the actual radioactivity, they could be reused.Tests were set up in duplicate, and 50 Al of superna-tant fluid was counted for 20 min in the antigen-binding assays. Controls consisted of antigen plusnormal human serum diluted 1:40 in TS-BSA. Theerror of the count was 3%, and the counting effi-ciency for '25I in the instrument was 86%.

Titration of biological activity of ET. SwissWebster mice less than 5 days old were randomizedand injected intracutaneously in the scapular regionwith 0.05 ml of a dilution of the ET preparationbeing titered. Ten mice per dilution were used, andmice were examined 18 h after injection for evidenceof a positive Nikolsky sign as an indication of exfol-iative activity. Biological activity was expressed asexfoliative units per milliliter or units per milli-gram of protein defined as the reciprocal of thehighest dilution of ET that produced a positive Ni-kolsky sign in 50% of the mice after 18 h.

Immunization. Young, 5- to 7-lb (ca. 2.27 to 3.18kg) New Zealand white rabbits were immunizedwith purified electrofocused ET over a period of 1year. Three initial injections of 100 ,ug each weregiven intravenously; then subsequent injectionswere given intramuscularly in Freund complete ad-juvant until a total of 5 mg of ET protein had beeninjected. Animals were bled weekly from a marginalear vein, and the serum was separated from the clotafter standing at 4 C for 18 h.

RESULTSS. aureus strain UT 0002 regularly yielded

crude toxin preparations with 20,000 exfoliat-ing units/ml when cultured on a shaker in HIBflushed periodically with 100% CO2. Electrofo-cusing of these crude ET preparations usually

INFECT. IMMUN.

on May 28, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

Page 3: Staphylococcal Scalded-Skin Syndrome: Development of … · causing staphylococcal scalded-skin syndrome was purified by electrofocusing. Ampholytes andsalts were removed fromthe

STAPHYLOCOCCAL SCALDED-SKIN SYNDROME 515

yielded a preparation with 2,000 to 10,000units/ml or 3,000 to 10,000 units/mg of protein.The mean molecular weights of the focusedpreparations were 32,400 ± 3,100, based on 10different preparations. Polyacrylamide gel elec-trophoresis of the focused preparations inamounts of 15, 30, or 60 ,ug of protein regularlyexhibited single bands after staining withCoomassie blue (Fig. 1). Since only a singleprotein appeared to be present, even when 60jpg of ET protein was electrophoresed, an at-tempt was made to radioiodinate the ET. Aradioactive ET with a specific activity of 0.013,LCi/,ug was obtained by using the procedure ofMarchalonis (13). A variety of precipitatingagents were tried to test the precipitability of

the labeled ET to provide us with data concern-ing the quantity of unbound label. Our firstexperiments indicated only about 40% of thelabeled ET could be precipitated by 10 or 20%aqueous trichloroacetic acid, even though thelabeled ET had been subjected to column chro-matography to free it of unbound 125I. A varietyof precipitants were added to portions of ET,and the mixtures were left at 4 C for 18 h,followed by centrifugation for 30 min at 4 C at2,500 x g in a Beckman J-21 refrigerated cen-trifuge with a JS-13 rotor. A 100-pu1 portion ofthe supernatant fluid from the precipitationswas removed and counted in a Biogammacounter. The count of the entire supernatantfluid was obtained by multiplying by a factor to

FIG. 1. Coomassie blue-stained sodium dodecyl sulfate-polyacrylamide gels (left to right) of 15, 30, and 60pg of isoelectrofocused ET isolated from cultures of S. aureus UT 0002.

VOL. 13, 1976

on May 28, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

Page 4: Staphylococcal Scalded-Skin Syndrome: Development of … · causing staphylococcal scalded-skin syndrome was purified by electrofocusing. Ampholytes andsalts were removed fromthe

516 WILEY, GLASGOW, AND ROGOLSKY

convert to total volume (usually 2, since thetotal volumes were 200 ADl). The count of thesediment was calculated by subtracting the to-tal supernatant fluid count from the total countof the tube. The percent precipitability was cal-culated as the percentage of the total count inthe sediment. The results are shown in Table 1.Although the 10 or 20% trichloroacetic acidfailed to precipitate the majority of the labeledprotein, 95% ethyl alcohol containing 0.02 Mbarium acetate precipitated 86% of the labeledET, indicating the likelihood that most of the1251 was bound to ET.

Initial development of the primary antigen-binding test was carried out employing rabbitanti-ET serum. To make the test highly sensi-tive, antigen with a count of approximately 33counts/min per piL was used in 25-jul quantitiesand counted for 20 min. This yielded antigencontrol counts of approximately 8,000 counts/50gl per 20 min. Antigen was diluted in the TS-BSA, and the rabbit anti-serum was diluted innormal rabbit serum 1:40 in TS-BSA. Initialtests using goat anti-7S rabbit globulin (Hy-land) produced disappointingly low titers andwere expensive to carry out since the antiglobu-lin had to be used at a low dilution. Subse-quently, it was ascertained that the Farr (7)technique could be utilized to precipitate theantigen-antibody complexes. Preliminary ex-periments showed that addition of aqueous, ice-cold saturated (NH4)2S04 to a final concentra-tion of 50% precipitated the bound, but not theunbound, ET. The antigen-binding assay wasset up as follows: rabbit antiserum dilutions in1:40 normal rabbit serum, 25 jul, and '251-la-beled ET diluted to 33 counts/! l per min in TS-BSA, 25 ,ul, were added to disposable glasstubes (10 by 60 mm), and the reactants weremixed on a Vortex mixer. The tubes werecapped with parafilm and incubated at 37 C for2 h, followed by 4 C for 18 h. To each tube wasthen added 50 Al of ice-cold, aqueous saturated

TABLE 1. Precipitability of '2I-labeledET by proteinprecipitants

Precipitant" Ratio (vol/vol) of ET precipitableET/precipitant

10% TCA 1:1 19.8720% TCA 1:1 20.625% PCA 1:1 11.225% PTA 1:1 33.44

Acetone 15:1 52.2895% ETOH + 20:1 86.20.02 M BaAc" TCA, Trichloroacetic acid; PCA, perchloric acid;

PTA, phosphotungstic acid; ETOH + 0.02 BaAc,0.02 M barium acetate in 95% ethyl alcohol.

(NH4)2SO4, which was kept in an ice bath dur-ing addition. The reactants were mixed asabove and left at 4 C for 18 h. The contents ofthe tubes were again mixed on a Vortex mixerand centrifuged at 2,500 x g at 4 C in a Beck-man J-21 centrifuge with a JS-13 rotor. Aftercentrifugation, 50 ,ul of the supernatant fluidfrom each tube was removed with a microliterpipette, placed in a clean, disposable tube (10 by60 mm) capped with parafilm, placed in a Bio-gamma counting vial, and counted for 20 minas described above. In addition to tube blanksfor determination of the background count, an-tigen controls with normal rabbit serum 1:40instead of antiserum were included with eachserum. Percent binding was calculated fromthe following formula devised by Brunner andChanock (2): Sc - St/Sc x 100 = % binding,where Sc is the corrected count of a controlsupernatant fluid without antiserum and St isthe corrected count of a test supernatant fluidwith antiserum. The results of a titration on asample of rabbit antiserum are shown in Fig. 2.Maximum binding was 42% at a final serumdilution of 1:8, but significant binding (10% orhigher) occurred at a final serum dilution of1:1,024.The primary antigen-binding test seemed ap-

plicable to human serum with only slight modi-fication. Volumes were the same as in the rab-bit antiserum test, but serum was diluted in a1:40 dilution of "normal" human serum in TS-BSA. Serum from one of us (B.W.) to be used asthe diluent was first tested, using TS-BSA as adiluent. This serum bound antigen at a dilutionof 1:32 (13.26%) but not at 1:40 (<10% binding).The results of tests done on a group of 11 hu-mans are shown in Table 2. Patient S.A. was ahealthy, normal graduate student who volun-tarily injected himself intradermally with 100mouse exfoliating units of ET. A description ofthe progression of events after this injection

60o

z

o -

40

atrt20-

2 4 8 16 32 64 128 256 512 iO24DILUTION OF RA881T SERUM (Rep,Focw)

FIG. 2. Percent 2 I1-labeled ET bound by 25 p1 ofrabbit hyperimmune serum. Bound antigen precipi-tated by addition of saturated (NHJ)2SO4.

INFECT. IMMUN.

I

on May 28, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

Page 5: Staphylococcal Scalded-Skin Syndrome: Development of … · causing staphylococcal scalded-skin syndrome was purified by electrofocusing. Ampholytes andsalts were removed fromthe

STAPHYLOCOCCAL SCALDED-SKIN SYNDROME

TABLE 2. Primary antigen-binding titers of sera from patients with staphylococcal scalded-skin syndrome

Staphylococcal Primary antigen binding ti-Patient Age Underlying disease scalded-skin syn- ter(serum)

drome confirmed t

S.A. 27 yr None + 16 Convalescent phaseO.W. 39 yr Acute myelogenous leukemia + 4 Acute-phaseBaby L 3 mo None + 128 Convalescent phasecMother L NAd None - 128 Presumably normalBaby C 1 mo None + 8 Convalescent phasecMother C NA None - 32 Presumably normalA.C. 14 days None + 32 Acute-phaseR.W. 19 yr Chronic glomerular nephritis + 16 Acute-phaseL.C. 54 yr Uremia + 8 Acute-phaseL.C. 54 yr Uremia + 8 Convalescent phaseF.T. 70 yr Diabetes + 256 Convalescent phaseMrs. T. 55 yr Lymphosarcoma + 512 Convalescent phase

" Staphylococcal scalded-skin syndrome was confirmed by isolation of S. aureus in phage group II, whichcaused exfoliation in suckling mice.

"1Titers expressed as the reciprocal of the highest final dilution of antiserum binding at least 10% of theradiolabeled ET. Antigen considered to dilute initial dilutions of antiserum twofold.

Presumably convalescent serum; precise dates not available.( NA, Not available.

was previously described (24). Unfortunately,in most instances only a convalescent serumwas available for testing. Ages of the patientsvaried from 1 month to over 70 years. Titers areshown as the reciprocal of the highest finaldilution of serum that bound at least 10% of theradiolabeled ET. It was considered that antigenfurther diluted the initial dilution of antiserumtwofold. A group of sera from 64 randomly cho-sen patients of all ages from less than 1 year to69 years was next tested by the primary anti-gen-binding test. The results of these assaysare shown in Fig. 3. They ranged from serawith titers of less than 4 to one that had a titerof 128. No particular age distribution of eithernegatives (titers less than 4) or positives (titersof 4 or higher) can be discerned, and there is noevidence of increasing titers with increasedage, as might be expected if repeated subclini-cal colonization with toxin-producing strains ofstaphylococci provides antigenic stimulation. Itis evident, however, that 73% of individualshave some antibody to ET in their serum. Someinsight into the relationship between the bind-ing titer and neutralizing ability of the serumcan be obtained from considering the followingexperiment. ET was diluted in PBS to either 10or 20 units/0. 1 ml and mixed with an equalvolume of undiluted antiserum or varying two-fold dilutions of antiserum in PBS. The mix-tures were incubated at 37 C for 2 h and thenstored at 4 C in the refrigerator for 48 h. Thenthey were centrifuged at 2,500 x g at 4 C, andthe supernatant fluids were removed and in-jected into groups of 10 suckling mice (less than5 days old) per dilution. A control consisting of

ANTIBODY TO EXFOLIATIVE TOXIN IN HUMANS1281

64-

32-

'a": 16l-

8- *

4-

less fthn 4- 0

I0 * 0I 0 0

I

II. 1 . 0

Ia* I 000 @00

II I

less han 10 12 14 16 18 20oev20AGE n yeorsFIG. 3. Scattergram of distribution of ET anti-

body titers determined by the primary antigen-bind-ing assay test among 64 randomly chosen subjects.Titers are reported as reciprocals of the final dilutionof serum binding at least 10% of the 1251-labeled ET,considering the antigen as diluting the antiserum anadditional twofold increment.

the same amounts of ET plus PBS, incubatedand treated as above, was also injected into agroup of 10 mice. Each mouse received 0.05 mlintradermally in the scapular area and wasexamined 18 h later for evidence of a positiveNikolsky sign, indicating scalding and, there-fore, failure of the antiserum to neutralize theET. The reciprocals of the highest final serumdilution protecting 50% of the mice from scald-ing were considered to be the titers. The resultsof these tests carried out with three human seraand two rabbit hyperimmune sera are shown inTable 3. The first two sera were from S.A., whoinjected ET into himself intradermally, and a"tnormal" human (B.W.) and were tested with afinal concentration of 2.5 ET units. The lastthree sera were from a normal human (K.O.)

VOL. 13, 1976 517

* * * * a

on May 28, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

Page 6: Staphylococcal Scalded-Skin Syndrome: Development of … · causing staphylococcal scalded-skin syndrome was purified by electrofocusing. Ampholytes andsalts were removed fromthe

518 WILEY, GLASGOW, AND ROGOLSKY

TABLE 3. Comparisonofneutralizationandantigen-binding titers of human and rabbit sera

Seruma/ ET unitsb Titerc Antitoxind Primarysubject (units/ml) bindingteassay titere

1. S.A. 2.5 2 100 162. B.W. 2.5 5 250 643. K.G. 5 4 400 644. R3 5 16 1,600 1,0245. R4 5 16 1,600 1,024

" Sera 1, 2, and 3 are of human origin, and sera 4and 5 are rabbit sera.

b Units of ET neutralized by 0.05 ml of serum.c Reciprocal of the final dilution of serum that

protected 50%O of the suckling mice from exfoliationby the indicated units of ET.

d Antitoxin units per milliliter of undiluted se-rum, considering the actual amount of diluted se-rum required to protect 50% of the suckling micefrom scalding; i.e., 0.05 ml ofserum 1 neutralized 2.5units of ET when diluted 1:2; hence, 0.02 ml neu-tralized 1 unit and, therefore, there were 50 units ofantitoxin/ml of 1:2 or 100 units/ml of undiluted se-rum.

e Reciprocal of the final dilution of serum bindingat least 10% of the radioiodinated ET in the radioim-munobinding test.

and two hyperimmune rabbits (R3 and R4) un-der immunization for over 1 year with electrofo-cused ET. The last three sera were tested witha final concentration of 5 ET units. There wasgood correlation between neutralizing titers,binding titers, and antitoxin content of thesera. The least sensitive method of measuringanti-ET was in the in vitro neutralization test.The primary antigen-binding test using 125I1labeled ET was much more sensitive as a detec-tor of anti-ET. Differences in antitoxin contentof the sera reflect the differing amounts of ETused in the neutralization tests; thus, "normal"human serum K.O. neutralized 5 units ofET ata dilution of 1:4 and, hence, it contained moreantitoxin than serum B.W., which neutralizedonly 2.5 units of ET.

DISCUSSIONIt appears that we successfully radioiodi-

nated the purified ET used in the work de-scribed here. An experiment to assure thatbinding of '25I-labeled ET was specific wasfirst carried out using purified human Fc frag-ment from immunoglobulin G. It is known thatprotein A in S. aureus non-specifically com-bines with the Fc portion of human or rabbitimmunoglobulin G so, if our ET preparationhad contained contaminating protein A, wemight have obtained nonspecific binding of the

INFECT. IMMUN.

labeled ET. There was no evidence ofbinding ofour '25I-labeled ET preparation by purified Fcfragment; hence this possibility was ruled out.Inhibition tests further revealed that the reac-tion was specific by showing that 372 ng ofunlabeled ET significantly inhibited the reac-tion between 125I-labeled ET and rabbit anti-body to ET.

Patients with confirmed staphylococcalscalded-skin syndrome do have detectable anti-body to staphylococcal ET. Two confirmedcases, F.T. and Mrs. T, had elevated titers ofanti-ET compared with most of the other pa-tients in our series. One normal subject (motherL) also had an elevated antibody titer. Cer-tainly, randomly chosen subjects also have an-tibody to ET in amounts comparable to many ofthe patients with confirmed staphylococcalscalded-skin syndrome. In our series of 64 ran-domly chosen subjects, no differences betweenadults and children regarding antibody to ETwere discernible. Since staphylococcal carriageaverages around 50% (23) and since phagegroup II strains comprise up to 11% of staphylo-coccal strains (22), it is not surprising that ahigh percentage of humans has antibody to ET.The presence of antibodies to ET in the serumof the two babies in our series of patients couldhave been the result of transplacental transferof maternal antibodies since they had amountssimilar to those of their mothers. It was to beexpected that an assay based on use of a radio-active tracer for detecting soluble antigen-anti-body complexes would be more sensitive thanone based on precipitation of the antigen. Theexperience of Brunner and Chanock (2) re-vealed that a radioimmunoprecipitation testwas more sensitive than complement fixation,immunofluorescence, or metabolic inhibitiontests in detecting antibody to Mycoplasmapneumoniae in human sera. Our test was re-producible in that, when sera were tested onseparate days, titers did not differ more thantwofold. The increased sensitivity of our pri-mary antigen-binding test was possible becausethe actual amount of protein used in the anti-gen-binding assay was 71.5 ng, which repre-sented one-fourth of a mouse exfoliating unit ofET. The maximum binding of radioiodinatedantigen in our tests was not as high as thatreported by Brunner and Chanock (2). Themaximum binding detected in our system wasin the 30 to 40% range, whereas Brunner andChanock (2) reported binding of 80% of theirlabeled mycoplasma antigen. Previously we re-ported the possibility of the ET being a glyco-protein (18). A possible explanation for thelower binding figures we obtained may be as-

on May 28, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

Page 7: Staphylococcal Scalded-Skin Syndrome: Development of … · causing staphylococcal scalded-skin syndrome was purified by electrofocusing. Ampholytes andsalts were removed fromthe

STAPHYLOCOCCAL SCALDED-SKIN SYNDROME

cribable to the glycoprotein nature of our anti-gen. Graham (8) described the solubility of atypical glycoprotein, fetuin, found in fetal calfserum. His purified glycoprotein was soluble in35% (NH4)2SO4, 7% trichloroacetic acid, and 1.2M perchloric acid and was precipitated only by95% ethyl alcohol that was 0.02 M with respectto barium acetate. It is possible that antigen-antibody complexes near the end point may nothave precipitated in the (NH4)2SO4 at a finalconcentration of 50%, thus lowering the bind-ing figures. Binding of less than 10% was notconsidered significant in our assay.On the other hand both the labeled ET and

the antiserum may be responsible for the lowbinding capacity in our system. If the quantityofantibody is limited and if only a portion of theET molecules are actually labeled, then onecould account for the binding percentages weobtained on the basis of competition of labeledand unlabeled ET for the limited antibody bind-ing sites. Alternatively, the labeling proceduremay have immunologically altered the ET mol-ecules in such a way that only a portion ofthemnow react with the antibody produced againstpurified unlabeled ET. Bolton and Hunter (1)reported loss of immunoreactivity of human-luteinizing hormone when labeled by the chlor-amine T procedure of Greenwood et al. (9), ascompared with a labeling procedure they devel-oped in which the protein being labeled is notdirectly exposed to the 125I solution or reagentsused in the iodination. We tried to more effi-ciently trap the antigen-antibody complexes byemploying normal serum at 1:8 instead of 1:40,reasoning that a more voluminous precipitateof globulins after addition of the saturated(NH4)2SO4 would more efficiently trap the com-plexes. The higher concentration of normal se-rum, however, did not result in higher bindingof labeled ET.From a consideration of the specific activities

of radiolabeled antigens employed by others (1,2, 9), it must be concluded that our radioiodi-nated antigen had low specific activity. With-out an amino acid analysis, we do not knowwhether this low specific activity was the resultof insufficient iodination or a low content oftyrosine and histidine in the ET. We hope tosolve this problem by attempting to raise thetiter of antibodies in our animals and to in-crease the specific activity of the ET by utiliz-ing more rigorous iodination.The question of the relationship between an-

tibody detected in the primary antigen-bindingtest and neutralizing antibody was partiallyanswered by the experiment showing thatthere was about a 64-fold reduction in antibody

titer when estimated by a neutralization te-stemploying 2.5 to 5 mouse exfoliating units ofET. We have not yet attempted in vivo neutral-ization of ET by injecting antitoxin first andthen challenging with ET. We have estab-lished, however, that at least a portion of theantibody detected in the antigen-binding ra-dioimmunoassay is neutralizing antibody.As to whether antitoxin plays a definitive

role in host resistance, one must consider thatthe lesions in staphylococcal scalded-skin syn-drome may be at a site distant from the site ofinfection; hence, ET may be transported via thecirculation to epidermal target cells. In otherinstances, e.g., bullous impetigo, the toxin isproduced locally and may act locally. At thepresent time we do not have the capability ofdetecting free ET in plasma or the fluid ofbullae. In our culture system, some phagegroup II staphylococci, especially those with anET plasmid, regularly elaborate over 20,000mouse exfoliating units of toxin per ml of cul-ture in 18 h. One could reasonably postulatethat cultivation in artificial medium representsonly an approximation of the in vivo conditions;hence, large amounts of toxin may be producedin patients with staphylococcal scalded-skinsyndrome. If this postulation is reasonable,then large amounts of antitoxin might be nec-essary to neutralize the ET sufficiently to pre-vent exfoliation. Even our hyperimmune rabbitserum neutralized only 5 units of toxin and adilution of only 1:16. It does not seem likely,therefore, that the ET antibody in the amountsdetected in our neutralization tests could ex-plain host resistance to the staphylococcalscalded-skin syndrome. In a previous paper(24), we showed that immunosuppression didnot affect susceptibility of suckling mice to theisolated ET but that it increased their suscepti-bility to infection by toxigenic strains of S.aureus, allowing development of the scalded-skin syndrome from smaller inocula of orga-nisms. These data indicate that antibacterialfactors clearly are determinants of host resist-ance in the mouse model and suggest that asimilar situation may occur in humans. Half ofthe subjects in the proven cases of staphylococ-cal scalded-skin syndrome had compromisingconditions at the time of development of thedisease. These included immunosuppression,diabetes, presence of neoplastic disease, andpossible immunological immaturity. Otherworkers described cases of staphylococcalscalded-skin syndrome in compromised adults(10, 11, 16, 19). The relatively low incidence ofthis syndrome may be due to failure to recog-nize it in its protean manifestations or to a

VOL. 13, 1976 519

on May 28, 2019 by guest

http://iai.asm.org/

Dow

nloaded from

Page 8: Staphylococcal Scalded-Skin Syndrome: Development of … · causing staphylococcal scalded-skin syndrome was purified by electrofocusing. Ampholytes andsalts were removed fromthe

520 WILEY, GLASGOW, AND ROGOLSKY

requirement that depressed host defenses bepresent to allow initiation of infection in mostadults.

ACKNOWLEDGMENTSThis work was supported by Public Health Service grant

AI 10949 from the National Institute of Allergy and Infec-tious Diseases.We wish to thank K. Maxwell of the Hyland Labs, Los

Angeles, for donating the human Fc fraction used in thiswork.

LITERATURE CITED1. Bolton, A., and W. Hunter. 1973. The labelling of pro-

teins to high specific radioactivities by conjugation toa '25I containing acylating agent. Biochem. J.133:529-539.

2. Brunner, H., and R. Chanock. 1973. A radioimmuno-precipitation test for detection of Mycoplasma pneu-moniae antibody. Proc. Soc. Exp. Biol. Med. 143:97-105.

3. Elias, P. M., P. Firtsch, G. Tappeiner, H. Mittermayer,and K. Wolff. 1974. Experimental staphylococcaltoxic epidermal necrolysis (TEN) in adult humansand mice. J. Lab. Clin. Med. 84:414-424.

4. Elias, P. M., H. Mittermayer, G. Tappeiner, P. Fritsch,and K. Wolff. 1974. Staphylococcal toxic epidermalnecrolysis (TEN): the expanded mouse model. J. In-vest. Dermatol. 63:467-475.

5. Epstein, E., P. Flynn, and R. Davis. 1974. Adult toxicepidermal necrolysis with fatal staphylococcal septi-cemia. J. Am. Med. Assoc. 229:425-427.

6. Fairbanks, G., T. Steck, and D. Wallach. 1971. Electro-phoretic analysis of the major polypeptides of thehuman erythrocyte membrane. Biochemistry10:2606-2617.

7. Farr, R. S. 1958. A quantitative immunochemical meas-ure of the primary interaction between I*BSA andantibody. J. Infect. Dis. 103:239-262.

8. Graham, E. R. B. 1972. Fetuin, p. 718. In A. Gottschalk(ed.), The glycoproteins, part A. Elsevier PublishingCo., Amsterdam.

9. Greenwood, J. C., W. M. Hunter, and J. S. Glover.1963. The preparation of 131I labeled human growthhormone of high specific radioactivity. Biochem. J.89:114-123.

10. Hawley, H. B., and M. D. Aronson. 1973. Scalded skin

INFECT. IMMUN.

syndrome in adults. Letter to the editor. N. Engl. J.Med. 288:1130.

11. Levine, G., and C. W. Norden. 1972. Staphylococcalscalded skin syndrome in an adult. N. Engl. J. Med.287:1339-1340.

12. Lowry, O., N. Rosebrough, A. Farr, and R. Randall.1951. Protein measurement with the Folin phenolreagent. J. Biol. Chem. 193:265-275.

13. Marchalonis, J. J. 1969. An enzymic method for thetrace iodination of immunoglobulins and other pro-teins. Biochem. J. 113:299-305.

14. Melish, M. E., and L. A. Glasgow. 1970. The staphylo-coccal scalded skin syndrome: development of an ex-perimental model. N. Engl. J. Med. 282:1114-1119.

15. Melish, M. E., and L. S. Glasgow. 1971. Staphylococcalscalded skin syndrome: the expanded clinical syn-drome. J. Pediatr. 78:958-967.

16. Reid, L. H., W. L. Weston, and J. R. Humbert. 1974.Staphylococcal scalded skin syndrome; adult onset ina patient with deficient cell-mediated immunity.Arch. Dermatol. 109:239-241.

17. Rogolsky, M., R. Warren, B. B. Wiley, H. Nakamura,and L. A. Glasgow. 1974. Nature of the genetic deter-minant controlling exfoliative toxin production inStaphylococcus aureus. J. Bacteriol. 117:157-165.

18. Rogolsky, M., B. B. Wiley, M. Keyhani, and L. A.Glasgow. 1974. Interaction of staphyloccal exfoliativetoxin with concanavalin A. Infect. Immun. 10:1260-1265.

19. Rothenberg, R., F. S. Renna, T. M. Drew, and D. S.Feingold. 1973. Staphylococcal scalded skin syndromein an adult. Arch. Dermatol. 108:408-410.

20. Seehafer, J., and R. Weil. 1974. Synthesis of polymervirus structural polypeptides in mouse kidney cells.Virology 58:75-85.

21. Weber, K., and M. Osborn. 1969. The reliability ofmolecular weight determinations by dodecyl sulfate-polyacrylamide gel electrophoresis. J. Biol. Chem.244:4406-4412.

22. Wentworth, B. 1963. Bacteriophage typing of the staph-ylococci. Bacteriol. Rev. 27:253-272.

23. Wiley, B. 1963. The incidence of encapsulated staphy-locci and anticapsular antibodies in normal humans.Can. J. Microbiol. 9:27-32.

24. Wiley, B., S. Allman, M. Rogolsky, C. Norden, and L.Glasgow. 1974. Staphylococcal scalded skin syn-drome: potentiation by immunosuppression in mice;toxin-mediated exfoliation in a healthy adult. Infect.Immun. 9:636-640.

on May 28, 2019 by guest

http://iai.asm.org/

Dow

nloaded from