Assessment of group B streptococcal opsoninsin human and rabbit serum by neutrophilchemiluminescence.

V G Hemming, … , A O Shigeoka, H R Hill

J Clin Invest. 1976;58(6):1379-1387. https://doi.org/10.1172/JCI108593.

The factors important in host defense against group B streptococci are not well understood.The role of antibody and complement in the prevention of serious infection by theseorganisms is not known because, to date, a reliable measure of functional opsonic activityhas not been developed. Recently, it has been shown that neutrophils produce achemiluminescence after ingestion of particulate matter, and that this event can be detectedand quantitated in a liquid scintillation system. We have adapted the chemiluminescenceprocedure to examine rabbit hyperimmune and human serum for the presence of group Bstreptococcal opsonins. Group B streptococci of types Ia, II, and III that were opsonized inhomologous but not heterologous type serum produced a peak in chemiluminescence whenadded to normal human neutrophils. Such activity was correlated, in each instance, withingestion of bacteria by neutrophils and deposition of immunoglobulin and C3 on thebacterial surface as detected by indirect immunofluorescence. With this assay, we haveexamined sera from colonized and diseased patients for the presence of opsonins to typesIa, II, and III group B streptococci. Maternal sera often contained type-specific opsoninswhich resided in the IgG fraction and which crossed the placenta to appear in paired cordspecimens. 63% of patients colonized with group B streptococci had serum opsonins totheir colonizing type of organism. […]

Research Article

Find the latest version:

http://jci.me/108593/pdf

Pdf

Assessment of Group B Streptococcal Opsonins in Human

and Rabbit Serum by Neutrophil Chemiluminescence

VAL G. HEMMING,ROBERTT. HALL, PHILIP G. RHODES,ANN0. SHIGEOKA,and HARRYR. HILL

From the Division of Clinical Immunology, the Department of Pediatrics and the Department ofPathology, University of Utah College of Medicine, Salt Lake City, Utah 84132 and the Children'sMercy Hospital, University of Missouri, Kansas City, Missouri 64108

A B S T RA C T The factors important in host defenseagainst group B streptococci are not well understood.The role of antibody and complement in the preventionof serious infection by these organisms is not knownbecause, to date, a reliable measure of functionalopsonic activity has not been developed. Recently, ithas been shown that neutrophils produce a chemilumi-nescence after ingestion of particulate matter, and thatthis event can be detected and quantitated in a liquidscintillation system. Wehave adapted the chemilumi-nescence procedure to examine rabbit hyperimmuneand human serum for the presence of group B strepto-coccal opsonins. Group B streptococci of types Ia, II,and III that were opsonized in homologous but notheterologous type serum produced a peak in chem-iluminescence when added to normal human neutro-phils. Such activity was correlated, in each instance,with ingestion of bacteria by neutrophils and deposi-tion of immunoglobulin and C3 on the bacterial surfaceas detected by indirect immunofluorescence.

With this assay, we have examined sera fromcolonized and diseased patients for the presence ofopsonins to types Ia, II, and III group B streptococci.Maternal sera often contained type-specific opsoninswhich resided in the IgG fraction and which crossedthe placenta to appear in paired cord specimens.63% of patients colonized with group B streptococcihad serum opsonins to their colonizing type oforganism. In contrast, none of the 15 patients withsepsis or meningitis had opsonins directed againsttheir infecting strain. These data suggest that the

This work was presented, in part, at the Annual Meeting ofthe Society for Pediatric Research, St. Louis, Mo., 29 April1976.

Dr. Hill is an Investigator of the Howard Hughes MedicalInstitute.

Received for publication 24 March 1976 and in revisedform 30 August 1976.

lack of type-specific opsonins to group B streptococcimay be one of the important factors in determininghost susceptibility to systemic infection with strains ofthis group.

INTRODUCTION

In recent years streptococci of Lancefield's group Bhave become a leading cause of serious infections ininfancy. Although a number of clinical and epidemio-logical reports on infection due to these organisms haveappeared in the literature (1-8), there is little defini-tive information on the factors important in the hostdefense mechanism against the strains of this group.Lancefield and co-workers have shown that rabbithyperimmune group B streptococcal antiserum willpassively protect mice injected with virulent strainsof group B streptococci (9-10). Since such antiserumdoes not appear to possess direct bactericidalactivity, mouse protection most likely results fromopsonic antibodies present in the serum. The role ofspecific and nonspecific opsonins in the prevention ofinfection due to group B streptococci has not beenadequately investigated, however, because of the lackof a reliable method for detecting and quantitatingfunctional opsonins for these organisms.

Allen and associates (11-15) have reported thathuman neutrophils emit a chemiluminescence afterphagocytosis of opsonized bacteria. The chemi-luminescence, which can be detected and quanti-tated in a liquid scintillation counter, appears toresult from decay to the ground state of electronicallyexcited carbonyl groups thought to be generated duringsinglet oxygen mediated oxidation of phagocytizedsubstrate. In the studies reported here, we haveadapted the chemiluminescence procedure to examinerabbit hyperimmune and human serum for thepresence of group B streptococcal opsonins. In

The Journal of Clinical Investigation Volume 58 December 1976 -1379-1387 1379

addition, we used this technique to assess opsonicactivity in serum from patients with group B strepto-coccal disease as well as colonized aind uncolonizedindividuals in an attempt to define what role fune-tional opsonic antibodies have in preventing clinicalinfection due to streptococci of Lancefield's group B.

METHODSPreparation of organisms. Reference strains of group B

streptococci type Ia (090), II (18RS21), and III (D136C)(kindly supplied by Dr. Rebecca C. Lancefield of the Rocke-feller University, New York) or group B streptococcalstrains isolated from infected or colonized patients (wildstrains) were cultured at 37°C in 50 ml of Todd-Hewitt(Difco Laboratories, Detroit, Mich.) broth for 20 h. Inselected experiments, the strains were grown in the buffered,glucose-enriched broth described by Baker et al. (16), andharvesting was carried out after 4 h of incubation during logphase growth. The organisms in the broth cultures weresedimented by centrifugation, the supernatant broth wasremoved and discarded, and the concentrated organisms werewashed three times in sterile phosphate-buffered saline(PBS)l (4,500 ml distilled water, 5.2 g Na2HPO4, 0.9 g KCl,0.9 g KH2PO4, 36 g NaCI). Standard suspensions of organ-isms of each reference type were prepared by diluting theconcentrated, washed organisms in sterile PBS to an opticaldensity (OD) of 0.9 at a wave length of 620 nm (Spectronic20, Bausch & Lomb Inc., Rochester, N. Y.). These standardsuspensions contained from 5.0 x 108 to 1.0 x 109 colony-forming U/ml.

Opsonization procedure. Strains of group B streptococci(types Ia, II, and III) were opsonized by mixing 0.5 ml ofthe standard solution (OD 0.9 at 620 nm) with 0.1 ml of thetest serum (1:6 dilution) in a 12 x 75 mmsterile capped tube(Falcon Plastics, Div. of BioQuest, Oxnard, Calif.). In selectedexperiments with rabbit hyperimmune serum and in most ofthe studies with human serum, 0.025 ml of whole humancomplement (Cordis Laboratories, Miami, Fla.) was added tothe reaction mixture. This volume of the preparation wasfound to maximally enhance the effect of heated rabbit orhuman serum. Addition of a larger volume had no moreeffect than 0.025 ml. The phagocytic mixture was thenrotated for 30 min at 37°C. After rotation, each tubewas centrifuged. The cell button was washed twice in PBS,then resuspended to its original volume (0.5 ml) in PBS.

Preparation of polymorphonuclear leukocytes. Humanpolymorphonuclear leukocyte (PMN) suspensions wereprepared from fresh samples of heparinized (10 U/ml)whole blood obtained from healthy adult volunteers. Aftersedimentation of the erythrocytes, the leukocyte-rich plasmawas removed, centrifuged, and the leukocyte button waswashed twice in PBS. The washed leukocytes were sus-pended in PBS to a concentration of 1.0 x 107 PMNs/ml.

Preparation of precipitin antigen. Hot HCI and TCAextracts of the reference strains of group B streptococciwere prepared by slight modifications of the procedures ofLancefield and Freimer (9, 17, 18).

Scintillation counting. Beckman Poly Q (Beckman In-struments, Inc., Fullerton, Calif.) vials were wrapped inaluminum foil and stored in darkness for at least 18 h beforeuse. Scintillation counting was performed in ambient light and

' Abbreviations used in this paper: CL, chemiluminescence;PBS, phosphate-buffered saline; PMN, polymorphonuclearleukocyte.

temperature in a Beckman LS 100c liquid scintillationsystem (Beckman Instruments, Inc.) out of phase, with onephoto multiplier tube disconnected. 0.5-mi aliquots of thePMNs and 0.5 ml of the bacterial suspensions were mi.lxedand the volume adjusted to 3.5 ml with sterile PBS. The vialwas capped, mixed well, placed in the scintillationcounter, and counted for 1 min at approximately 10-minintervals for a total of 100 min. The chemiluminesceicedetected by the counter is expressed in counts per minute.

Visual microscopic examination of reaction mixtures. Afterscintillation counting, each reaction mixture was centrifugedat 5,000 rpm for 5 min, the supernate was removed, and thecell button was resuspended in a small volume of PBS.An aliquot of this mixture was placed on a clean microscopeslide and air dried. The slides were fixed for 10 mIi in100%Y methanol and stained with Giemsa (Harleco, Gibbstown,N. J., 1:50 in distilled water) for 1 h. Each slide was examinedmicroscopically under oil immersion to visually accessphagocytosis of the organisms. We found that attempting toquantitate the degree of phagocytosis visuallv by countingthe percentage of PMNs ingesting or the number of bacteriaingested per PMNwas extremely difficult. On the other hand,there was a clearly visable distinction between the appearanceof preparations in which there had been phagocytosisand those in which no ingestion had occurred. In positivesmears, the majority of the organisms appeared to be intra-cellular and almost all PMNshad numerous ingested bacteria.In contrast negative preparations contained bacteria whichwere predominantly unassociated with PMNs and only anloccasional leukocyte contained organisms. All slides wvereinterpreted without the knowledge of what made up thephagocytic mixture. Moreover, there was excellent correla-tion in interpreting phagocytosis vs. Ino phagocvtosis bythree individual observers.

Preparation of rabbit antiserum. Rabbit hyperiiimimluneserum was prepared by the method of Lancefield withformalinized whole cell vaccines (17). Additional referenceantiserum (Ia, II, and III) was kindly supplied by Dr. Lance-field and by Dr. R. R. Facklam and Dr. H. Wilkinson of theCenter for Disease Control, Atlanta, Ga. Group B strepto-coccal grouping antiserum was obtained from BurroughsWellcome Co., (Research Triangle Park, N. C.). This latterantiserum gives excellent capillary precipitin reactions withgroup B streptococcal extracts and forms heavy precipitinlines with group B cultu-re filtrates in counterimiimunio-electrophoresis (19).

Immunofluorescence procedure. Fluorescein isothio-cyanate conjugated antirabbit globulin (Behring Diagnostics,American Hoechst Corp., Somerville, N. J.) or antihuman C3antiserum (Behring Diagnostics) (0.1 ml of a 1:3 dilution)was added to the cell buttons of opsonized or unopsonizedbacteria. Each tube was mixed, incubated on ice for 30 mmi,and then washed three times with PBS. After washing, 0.1 mlof a 1:2 mixture of normal saline and glycerine was addedto the bacterial button and the tube was mixed. One drop ofthe mixture was placed on a clean microscope slide, acover glass was placed over the drop, and the mixturewas examined immediately with a Zeiss ultraviolet micro-scope using an epi-illuminator (Carl Zeiss, Inc., New York).The amount of fluorescence observed on the surface of thebacteria was graded from trace (small) amount of fluorescenceto 4+ (heavy fluorescence).

Agglutination procedure. Rabbit antisera were tested forthe presence of type-specific agglutinins by mixing a drop ofthe bacterial cell suspension with a drop of the test serumin a sterile glass test tube. These preparations were examinedfor the presence of agglutination of the organisms on aninverted microscope.

1380 V. G. Hemming, R. T. Hall, P. G. Rhodes, A. 0. Shigeoka, and H. R. Hill

Capillary precipitin procedure. Capillary precipitin testswere done by mixing equal volumes of the antiserum andantigen to be tested in sterile capillary tubes (1.2 mminsidediameter) (17). Typing reactions were read in 5-10 min;the formation of a precipitate was interpreted as a positiveprecipitin test.

Clinical specimens. Serum specimens were obtained from50 adults and 47 infants. Included in these sera weresamples from 18 colonized mothers, 9 colonized infants, 19septic infants and their mothers, 2 infected adults, and 11normal maternal cord pairs. The colonization status wasdetermined from throat, ear, umbilical, blood, or cerebrospinalfluid cultures in the colonized or diseased infants. Themothers of these patients had vaginal, cervical, and throatcultures taken. All swabs were plated onto 5% sheep bloodagar plates without the addition of selective antibiotics.After incubation for 18-24 h, typical streptococcal colonieswere identified; all strains were grouped and typed byDr. Lancefield's capillary precipitin technique (17) and bycounterimmunoelectrophoresis (19). The colonization statusof the maternal-cord infants pairs was not determined. Serawere frozen at -20' to -700C and thawed only once beforetesting. All sera were tested in the presence of a maximallyenhancing amount of whole human complement.

RESULTS

Chemiluminescence (CL). The addition of stockstrains of group B streptococci, pre-opsonized inhomologous type rabbit antiserum, to washed humanleukocytes resulted in an increase in CL whichwas easily measured in the scintillation counter(Fig. 1). No increase in CL was noted when un-opsonized bacteria were utilized. A small CL peakwas noted for type II, whereas higher peaks wereseen for types Ia and III. The rabbit antiseraused in this experiment were free of complement(no change in the CL elicited after heating at 56°Cfor 1 h). Addition of whole human complement to thepre-opsonization mixtures accentuated the CL peak fortype Ia by 3,500 cpm and for type II by 6,500 cpm. Thepeak for type III was not increased by the addition ofcomplement.

A recently isolated type III (wild) strain from aninfant with group B streptococcal sepsis was nextexamined to determine the effect of complementon its opsonization. The peak in CL observed whenthis organism was opsonized in heated type IIIrabbit antiserum (14.8 x 103 cpm) was not significantlydifferent from that obtained with the same serum andadded complement (14.7 x 103 cpm). Moreover, theCL curves observed with the two phagocytic mixtureswere essentially identical. No CL peak was notedwhen either stock or wild strains of group Bstreptococci were opsonized in the complementpreparations alone.

Additional studies were carried out to determinethe effects on CL of (a) using buffered, enrichedbroth which has been reported to enhance group Bcapsular polysaccharide production (16), (b) usinglog phase growth organisms rather than 20-h cultures,

20-

0

x

ECL7 1 5 -

wzwcnwZ 10-

-J7-I

2E0I- 5

wz

TYPE Ia + Io ANTISERUM

10 20 30 40 50 60 70 80 90TPME(MINUTES)

FIGURE 1 Neutrophil chemiluminescence elicited over a 90-min period by stock group B streptococcal strains of typesIa, II, and III which were opsonized in complement-freehomologous type rabbit hyperimmune serum. The controlsin this experiment are PMNsplus unopsonized bacteria, orthe PMNmixture without bacteria.

and (c) varying the number of organisms presentedto the cells.

The use of organisms grown in the buffered, enrichedbroth resulted in a higher rather than a lower peakin CL. The peak in CL produced by type III strainsgrown in this medium and opsonized in homologoustype rabbit or human serum was increased by approxi-mately 4,000 cpm (35-50%) over that produced bystrains grown in regular Todd-Hewitt broth and op-sonized in the same sera. Addition to PMNsof unop-sonized bacteria grown in the enriched broth or ofsuch organisms treated with serum lacking typespecific antibody did not result in a peak in CL.Moreover, the use of organisms in log phase growth(4 h incubation), which should presumably have moreantiphagocytic factors, did not result in a decrease inthe CL peak. Wild type III streptococci in log phasegrowth opsonized in type III rabbit antiserum andcomplement produced a slightly higher peak in CL(17.5 x 103 cpm) than did 20-h cultures opsonized inthe same opsonic mixture (16.0 x 103 cpm).

The number of organisms presented to the PMNshada significant effect on the resulting CL. Very low,almost undetectable peaks were obtained with lowernumbers of bacteria (2 logs below the usual number of5 x 108 to 1 x 109). In contrast, a 3-fold higher numberof organisms increase the peak in CL by an averageof 133%; 5-fold higher by 315%; and 10-fold by 333%.The number 5 x 108 to 1 x 108 organism (OD 0.9 at620 nm) was selected (somewhat arbitrarily) as aconvenient one that produced a significant but notmaximum CL response. With these conditions the

Group B Strep Opsonins 1381

°=20

w0zw

o15pw -z

-J

-J

0-or

No phagocytosisn = 47

:1,.

I .

phogocytosisnz 57

FIGURE 2 Peak neutrophil chemiluminescence elicited bygroup B streptococci correlated with the presence or absenceof phagocytosis as determined by microscopic examination ofeach of 104 neutrophil-bacteria reaction mixtures.

standard deviation of triplicate sampling averaged5% of the mean (five separate experiments) while thesame opsonizing serum tested on three differentoccasions had mean peaks in CL with a standarddeviation that averaged 7+6% of the mean (meanvariance 6+6% of the mean).

The data depicted in Fig. 2 are the results of 104separate observations of the bacteria-neutrophil prep-arations plotted by the peak in CL and the presenceor absence of bacteria inside neutrophils as determinedby microscope examination. In every case, where theCL peak exceeded 6,500 cpm, the smear showed thatthe majority of the bacteria were intracellular and thatmost PMNs had phagocytized. Conversely, phago-cytosis was not observed in any of the preparationswhich had a CL peak of less than 5,500 cpm. Therewere 12 instances, 11%, in which one could notpredict whether phagocytosis had taken place by re-ferring to the CL peak alone. These mixtures had CLpeaks between 5,500 and 6,500 cpm and showed onlyminimal phagocytic uptake. This led us to select6,500 cpm as the level of CL which indicates thepresence of functional opsonic antibodies.

Cross-opsonization and absorption studies. Todetermine if the group B streptococcal opsonic anti-bodies measured by the CL procedure were typespecific the following experiments were performed:(a) group B streptococci were pre-opsonized withheterologous type rabbit antiserum and tested for theirability to elicit a peak in CL; (b) an aliquot of

each type of rabbit antiserum was absorbed with anequal volume of live homologous or heterologoustype organisms; these absorbed sera were then used topre-opsonize type Ia, II, and III organisms; (c) analiquot of each type of rabbit antiserum was absorbedwith an equal volume of homologous type HC1 orTCA extracted antigen. The resulting precipitate wasremoved by centrifugation and this absorbed serumwas used to pre-opsonize both homologous and heter-ologous type organisms.

The results of several of these experiments aresummarized in Table I. CL peaks were observed foreach of the three streptococcal types opsonized withhomologous type rabbit antiserum. Absorption of thethree types of rabbit serum with homologous typeHC1 or TCA extracts had varying effects on theopsonic activity as measured by CL. Both HC1 andTCA extracts reduced the CL activity of the type laserum. In contrast, absorption of type II serum witheither of these extracts markedly enhanced its CLactivity. The opsonic activity of the type III serumwas unchanged when absorbed with HCl-extractedantigen, but was somewhat reduced when absorbedwith TCA extract. In no instance did homologousabsorption with either the HC1 or TCA extractcompletely remove the opsonic activity of any of the

TABLE IEffect of Homologous and Heterologous Type Absorption

of Group B Streptococcal Antiserum on Opsonizationof Homologous Type Group B Streptococci as

Measured by Chemiluminescence

Antiserum Absorption* Peak in CL

cpm x 103

Ia None 16.4Ia HCl Ia 7.7Ia TCA Ia 8.0Ia Ia cells 4.3Ia II cells 15.9Ia III cells 16.0

II None 6.0II HCl II 23.9II TCA II 16.3II II Cells 3.9II Ia cells 5.9II III cells 6.2

III None 13.4III HC1 III 13.1III TCA III 9.8III III cells 3.8III Ta cells 13.0III II cells 12.8

* Concentrated, live whole group B streptococcal cells, hotHC1, or TCA extracts of group B streptococcal cells.

1382 V. G. Hemming, R. T. Hall, P. G. Rhodes, A. 0. Shigeoka, and H. R. Hill

sera tested. This absorption did, however, remove alldetectable precipitins. Absorption of each serum withan equal volume of whole, homologous type organismsremoved all opsonic activity from that aliquot ofserum. However, absorption of serum with heter-ologous concentrated organisms did not decrease theserum's opsonic activity against the homologous strain.Type-specific agglutinins were present in all threetypes of rabbit antiserum. Absorption with homologoustype HCI and(or) TCA antigens removed these agglu-tinins from types Ia and III antisera, but not fromtype II.

Experiments were conducted to see if antibodies tothe group B streptococcal group-specific substancewere opsonically active in this system. Pre-opsonizationof each of the reference strains with group B strepto-coccal grouping serum with and without the addition ofhuman complement failed to elicit a CL peak or

visual evidence of phagocytosis in these experiments.CL assay of human serum. The CL opsonic assay

for group B streptococci has also been used to measure

opsonins in human serum. A representative experimentis shown in Fig. 3. Fresh human serum from a singledonor (V. G. H.) was found to contain opsonins fortypes II and III streptococci, but no opsonic activityfor type Ia. The opsonins against types II and IIIwere removed by homologous but not heterologoustype absorption with whole group B streptococcal cells(Table II). Furthermore, the serum could be titered bytwofold dilutions. (CL peak greater than 6,500 cpm:type 11-1:64, type III-1:32).

Indirect immunofluorescence studies. Group Bstreptococci of types Ia, II, and III were pre-opsonized in homologous and heterologous type rabbitantiserum. These preparations were then incubatedwith antirabbit globulin (goat) conjugated withfluorescein isothiocyanate. Organisms pre-opsonizedin homologous type antiserum gave strong fluorescentreactions while those pre-opsonized in heterologous

Uz

_(-)

C,Jz

-J o2 -xLAJ

I E

-J

a.

0

I

z

15-

10-

5-

TYPE II + HUMANSERUM

TY/YPE m+ HUMANSERUM

tUNOS UNNOEDTYPESONIZEDTYPERm

TYPE Io + HUMANSERUM

10 20 30 40 50 60 70 80 90TIME (MINUTES)

FIGURE 3 Neutrophil chemiluminescence elicited by groupB streptococci of types Ia, II, and III opsonized in fresh adulthuman serum (V. G. H.). Unopsonized type III organismswere used as controls.

TABLE IIEffect of Homologous and Heterologous Type Absorption

on Group B Streptococcal Opsonins in Human Serum

Peak in CLproduced by grotup B

streptococcal typesAbsorbed

Opsonizing serum with* Ia II III

cpm x 103

Humanadult (V. G. H.) Nothing 4.8 11.6 10.8Ia cells 4.3 10.7 11.0II cells 4.4 3.9 10.6III cells 4.4 10.1 4.3

* Absorbed with an equal volume of packed live organismsat 37'C for 30 min.

type serum showed little or no fluorescence. Similarresults were obtained when organisms preopsonized inhuman (V. G. H.) serum were examined. Referenceorganisms (Ia, II, and III) were pre-opsonized, thenincubated with rabbit antihuman IgG, IgA, IgM, andanti-C3. No indirect immunofluorescence was ob-served in the anti-IgM or anti-IgA preparations.However, organisms of types II and III showed4+ immunofluorescence when incubated with anti-IgG. Pre-opsonized Ia organisms, which did not elicita peak in CL, showed only trace amounts offluorescence with the fluorescein-conjugated anti-IgG.Each of the three streptococcal types opsonized inthe fresh human serum had C3 on their surface asdetermined by indirect immunofluorescence. Theamount of C3 on the surface of the Ta organisms,which showed no fluorescence with anti-IgG andwhich failed to produce a CL peak, was not markedlydifferent from that on the surface of the type II andIII strains (3+ to 4+).

Prevalence of group B streptococcal opsonins inhuman sera. 11 maternal-cord paired sera were col-lected and examined by the CL assay for opsoninsagainst the Ia, II, and III reference strains of groupB streptococci. The colonization status of theseindividuals, unfortunately, was not known. As shown inFig. 4, none of the maternal-infant sera had signif-icant opsonic activity (.6,500 cpm) against Ia or-ganisms and only 18% of the infants and 27% of themothers had serum opsonins to type II. In contrast,73% of the cord sera and 82% of the maternalspecimens possessed opsonic activity to group B strep-tococci of type III. When present, opsonic activitywas usually slightly higher in the maternal serum thanin the corresponding cord specimen. Indirect immuno-fluorescent studies showed significant (3+ or greater)deposition of IgG and C3 on the surface of bacteriaincubated with maternal and cord specimens posses-sing opsonic activitv. No IgA or IgM could be detected.

Group B Strep Opsonins 1383

wzzwcnwz

mr)J o

a -

XI: E

Ic

0

z

16-

14-

12-

10-

8-

6-

4-

2-

Infont Mother

--Per cent with 0%F%CL >6,500 1%_ 0%

Reference type Io

- 16-

- 14-

*12-

* 10-

- 8-

6

- 4-

Infant Mother

v-FI

Z18%=1 1

FIGURE 4 Opsonins to group B streptococci in pairedmaternal and cord sera as measured by the CL assay. (Phago-cytic uptake associated with CL peak of .6,500 cpm).

In sera lacking opsonic activity only trace amounts ofIgG and C3 were found on the bacterial surface byindirect immunofluorescence.

A total of 50 adult sera and 47 infant or cord serahave been examined for opsonic activity to types Ta,IT, and III group B streptococcal reference strains.These specimens include ones from the diseased andcolonized infants and their mothers, whose coloniza-tion status is known, the maternal cord pairs justdiscussed, and several adult volunteers who were notcolonized. The sera were examined with the additionof whole human complement. 74% of adults and 58%of infants were found to have opsonic activity fortype III group B streptococci, while opsonins to type IIwere present in 40%o of the adult and 28% of theinfant sera. Only 12% of the adult and 13% of theinfant sera contained opsonic activity for type Ta strains.

Included in the 97 sera examined thus far are 67specimens from patients with group B streptococcaldisease or ones colonized with strains of this group.62% of the patients were colonized or had disease dueto type II strains, 25% with type III, and 13% withtype Ia strains. The clinical data from the 19 diseasedinfants and two diseased adults are shown in Table TTT.18 of the infants had blood cultures positive for group Bstreptococci within 48 h of birth. Six of the patientsalso had roentgenographic or autopsy evidence ofpneumonia. Only one of these early onset cases hadmeningitis. The infant who had late onset disease(10 days) had meningitis with positive cerebrospinalfluid cultures for type III group B streptococci. Thetwo adults include a 22-yr-old female with puerperalsepsis (type II) who survived and an 80-yr-old female

who developed fatal group B streptococcal sepsis andmeningitis (type III).

All sera were tested against the homologous typereference strain, and when available against the infect-ing or colonizing strain isolated from the patient.(Unfortunately, some of these organisms were lostbefore initiation of these antibody studies.)

56% of sera from the 18 colonized mothers (whoseinfants did not become ill) contained opsonins to thehomologous type reference strain (Table IV). Nineinfants from these 18 colonized mothers were alsocolonized. Unfortunately, the organisms isolated fromonly four of these maternal infant pairs were availablefor testing. Two of the pairs (50%o) contained high levelsof opsonins to the offending organism. Seven of thenine colonized infants (78%) had serum opsonins tothe homologous type reference strain. In contrast noneof the 13 septicemic infants or the 2 septicemicadults from whomorganisms were available had serumopsonic activity (CL 2 6,500) to their infecting organ-ism (Tables III and IV). 3 of the 19 diseased infants andone diseased adult did have antibody to the homolo-gous type reference strain. The differences in antibodyprevalence between the colonized and diseasedpatients against their own strain (P c 0.023) or thehomologous type reference strain (P c 0.005) wassignificant by the Fischer Exact test.

DISCUSSION

Type-specific opsonins for group B streptococcihave been demonstrated in human adult and newborninfant serum. Maternal opsonic antibody, which ap-pears to reside in the IgG fraction (as determined byindirect immunofluorescence), crosses the placentaand appears in the newborn infant in most instances.Opsonins to type III group B streptococci werepresent in one-half to three-fourths of sera testedwhile antibodies to type II and Ta were less common.Opsonins directed toward the colonizing strain or ahomologous type reference strain of group B strepto-cocci were often present in the serum of asymptomatic,colonized patients. In contrast, none of 15 diseasedpatients had serum opsonic activity for their infectingstrain and only 4 of 21 (19%) had opsonins to thehomologous type reference strain. These data suggestthat the lack of type-specific serum opsonins for groupB streptococci may be a factor in the development ofserious infection with these organisms.

The antigenic determinants responsible for viru-lence and the production of opsonic antibodies forthe different types of group B streptococci are notprecisely defined. The type I strains (Ia, Tb, and Ic)have several known detectable antigens which includethe following: (a) Ia and Ic have a shared, major

1384 V. G. Hemming, R. T. Hall, P. G. Rhodes, A. 0. Shigeoka, and H. R. Hill

II

TABLE IIIClinical and Bacteriologic Data and Serum Opsonic Activity in Patients

with Systemic Group B Streptococcal Disease

Serum opsonic activity(CL peak) of the patient's

serum against

Age at Type Infecting Homologous typeSubject Weight diagnosis Diagnosis Outcomet recovered strain reference strain

g cpm x 103

1 1,200 NB* Pneumonia and sepsis D II 6.4 (60) 5.1 (60)2 3,000 NB Sepsis S II 5.1 (75) 5.1 (60)3 2,900 NB Sepsis D II 4.7 (60) 5.9 (75)4 2,800 NB Pneumonia, sepsis D II 6.2 (56) 7.5 (68)5 3,700 NB Sepsis D II 6.1 (60) 6.4 (75)6 2,540 NB Sepsis S III 5.2 (60) 6.2 (60)7 3,400 NB Sepsis and pneumonia D II 3.4 (56) 3.9 (56)8 2,650 NB Sepsis, meningitis pyarthrosis S Ia 4.9 (75) 9.3 (70)9 2,230 NB Sepsis S II 6.1 (56) 6.3 (56)

10 3,070 NB Sepsis, pneumonia D II 6.2 (60) 6.0 (45)11 2,060 NB Pneumonia, sepsis S II 5.4 (45) 7.0 (60)12 2,630 NB Sepsis S III 4.5 (60)13 1,600 NB Sepsis D II 4.1 (60)14 1,750 NB Pneumonia, sepsis S III 9.2 (75)15 2,600 NB Sepsis S II - 4.3 (72)16 3,100 NB Sepsis S II 3.9 (60) 5.9 (75)17 2,100 NB Sepsis S II 4.2 (60)18 3,300 10 days Meningitis S III 4.0 (60) 5.2 (60)19 2,800 NB Sepsis S Ia 4.8 (60)20 22 yr Puerperal sepsis S II 5.7 (60) 5.9 (75)21 80 yr Sepsis, meningitis D III 4.1 (75) 7.3 (75)

*NB, symptomatic in first 48 h of life.4 S, survived; D, died.

) = Time in minutes to peak in CL.

polysaccharide, capsular antigen; (b) lb and Ic have a

shared protein antigen; (c) Ib has a major, capsularpolysaccharide antigen; and (d) Ia, Ib, and Ic have a

shared minor polysaccharide antigen which has beentermed the Ia, b, c antigen (10). Because of thesedifferent antigenic determinants and the cross re-

activity between strains, we chose to examine only theIa strain. The major antigens of the type II and III

group B streptococci appear to be capsular poly-saccharides without significant cross reactivity. Lance-field and Freimer (18) have indicated that extractionof the group B capsular polysaccharide antigen withTCA yields a more complete antigen containinggalactose, glucose, glucosamine, and a labile sialicacid-containing component. Treatment of this TCAex-

tract or whole organisms with hot HCI yields a

TABLE IVGroup B Streptococcal Opsonins in Serum from Infected and Colonized Patients

Homologous reference strain Strain from patient

Patient group Tested Opsonins Tested Opsonins

n % P value* n % Pvalue*

Colonized mothers 18 56 <0.005 4 50 <0.023Colonized infants 9 78 4 50Infected infants 19 16 13 0Infected adults 2 50 2 0

* Fisher Exact test, comparison of infected patients to colonized patients.

Group B Strep Opsonins 1385

degraded "partial" capsular antigen. In their studies,absorption of rabbit hyperimmune antiserum witheither the HCI or the TCA extract removed some, butnot all, of the mouse protective activity. This sug-gested the possibility of another antigen-antibodysystem important in the virulence of these organisms.

In the present studies, absorption of rabbit hyper-immune antiserum with the homologous TCA antigenextract removed all detectable precipitins but only aportion of the opsonic antibody with types Ia and ITT.Absorption with "partial" HCI antigen had varyingresults on the three types. It significantly decreasedthe opsonic activity of the la antiserum, suggestingthat this determinant is a major one in strains of thistype. Paradoxically, absorption of type II rabbit anti-serum with homologous type HCl and TCAextracts sothat all detectable precipitins were removed con-sistently enhanced opsonization as measured by CL.The reason for this extremely reproducible phenome-non is unknown. One possible explanation is that theunabsorbed type II antiserum caused agglutinationof the organisms resulting in fewer individualorganisms being available to the PMNs for phago-cytosis. The rabbit antiserum to each of the typespossessed agglutinins, however, and absorption withthe HCI extracts did not remove this activity. It isconceivable that a prozone phenomena was operativehere, but to our knowledge this has not beendescribed in an opsonic antibody system. Anotherpossible explanation is that antibodies directed againstthe HCI and TCA determinant of the type TIstrain compete for the binding site (or a portion ofthe site) needed for the opsonic antibody.

The failure of complete absorption of opsonic anti-body after removal of all detectable precipitins witheither the HCI or the TCA extract may indicate thepresence of another antigenic determinant on wholecells of group B streptococci, as has been suggestedby Lancefield and Freimer (18). Characterization ofthis determinant, if it exists, and the antibodiesdirected against it may be of utmost importance inefforts aimed at developing a vaccine for theseorganisms.

Of interest is the fact that recently isolated "wild"strains of group B streptococci resist opsonization byantibody and complement more than stock strains of thesame type. The reason for this phenomena is unknown,but may be related to the quantity of antiphagocyticfactors being produced. We were unable to confirm,however, that manipulations designed to enhancecapsular polysaccharide production, such as the use ofbuffered enriched medium or log phase growthcultures, resulted in a decrease in phagocytic uptakewhen specific antibody was present.

The role of complement, both classic and alternativepathways, in the opsonization of group B strepto-

cocci also needs further study. Wewere consistentlyable to significantly enhance opsonization of types Iand II strains with whole human complement, but thispreparation had no effect on type III organisms. Thiswas true if stock or wild group B streptococcalstrains were used. In addition, type Ta strainsincubated in fresh humian serum demonstrated signifi-cant amounts of comiiplement on their surface but didnot elicit a peak in CL. This suggests that thealternative pathway mav not play a major role in op-sonizing these bacteria.

Dossett et al. detected opsonins to group Bstreptococci in whole blood and serum in 1969 (20).Subsequently Klesius et al. (21) and Mathews et al.(22) reported that serum from humans and experi-mentally infected baboons contained opsonins directedagainst types Ta, Tb, Ic, II, and III organisms. Theirstudies, which were carried out by visually deter-mining phagocytic uptake and the percentage ofPMNsingesting microorganisms, suggested that typesTb, Ic, II, and III streptococci were naturally andnonspecifically opsonized by 95% of human andbaboon sera. Absorption of such serum with homolo-gous type whole organisms did not decrease opsonicactivity. In contrast our studies and those of Lance-field and co-workers (10, 18) would suggest that opsonicor mouse protective antibody is type specific, sincehomologous but not heterologous type absorption ofserum with whole cells removes this activity. Itshould be noted, however, that antiserum containingexcellent precipitins against the group B groupspecific carbohydrate does not possess opsonic activityas measured in our system or mouse protective anti-body as measured in Dr. Lancefield's procedure.

Recently Baker and Kasper (23) have developed aradioactive antigen-binding assay for detectingantibody directed against a polysaccharide antigencontaining both group-specific carbohydrate determi-nants and the type III capsular polysaccharide ofgroup B streptococci. This test apparently measuresall antibodies (precipitin, agglutinin, opsonic, etc.)directed against this combined antigen. Our results andthose of Lancefield and co-workers (10, 18) would sug-gest that precipitins or agglutinins directed against thegroup-specific carbohydrate have no role in protection.In addition, antisera against other types of group Bstreptococci show cross reactivity with the combinedantigen used in the radioimmunoassay. Using this tech-nique, they have found that patients who develop typeIII group B streptococcal disease (predominantly lateonset as vs. early onset in our series) usually lackantibody to the combined type III and group B poly-saccharide antigen (23). Although these results arevery similar to ours, the radioimmunoassay does not,of course, assess functional opsonic activity.

The most striking finding in the data reported here

1386 V. G. Hemming, R. T. Hall, P. G. Rhodes, A. 0. Shigeoka, and H. R. Hill

is the absence of opsonins to the infecting organism in12 of our early onset bacteremic newborns, 1 caseof late onset meningitis, and 2 bacteremic adults. Astatistical comparison of these data reveal a significantdifference between the prevalence of opsonins incolonized patients vs. those who became infected.Clearly, the absence of type-specific opsonic activityappears to be one risk factor in the development ofsystemic infection.

ACKNOWLEDGMENTS

We thank Nancy Hogan, Mary Portas, and Gay Kurth fortechnical assistance and Marcy Pixton for secretarial aid.

This investigation was supported, in part, by U. S. PublicHealth Service grants AM18354 and AI-13150.

REFERENCES

1. Hood, M., A. Janney, and G. Dameron. 1961. Betahemolytic streptococcus group B associated withproblems in the neonatal period. Am. J. Obstet. Gynecol.82: 809-818.

2. Eickhoff, T. C., J. 0. Klein, A. K. Daly, D. Ingall,and M. Finland. 1964. Neonatal sepsis and other in-fections due to group B beta-hemolytic streptococci.N. Engl. J. Med. 271: 1221-1228.

3. Baker, C. J., F. F. Barrett, R. C. Gordon, and M. D. Yow.1973. Suppurative meningitis due to streptococci ofLancefield group B: A study of 33 infants. J. Pediatr.82: 724-729.

4. Franciosi, R. A., J. D. Knostman, and R. A. Zimmerman.1973. Group B streptococcal neonatal and infant infec-tions. J. Pediatr. 82: 707-718.

5. Baker, C. J., and F. F. Barrett. 1973. Transmission of groupB streptococci among parturient women and theirneonates.J. Pediatr. 83: 919-925.

6. Howard, J. B., and G. H. McCracken, Jr. 1974. Thespectrum of group B streptococcal infections in infancy.Am. J. Dis. Child. 128: 815-818.

7. Wilkinson, H. W., R. R. Facklam, and E. C. Wortham.1973. Distribution by serological type of group Bstreptococci isolated from a variety of clinical materialover a five-year period with special reference to neonatalsepsis and meningitis. Infect. Immun. 8: 228-235.

8. Hemming, V. G., D. W. McCloskey, and H. R. Hill. 1976.Pneumonia in the neonate associated with group Bstreptococcal septicemia. Am. J. Dis. Child. In press.

9. Lancefield, R. C. 1934. A serological differentiation ofspecific types of bovine hemolytic streptococci (group B).

J. Exp. Med. 59: 441-458.10. Lancefield, R. C., M. McCarty, and W. N. Everly. 1975.

Multiple mouse-protective antibodies directed againstgroup B streptococci.J. Exp. Med. 142: 165-179.

11. Allen, R. C., R. L. Stjernholm, and R. H. Steele. 1972.Evidence for the generation of an electronic excitationstate (s) in human polymorphonuclear leukocytes and itsparticipation in bactericidal activity. Biochem. Biophys.Res. Commun. 47: 679-684.

12. Stjernholm, R. L., R. C. Allen, R. H. Steele, W. W.Waring, and J. A. Harris. 1973. Impaired chemilumi-nescence during phagocytosis of opsonized bacteria.Infect. Immun. 7: 313-314.

13. Allen, R. C. 1975. Halide dependence of the myeloperoxi-dase mediated antimicrobial system of the polymor-phonuclear leukocyte in the phenomenon of electronicexcitation. Biochem. Biophys. Res. Commun. 63: 675-683.

14. Allen, R. C. 1975. The role of pH in the chemilumi-nescent response of the myeloperoxidase-halide-HOOHantimicrobial system. Biochem. Biophys. Res.Commun. 63: 684-691.

15. Allen, R. C., S. J. Yevich, R. W. Orth, and R. H.Steele. 1974. The superoxide anion and singlet molecularoxygen: their role in the microbiocidal activity of thepolymorphonuclear leukocyte. Biochem. Biophys. Res.Commun. 60: 909-917.

16. Baker, C. J., and D. L. Kasper. 1976. Microcapsule oftype III strains of group B streptococcus: production andmorphology. Infect. Immun. 13: 189-194.

17. Lancefield, R. C. 1938. A micro-precipitin-technique forclassifying hemolytic streptococci and improved methodsfor producing antisera. Proc. Soc. Exp. Biol. Med. 38:473-478.

18. Lancefield, R. C., and E. H. Freimer. 1966. Type-specific polysaccharide antigens of group B strepto-cocci. J. Hyg. 64: 191-203.

19. Hill, H. R., M. E. Riter, S. K. Menge, D. R. Johnson,and J. M. Matsen. 1975. Rapid identification of group Bstreptococci by counter-immunoelectrophoresis. J. Clin.Microbiol. 1: 188-191.

20. Dossett, J. H., R. C. Williams, Jr., and P. G. Quie. 1969.Studies on the interaction of bacteria, serum factors andpolymorphonuclear leukocytes in mothers and newborns.Pediatrics. 44: 49-57.

21. Klesius, P. H., R. A. Zimmerman, J. H. Mathews, andC. H. Krushak. 1973. Cellular and humoral immuneresponse to group B streptococci. J. Pediatr. 83:926-932.

22. Mathews, J. H., P. H. Klesius, and R. A. Zimmerman.1974. Opsonin system of the group B streptococcus.Infect. Immun. 10: 1315-1320.

23. Baker, C. J., and D. L. Kasper. 1976. Correlation ofmaternal antibody deficiency with susceptibility to neo-natal group B streptococcal infection. N. Engl. J. Med.294: 753-756.

Group B Strep Opsonins 1387