Antibody response to teichoic acid and peptidoglycan in Staphylococcus aureus osteomyelitis

(1)

Antibody

Response

to

Teichoic

Acid

and

Peptidoglycan

in

Staphylococcus

aureus

Osteomyelitis

ELLIOT JACOB,lt* LINDAC. DURHAM,' MICHAEL C. FALK,' TAFFY J. WILLIAMS,'

AND LAWRENCE J. WHEAT23

NavalMedical Research Institute, Bethesda, Maryland 20814,1 and Indiana UniversityMedicalCenter, Wishard

Memorial Hospital,2 and Indianapolis Veterans Administration Hospital,3 Indianapolis, Indiana46202 Received 25 July 1986/Accepted 7 October 1986

An enzyme-linked immunosorbent assay was used to evaluate the immunoglobulin G (IgG) response to

Staphylococcus aureus crude teichoic acid (TA) and peptidoglycan (PG) in both rabbits and patients with

osteomyelitis. Inrabbits withexperimental S. aureusosteomyelitis,elevated levelsofIgGtoTAwerepresent in13/18 (72%) of theserumsamplesobtainedat4and 10 weeks postinfection. Incontrast,only5/18(28%) of

thesesera were foundtobepositive for antibodiestoPG. Ofatotal of 39 patients with confirmed S. aureus osteomyelitis(11acute,28chronic), IgGtoTAwaselevatedin 17(44%), whereas antibodiestoPGwerefound

tobeincreasedinonly 1(3%). Cross-reactingantibodiestoS. aureusTAweredetected in only1/18(6%)of

thepatientswithosteomyelitis caused by organisms other than S.aureus.Thesestudiesindicate that IgGtoTA

is more prevalent thanIgG to PG in patients with staphylococcal osteomyelitis. Although these resultsare

encouraging, a larger number of patients is required for an adequate evaluation ofthe TA enzyme-linked

immunosorbentassayforthediagnosis andmanagementof suspected S. aureusosteomyelitis.

Osteomyelitis, secondary to trauma, surgical prosthesis implantation, oracontiguous focus of infection, iscommon

and may causeserious permanentdisability (9). Despite the

use of agressive antibiotic therapy, acute osteomyelitis is difficult to treat and progresses to the chronic stage in

approximately 15 to 30% of patients (16). Staphylococcus

aureusisthemostcommonetiologicalagentof osteomyelitis and is isolated in about 60% of allcases (29). Although the diagnosis ofosteomyelitis isreadily established inthe

pres-enceofclinical and radiographic evidence ofosseous

infec-tion, identification of the etiologicalagentby bacteriological

cultures is often difficult. In some patients with acute

osteomyelitis, blood cultures may be negative; cultures of

bacteriafromopenwoundsordraining sinustractsareoften unreliable because the organisms isolated may represent

only superficial contaminants rather than the actual

etiolog-icalagentsof thebone infection(34). Moreover, it has been estimated thataspecificbacterial etiology canbeisolated in

only50to 70% ofcases (19).

Becauseofthedifficulties associatedwith theisolationof thecausative agent ofosseous infections by routine

bacte-riological techniques, emphasis has been given recently to thedevelopment of serologicalassaysas anaidtodiagnosis.

Thetechniques used for the serodiagnosis of serious staph-ylococcal infections have included gel diffusion (25, 34), immunoelectrophoresis (11, 25), radioimmunoassay (33,34),

andtheenzyme-linkedimmunosorbentassayELISA(8, 10, 15, 36). The most commonly used staphylococcal antigens

for these assaysare either ultrasonicpreparations of whole

cells, ribitol teichoic (TA) acid, or the peptidoglycan (PG)

moiety of the cell wall. Unfortunately, differences in the

preparation of antigens, as well as in the assays used for

antibody detection, have generated a certain degree of

variability in the datareportedfromlaboratorytolaboratory (22, 31).

*Correspondingauthor.

tPresent address: U.S. Army Institute of Dental Research,

WalterReed ArmyMedicalCenter, Washington, DC 20307-5300.

In a previous investigation, we reported on the

develop-mentofanexperimentalstaphylococcal osteomyelitis model

for rabbits thatclosely resembles theclinical, radiographic,

and histological features associated with this type of

infec-tion in humans(12).Thepurposeofthepresentstudywasto

evaluate theimmunoglobulin G(IgG) responseto S.aureus

crude TA and PG in both rabbits and humans with

osteomyelitis by an ELISA.

MATERIALS ANDMETHODS

Bacterial strains. TheWashington Hospital Center (WHC)

strain of S. aureus, originally isolated from a patient with

mandibularosteomyelitis,wasobtained from James Perry of

the Naval Medical Research Institute, Bethesda, Md.

Cul-tures for inoculation were prepared by growing the

orga-nisms at 37°C for 18 h in Trypticase soy broth (BBL Microbiology Systems, Cockeysville, Md.). The cells were

harvestedat 8,000 x gfor 20 min, washed three times with

sterile 0.85%NaCl,andresuspendedtoafinalconcentration of 5 x 107 CFU/ml. S. aureus Wood 46 ATCC 10832 was

obtained from the American Type Culture Collection, Rockville, Md.

Animals. New Zealand White rabbits (Dutchland

Labora-tories, Denver, Pa.), weighingbetween 3 and 4kg each,were

used. The rabbitswere housed in individualcagesandwere

maintained on a standard laboratory diet. The experiments

described in this paper were conducted according to the

principlessetforth in thecurrentedition of the Guideforthe Care and Use of LaboratoryAnimals (3).

Osteomyelitis model. The method used to create

osteo-myelitishas beenpreviouslydescribed in detail(12). Briefly,

the right midfemur of each anesthetizedrabbitwasexposed

surgically, and a duct was created in the medullary cavity

withan activated high-speed drill. An 18-guage needlewas

used to inject 0.5 ml of a sclerosing agent (3% sodium

tetradecyl sulfate; Elkins-Sinn,Cherry Hill, N.J.)-0.1ml(5 x 106CFU) of S. aureus WHC into the medullary cavity.

The hole in the bonewascappedwith sterile bonewax,and the incision was closed with interrupted sutures. The ani-122

on April 11, 2020 by guest

http://jcm.asm.org/

(2)

mals were observed for 10

weeks,

during

which

biweekly

radiographs

were obtained and serum

samples

were drawn for

antibody

analysis.

The

diagnosis

of

osteomyelitis

was

based upon the criteria established

by

Norden

(17),

which included

radiographic

evidence of

infection,

as well as

positive

S. aureus cultures obtained from the infected fe-murs atthetime ofsacrifice.

Patients. (i) Acute S. aureus osteomyelitis. Each of 11

patients

in this category had

complaints

oflessthan 6 weeks and

roentgenographic

or radionuclide (or both types of

)

evidence of

osteomyelitis,

aswellas

positive

S. aureusbone cultures.

(ii)

Chronic S. aureus osteomyelitis. In this group, 28

patients

had boneinfections forat least6

weeks,

roentgen-ographic

evidence of

osteomyelitis,

and

positive

S. aureus bone cultures.

(iii)Non-S.aureusosteomyelitis.Atotalof 18

patients

were included in this category who had

osteomyelitis

due to

organisms

otherthan S. aureus. These

subjects

included10

patients

with

gram-negative

and 8

patients

with

gram-positive

bone infections.

(iv) Controls. Sera obtained from 50

healthy

marines

during

routine

physical

examinationsserved ascontrolsfor

determining

base-line

IgG

levels to S. aureus

antigens.

These serawereobtained

through

the courtesyofJamesG. Olson ofthe Naval Medical Research

Institute, Bethesda,

Md.

Extraction ofantigens. CrudeTAwasextractedfrom both the WHC andWood46strains ofS.aureus

by

amodification of the

technique originally

described

by

CrowderandWhite

(4).

Briefly,

cells froman 18-hbroth culturewere

harvested,

washedtwice with deionized waterandtwice with acetone, andallowed toair

dry.

The cellswere then

disrupted

for20

min in the presence of

glass

beads with a Bead Beater

shaking

apparatus

(Biospec Products, Bartlesville, Okla.).

The

disrupted

cellswere

passed through

acoarse

sinter-glass

filter,

and the filtrate was

centrifuged

at 13,000 x g for 20

min. The supernatant,

containing

crude TA, was passed

through

a 0.45-,um-pore membrane filter and then lyophi-lized.

Cell walls were

prepared

according

to the method of Peterson et al.

(20).

PG was extracted from the cell walls

with 10% trichloroacetic acid at

60°C

for 90 min. After

centrifugation

at

13,000

x gfor 20

min,

the supernatantwas discarded and the PG

pellet

was washed three times with deionized waterand

lyophilized.

Chemical

analysis

of crude TA and PG.

Carbohydrate

determinationswere

performed by

gas

chromatography

with

myoinositol

used as the internal standard. Samples were

prepared

as described

by

Porter (21). Thermal decomposi-tion of amino sugars was prevented

by

the avoidance of

heating

after the

hydrolysis

step. A Varian 3700 gas

chro-matograph equipped

with CDS-111

integrated

microproces-sor and a 3% SP-2330 coated 100/120 Supelcoport glass

column

(183

cm

by

2 mm) were used for the analysis.

Temperatures

were

programmed

from 170 to210°C over16 min.

Aminoacidanalysiswasperformedafter the

hydrolysis

of

sampleswith 6 N HClin vacuo for22h. For the

determina-tionof

glucosamine

andmuramicacid, sampleswere

hydro-lyzed

undersimilar conditionsbutonlyfor4 h.Analysis was

performed

with a model 4400 amino acid analyzer (LKB

Instruments,

Inc., Rockville, Md.)

Immunodiffusion.Thepresence ofTAin the crudeantigen

preparation

was demonstrated by immunodiffusion by the

Endo-Staph

TA antibody test (Meridian Diagnostics, Inc.,

Cincinnati, Ohio). This commercially available kit includes

preformed agarose plates, a partially purified S. aureus

(Lafferty strain) ribitol TAantigen, anda rabbitanti-ribitol TA serum. A 10-pil amount of the crude (100 ,ug/ml) or

purified (30,ug/ml)TApreparationswas added to the periph-eral wells of agarose plates, and the anti-TA serum was

added to the central well. After48 hat room temperature, theplate was immersed in a 0.85% NaCI solution for 24

h,

dried,andwashed withseveralchanges of deionizedwater.

The plate was stained with 0.5% amido black 10B (Sigma ChemicalCo., St. Louis, Mo.) for 8 min and thendestained

overnight ina solution ofglacial acetic acid, methanol, and water (1:7:2).

ELISA. The microplate modification of the technique originally described by Engvall and Perlman (7)was usedto

detect IgG to crude TA and PG in both rabbit as well as

humansera.Adetaileddescription of the ELISA

methodol-ogy has beenprovided elsewhere (12). The wells of micro-titerplates were coated with either crudeTA (10 ,ugIml) or

PG (50 ,ug/ml) as determined to be optimal by a previous checkerboard titration. All sera were tested at a 1:5,000 dilution, and an alkaline phosphatase-conjugated goat

anti-rabbit IgG or anti-human IgG (Miles Laboratories, Inc.,

Elkhart, Ind.) was each used at a 1:3,000 dilution. Optical density values were recorded spectrophotometrically at a wavelength of 405 nm

(OD405)

after a 30-min enzyme-substrateincubationperiodat25°C. Seraweredesignatedas

positive for antibodiestocrudeTA orPGifthe OD405 values exceededthe meanplustwotimesthe standarddeviation of

the prebleed serumvalues for rabbits or the control serum valuesforhumans.

Specificity studies. To demonstrate that the ELISA was

specific forthedetection ofS. aureusantibodies, absorption studies were performed. Sera from five patients with S.

aureus osteomyelitis were heat inactivated at 56°C for 30 min. A 5-ml amountofa 1:2,500dilution ofeach serum was absorbedwith anequal volume of washed S. aureus Wood

46cellsthat had beenadjusted spectrophotometricallyto an

OD540

of0.25.Theserum-cellsuspensionswerekept at25°C

for 4 h withintermittent agitation every 15 minand centri-fugedat 8,000 x gfor20min. The supernatants,

represent-ingtheabsorbedsera, were thenpassed through a 0.45-,um-poremembrane filterto remove anyresidual bacteria. Both

theabsorbedandunabsorbed sera weretestedfor IgG to S. aureus crude TAby ELISAasdescribed above.

RESULTS

Osteomyelitis model. Radiographic evidence of osteomy-elitis was present in10/13 oftherabbits whose femurswere

inoculated with a combination of sclerosing agent and S.

aureus. Ingeneral, radiographic changes werefirst notedat

approximately4 weeks postinfectionand initially consisted of periosteal elevation and thickening. Because none of

these animals received treatment, the osseous infection progressed in severity, and by 6 to 8 weeks increased

medullary cavity densityand areas of bone destruction were evident. Gross examination of the infected femurs at the time ofsacrificerevealed thepresence of copious amounts of

purulent material within the medullary cavity, as well as extensive reactive bone formation (Fig. 1). S. aureus was isolated from the infectedfemursof 9/10 of theanimals with

radiographic evidence of osteomyelitis.

Chemicalanalysisofantigens. The results of the carbohy-drateandamino acidanalysis of crude TA and PG extracted from S. aureus WHC are presented in Table 1. With respect

on April 11, 2020 by guest

http://jcm.asm.org/

(3)

FIG. 1. Comparison of the infected (A) and noninfected(B)femurs ofarabbit withS. aureusosteomyelitis shown at the time ofsacrifice. Note the narrowing ofthemedullary cavityandextensive reactive boneformation, which arecharacteristicof the osseous infection.

to the crudeantigen preparation, although typical TA com-ponents such as ribose, glucosamine, and alanine were

detected, there was a significant degree of contamination with other amino acids which are not present in purified

preparations of TA (20). The PG preparation, on the other hand, was relatively pure as indicated by the presence of large quantities of cell wall backbone sugars (glucosamine

TABLE 1. Carbohydrate and amino acidanalysis of crudeTA and PGofS.aureus WHC

Amtof carbohydrateoramino acid

Compound (nmol/mg of antigen)

TA PG

Fucose 0.5 2.8

Ribose 88.7 7.5

Glucosamine 58.0 504.8

Galactosamine 0.2 1.1

Mannose 9.5

Galactose 4.6 45.1

Glucose 28.3 22.5

Muramic acid 431

Asparticacid 285 15

Threonine 104 13

Serine 82

Glutamic acid 312 446

Glycine 232 1,860

Alanine 620 791

Valine 192 20

Methionine 14 5

Isoleucine 160 8

Leucine 172 7

Tyrosine 48 3

Phenylalanine 88 4

Histidine 128 47

Lysine 418 502

Arginine 120 26

and muramic acid), tetrapeptide components (alanine, glutamic acid, and lysine), and glycine, which serves to cross-link the tetrapeptide subunits.

Immunodiffusion. The resultsofthe analysis ofthe crude

TAby immunodiffusionareshowninFig.2. Whencrude TA extracted from S. aureus WHC wasallowedto reactagainst

the commercially availableribitolTA in the presence ofan

anti-ribitolTAserum, asingle line of identitywasobserved. Similarly, a line of identity was noted when crude TA

extracted fromtheWHCstrain wasallowedto reactagainst

_)

FIG. 2. Results of immunodiffusion studies showing a line of

identity between the crude TA antigen extracted from S. aureus WHC (well 1) and a partially purified ribitol TA from S. aureus Lafferty (well 2). A single fineofidentity wasalso noted between crudeTAofthe WHC strain(well4) and the Wood 46 strain(well 5).

Rabbit anti-ribitol TA serum was addedtothecenterwell.

on April 11, 2020 by guest

http://jcm.asm.org/

(4)

TABLE 2. Results of ELISA fordetection ofIgGtocrude TAand PG in rabbits with S.aureusosteomyelitis

No.(%)of rabbitspositiveatweek:

Antigen

0 1 4 10

TA O 1(10) 6(67) 7(78)

PG O 1(10) 2(22) 3(33)

aDetermined by OD4,5 value exceeding the mean plus two times the standard deviation of the values for theprebleedsera(week0).

crude TA extracted from S. aureusWood 46. Althoughthe

analysis of the crude TA antigens by polyacrylamide gel electrophoresis revealed the presence ofapproximately 20

different protein staining bands(datanotshown),the results

ofthe immunodiffusion experiments clearly show the pres-ence of ribitolTA inthesepreparations.

Antibody response of rabbits with osteomyelitis. Table 2

shows the results of the ELISA for the detection of IgG antibodiestoS. aureusWHC crudeTAand PG in theseraof

rabbits with osteomyelitis thatwereobtainedthroughoutthe

10-weekexperimentalperiod. The datashowthat at 1 week postinfection, only 1 of 10 animals had an elevated IgG

response to each antigenpreparation. By 10 weeks postin-fection, however,mostoftheanimalswerefoundtoproduce elevated levels of antibodies to the crude TA but three

possessed increased antibody levels to PG. Ofa totalof14

rabbitserathat werepositivebyELISA forIgGtocrudeTA

extracted from S. aureus WHC, 10/14

(71%)

were also positive to TA ofthe Wood46 strain,

indicating

that these

antigens sharecommondeterminants. Elevatedlevels ofIgG

toeither the crude TA orPGwere notdetected inthe three

animals which did not develop radiographic evidence of osteomyelitis.

Antibody response of patients with osteomyelitis. The re-sultsoftheELISAfor detection of

IgG

tocrude TA andPG of S. aureus Wood 46 in patients with osteomyelitis are

shown in Table 3. Elevated levels ofIgG to the crudeTA

antigen were detected in 55% ofthe patients with acute S. aureus and in 39% of those with chronic S. aureus

osteomyelitis. Of a total of 39 patients with S. aureus

osteomyelitis (acute and chronic), 17 were positive for IgG

to the crude TA antigen. In contrast, only 1 of 18 patients with non-S. aureus osteomyelitis was

found

to have an

OD405 value that exceeded the established cutoff.

Whenallof thesesera weretestedfor IgGto S. aureusPG (Table 3), there were virtually no differences noted in the IgG levels of the healthy controlsandthose

of

patientswith

osteomyelitis. The dataindicate that only one patient with chronic S. aureusosteomyelitis was positive for antibodies

to PG. None ofthe patients with eitheracute S. aureus or

TABLE 3. IgGresponseof patients with

osteomnyelitis

tocrude TAandPGof S. aureus Wood46byELISA

No.(%)ofpatientspositiveto Osteomyelitis patient group indicated antigen

TA PG

Healthycontrols 1 (2) 2(4)

Acute S. aureus 6 (55) 0(0)

ChronicS. aureus 11(39) 1(4)

Non-S. aureus 1 (6) 0(0)

aDetermined by OD405 value exceeding the mean plus two times the

standarddeviationofthe values for thehealthycontrolsera.

TABLE 4. ELISAspecificity forIgGtocrude TAantigenafter absorption ofserafrompatients with S.aureusosteomyelitisa

OD405of serumsample

Patientno. % Reduction

Unabsorbed Absorbed

1 0.334 0.033 90.1

2 0.332 0.017 94.9

3 0.277 0.038 86.3

4 0.210 0.070 66.7

5 0.282 0.014 95.1

aSerawereabsorbedwith whole cellsof S.aureusWood 46 andtestedby

ELISAata1:5,000dilution for IgGtocrude TAantigen extracted from the

samestrain(seeMaterialsandMethods).

non-S. aureus osteomyelitis was found to possess signifi-cantly elevated antibodylevels to the PGantigen.

Specificity studies. To demonstrate the specificity of the assay, sera obtained from five patients with S. aureus osteomyelitis were absorbed with whole cells of S. aureus Wood 46 and then were tested by ELISA for IgGto crude TA antigen extracted from the same strain. Absorption of each serum resulted in asignificant decrease in

OD405

values for all sera tested (Table 4). The average reduction for all

sera was 86.6% and in 3/5 of the instances wasgreater than 90%.

DISCUSSION

The effective treatment ofosteomyelitis requires prompt

isolationandidentification of theetiologicagent and

appro-priate selection ofantimicrobial therapy. Thisconcept was

dramaticallyillustrated in thestudy byHarris(9),who noted a 15-foldincrease in the incidence ofchronic osteomyelitis when treatment wasdelayed3 or moredays aftertheonset of the acute phase of the illness. Because the initial treat-ment is crucial for preventing the establishment ofchronic osteomyelitis, the importance of obtaining bacteriological

cultures cannot be overemphasized. Yet despite the most

aggressive diagnostic workup, up to 40% of the cases of osteomyelitisdo notproduceaspecific etiologic microorgan-ism against which treatment can be directed (19). Prompt and accurate identification ofthe causative agent currently

requires bone biopsy for acute osteomyelitis if blood cul-tures arenegative,astheyhavebeenreported to be in nearly

50%of cases(5,28).Sinus tractculturesarenotreliablewith

less than a 50% correlation of sinus tract cultures with microorganisms recovered from bone material (14, 19).

Althoughbonebiopsyrepresentsthe procedureof choicefor

establishinga correct bacterial etiology,the results ofbone

cultureshave been reportedto be negative in 10 to 31% of

cases (5, 28). In patients with osteomyelitis underlying decubitus or diabetic foot ulcers, bone biopsy is rarely

indicated,

because organisms

colonizing

the ulcers usually contaminate the biopsy specimens (32). In addition, bone

biopsy forcultures represents a specialized procedure and mayoccasionallyrequire general anesthesia (14)subsequent

cultures are often required to monitor for reinfection and shifts in sensitivitypatterns, especiallyifthe patientfailsto show a satisfactory response (19). Therefore, the develop-ment of a serological assay for the establishment of a

bacteriological diagnosis in osteomyelitis would be

ex-tremely beneficial.

The serological diagnosisof deep-seated S. aureus infec-tions has been the subject of intensive investigation.

Whereas earlierassays measured antibodies to extracellular

on April 11, 2020 by guest

http://jcm.asm.org/

(5)

S. aureus, products, notably alpha and gamma hemolysins

(13, 23), more recently the focus has shifted toward the detection of antibodies to cell-associated components, in-cludingTA (8, 10, 15,25, 31, 36), PG (2, 26, 27, 30, 35), and cell walls (12). These previous studies have shown that significantly elevated levels of antibodiestoTAorPGcanbe

detected in thevast majority ofpatients withcomplicated S.

aureusbacteremiaas measuredby geldiffusion, ELISA,or

radioimmunoassay. In contrast, it is more difficult to

dem-onstrateanincreasedantibody responsetothese antigens in patients with S. aureus osteomyelitis, especially during the chronic phaseofinfection (8, 15, 25, 30, 34).

Although there have been several clinical investigations

which evaluated and compared the serologic response of

patients withseriousS.aureusinfectionstoboth TA and PG

(2, 26, 30), in each ofthese studies, patients with

osteomy-elitis were included in the general categories of either complicated or uncomplicated S. aureus bacteremia. As a

result, it is difficult to ascertain the proportion of patients withosteomyelitisthatwerefoundtohave elevatedantibody

responsestoTA, PG, orboth compared tothe responsesof

other patients without osteomyelitis that were included

within the same group. In the present study, we used the

ELISAtocomparethe IgGresponsesto TAandPG of both

rabbits withexperimentalS. aureusosteomyelitis,aswellas

patients with bone infections caused by this organism. As

shown above (Table 2), rabbits with experimental osteomyelitis produced elevated levels of IgG to S. aureus

crude TA morefrequently than in response to PG.

Thecrude TAantigenwasalsoshowntobemoresensitive than PG forthedetection of IgG in the seraofpatientswith confirmedS.aureusosteomyelitis(Table 3). With respectto specificity,only 1/18(6%) of the patients withnon-S. aureus

osteomyelitis hadanelevated responsetothe crude TA and

none waspositive tothePGantigen. Recently, Wheatetal.

(34)reportedthat52% of the patients withanacuteS.aureus

infection and 47%withchronic S. aureus osteomyelitis had

an increased IgG response to an ultrasonic extract of S.

aureusWood 46,asmeasured byaradioimmunoassay. They

also notedthat6/33 (18%) of the patientswith osteomyelitis caused byorganisms other than S.aureusgavefalse-positive reactions inthisassay. Itappears, therefore, that the

sensi-tivity of the TA ELISA closely approximates that of the

radioimmunoassayand that the ELISAisalsomorespecific for S. aureus antibodies.

Ourinability todemonstrate anincreased IgGresponseto

PG in patients with either acute or chronic S. aureus

osteomyelitis is somewhat surprising in view of previous studies(2, 26, 27, 30,35) that have shownthatIgG levelsto

PGaresignificantly elevated in patients with othertypes of serious S. aureus infections. Thereareseveral reasonsthat may explain theapparent differences in these results. One

possibility is that in these previous studies, elevated

anti-bodies to PG were detected primarily in patients with S.

aureus bacteremia and that the lack of PG antibodies in

patients with S.aureusosteomyelitis,in which bacteremiais

often absent, may reflect differences in the processing or

presentation (or both) of PG by macrophages in

osteomyelitis, asopposedtoS. aureusinfections associated with bacteremia. Another possibility that may explain the

different results is the methodofsolubilizingthe PG before itsuseintheassays. Whereas in the aforementionedstudies

PG was solubilized by either ultrasonication or digestion

with lysozyme, weelectedtoeliminate solubilization ofour

PGpreparation becauseof previous reports which indicated

that its immunological activitymaybe lost iffragmentationis

excessive

(18, 20).

The antigenicity ofour PG

preparation

was,

however,

verified in the ELISA with a

hyperimmune

rabbit anti-S. aureus serum inwhich an

approximately

five fold-increase in the

OD405

value

(0.700)

was

detected,

as

compared

with that

(=

0.145)

with serum obtained before immunization

(data

not

shown).

In

conclusion,

wehave useda

rapid

andsensitive ELISA to measure

IgG

tocrude TA

and

PGin rabbits and humans

with

S. aureus

osteomyelitis.

This crudeTA

antigen

prepa-ration wasfound tobe

significantly

more sensitive than PG for thedetection of

IgG

inboth rabbit and humansera. Itis of

particular

importance

thatelevated levelsofTA antibod-ies wereencountered with greater

frequency

inrabbits with

experimental

S. aureus

osteomyelitis

than in humans with clinical infections.

Of

the 18 total serum

samples

that were

obtained from these rabbitsat4and 10 weeks

postinfection,

13/18

(72%)

werefound tobe

positive

forTAantibodies

by

ELISA.In contrast,

IgG

toTAwas

significantly

elevated in

only

17/39

(44%)

of the

patients

with confirmed S. aureus

bone infections

(both

acute and

chronic).

Because others

have shown that

IgG

to TA may be present in

high

levels

early

during deep-seated

S.aureus infections andthen may

decline to normal values after the initiation of

appropriate

antibiotic

therapy

(1,

6,

27),

itisconceivable thatsomeofthe

patients

included in our

study

may have

responded

to

therapy.

Because none ofthe rabbits with

experimental

S.

aureus

osteomyelitis

was

placed

on a

therapeutic

regimen,

the infections continued to progress in

severity,

thereby

perhaps

accounting

for the

higher

frequency

of

elevated

TA antibodies.

Although

the resultsof thepresent

investigation suggest

thattheTAELISAmay beusefulfor theidentification ofS.

aureus in

patients

with

osteomyelitis,

additional studies are

required

involving

a

larger

patient

population

with sera

obtained beforeandafter

therapy,

tomore

adequately

eval-uate the

efficacy

of this

procedure

for the

diagnosis

and

management ofthese types ofinfections. ACKNOWLEDGMENTS

We thank Seigfried J.

Schaberg

and

Douglas

M. Arendt for outstanding

surgical

and

histopathological

support,

respectively.

Wealso thankT. L.Bredice for the

preparation

of this

manuscript.

Thiswork(researchtask3S162775A825AA083)was

supported

by

the Naval Medical Research and

Development

Command. LITERATURECITED

1. Bayer, A.S., D. B. Tillman, N.

Concepcion,

and L. B. Guze. 1980. Clinical value of teichoic acid

antibody

titers in the diagnosis and management of

staphylococcemias.

West. J. Med. 132:294-300.

2. Christensson,B.,F.Espersen,S. A.Hedstrom,andG.Kronvall. 1985.

Serological

assays

against Staphylococcus

aureus

pepti-doglycan, crude

staphylococcal antigen,

and

staphylolysin

in the diagnosisofserious S. aureusinfections. Scand. J. Infect. Dis. 17:47-53.

3. Committeeon theCare and Use ofLaboratory Animals. 1985. Guide for the care and use oflaboratory animals. Institute of

Laboratory Animal Resources, National Research Council, U.S. PublicHealthService,

Washington,

D. C.

4. Crowder,J.G.,and A.C. White.1972.Teichoic acidantibodies in staphylococcal and

nonstaphylococcal

endocarditis. Ann. Intern. Med. 77:87-90.

5.

Dich,

V. Q., J. D. Nelson, and K. C. Haltalin. 1975. Osteomyelitis in infants and children.

Amn.

J. Dis. Child. 129:1273-1278.

6. Eid, A. M.,H.Issa, andA.I. Deif. 1980. Some

immunological

aspects of

staphylococcal hematogenous

osteomyelitis.

Arch. Orthop. Trauma.

Surg.

96:221-224.

on April 11, 2020 by guest

http://jcm.asm.org/

(6)

7. Engvall, E., and P. Perlman. 1972. Enzyme-linked immunosor-bent assay (ELISA). III.Quantitation ofspecificantibodiesby enzyme-labeledanti-immunoglobulininantigencoated tubes. J. Immunol. 109:129-135.

8. Granstrom,M.,I. G. Julander, S. A.Hedstrom,and R.Mollby. 1983. Enzyme-linked immunosorbent assay for antibodies againstteichoic acid inpatientswith staphylococcal infections. J.Clin. Microbiol. 17:640-646.

9. Harris, N. H. 1960. Some problems in thediagnosis and treat-ment of acute osteomyelitis. J. Bone Jt. Surg. Br. Vol. 42B:535-541.

10. Herzog, C., H. C. Wood, I. Noel, and J. C. Booth. 1984. Comparison of a new enzyme-linked immunosorbent assay method with counterimmunoelectrophoresis for detection of teichoic acidantibodies inserafrom patients with Staphylococ-cusaureusinfections. J. Clin. Microbiol. 19:511-515.

11. Jackson, L. J., M.I. Sottile, F. G. Aguillar-Torres, T.H. Dee, and M. W. Rytel. 1978. Correlation ofantistaphylococcal anti-body titerswithseverity of staphylococcal disease.Am. J. Med. 64:629-633.

12. Jacob, E., D. M. Arendt,I.Brook, L. C. Durham, M. C. Falk, and S. J.Schaberg. 1985. Enzyme-linkedimmunosorbentassay for detection of antibodiestoStaphylococcus aureusceli walls inexperimental osteomyelitis. J.Clin. Microbiol. 22:547-552. 13. Lack, C. H., and A. G. Towers. 1962. Serological tests for

staphylococcal infection. Br. Med. J.2:1227-1231.

14. Mackowiak,P. A.,S. R. Jones, and J. W. Smith. 1978. Diagnos-tic value ofsinus-tract cultures in chronic osteomyelitis.J. Am. Med. Assoc.239:2772-2775.

15. Mackowiak, P. A., and J. A. Smith. 1978. Teichoic acid anti-bodies in chronic staphylococcal osteomyelitis. Ann. Intern. Med. 89:494-496.

16. Mollan, R. A. B., and J. Piggot. 1977. Acute osteomyelitis in children.J. Bone Jt. Surg. Br. Vol.59B:2-7.

17. Norden, C. W. 1970. Experimentalosteomyelitis: adescription of the model.J. Infect. Dis. 122:410-418.

18. Oken, M. M., P. K. Peterson, and B. J. Wilkinson. 1981. Endogenous pyrogen production by human blood monocytes stimulated by staphylococcal cell wall components. Infect. Immun. 31:208-213.

19. Pancoast, S. J., and H. C. Neu. 1980. Inosteomyelitis the initial treatment is crucial. Med.Times 108:56-75.

20. Peterson, P. K., B. J. Wilkinson, Y. Kim, D. Schmeling, S. D. Douglas, P. G. Quie, and J. Verhoef. 1978. The key role of peptidoglycan in the opsonizationofStaphylococcusaureus.J. Clin. Invest. 61:597-609.

21. Porter, W. H. 1975.Application of nitrousaciddeterminationof hexosaminestothesimultaneous GLCdetermination ofneutral andaminosugars inglycoproteins. Anal. Biochem. 63:27-43. 22. Sheagren, J. N. 1984.Guidelines forthe use of the teichoic acid

antibody assay.Arch. Intern. Med. 144:250-252.

23. Taylor, A. G., J. Cook, W.J. Fincham, andF. J. C. Millard. 1975.Serologictestsin thedifferentiationofstaphylococcaland tuberculous bone disease. J. Clin. Pathol. 28:284-288. 24. Tuazon,C.U., J.N.Sheagren,M. S.Choa,D.Marcus,andJ.A.

Curtin. 1978. Staphylococcus aureus bacteremia: relationship between formation of antibodies toteichoic acid and develop-mentof metastatic abscesses. J. Infect. Dis. 137:57-62. 25. Tuazon, C.V., andJ.N.Sheagren. 1976. Teichoic acid

antibod-ies in the diagnosisofserious infections with Staphylococcus aureus. Ann. Intern. Med.84:543-546.

26. Verbrugh, H. A., R. Peters, W. H. F. Goessens, and M. F. Michel. 1986. Distinguishing complicated from uncomplicated bacteremia caused by Staphylococcus aureus: the value of "new' and ''old' serological tests. J. Infect. Dis. 153:109-115.

27. Verbrugh, H. A., R. Peters, M. Rozenberg-Arska, P. K. Peterson,andJ.Verhoef. 1981. Antibodiestocellwall peptido-glycanofStaphylococcus aureus in patients with serious staph-ylococcal infections.J. Infect. Dis. 144:1-9.

28. Waldvogel, F. A., G. Medoff, and M. N. Swartz. 1970. Osteomyelitis: areview ofclinicalfeatures, therapeutic consid-erations and unusual aspects. N.Engl.J. Med.282:198-206. 29. Waldvogel, F. A., and H. Vasey. 1980. Osteomyelitis: the past

decade. N. Engl. J. Med.303:360-367.

30. Wergeland, H., C. Endresen, O. B. Natas, P. Aasjord, and P. Oeding. 1984. AntibodiestoStaphylococcusaureus peptidogly-canand lipoteichoic acid in sera from blood donors andpatients with staphylococcal infections. ActaPathol. Microbiol. Immu-nol.Scand. Sect. B 92:265-269.

31. West, T. E., N. M. Burdash, A. M. Boehm, and M. E. West. 1983. Evaluation of a commercial counterimmunoelectropho-resis kit for detection ofStaphvlococcusaureus teichoic acid antibodies J. Clin. Microbiol. 17:567-570.

32. Wheat, L. J. 1985.Diagnostic strategies in osteomyelitis. Am. J. Med. 78:218-223.

33. Wheat, L. J., R. B.Kohler, and A. C. White. 1978. Solid-phase radioimmunoassay for immunoglobulin G Staphylococcus au-reusantibody in serious staphylococcal infection. Ann. Intern. Med. 89:467-472.

34. Wheat, L. J., A. C. White, and C. Norden. 1985. Serological diagnosisofStaphylococcusaureus osteomyelitis. J. Clin. Mi-crobiol. 21:764-767.

35. Wheat, L. J. , B. J. Wilkinson, R. B. Kohler, and A. C. White. 1983.Antibody response to peptidoglycan during staphylococ-cal infections. J. Infect. Dis. 147:16-22.

36. Yamada, J. K., C. B. Inderlied, and R. K. Porschen. 1983. Detection ofantibody to Staphylococcusaureus teichoic acid by enzyme-linked immunosorbent assay. J. Clin. Microbiol. 17:898-905.