Vol. 23, No. 1JOURNAL OF CLINICAL MICROBIOLOGY, Jan. 1986, p. 56-610095-1137/86/010056-06$02.00/0Copyright ©) 1986, American Society for Microbiology

Relationships between Alternative Complement Pathway Activation,C-Reactive Protein, and Pneumococcal Infection

RUTH A. RABINOVITCH,'2t SUSAN M. KOETHE,3 JOHN H. KALBFLEISCH,"2 LAUREL C. PREHEIM,4 AND

MICHAEL W. RYTEL .2*Division of Infectious Diseases, Department of Medicine,'* and Department of Pathology,3 Medical College of

Wisconsin, Milwaukee, Wisc-onsin 53226; Veterans Administration Medical Center, Wood, Wisconsin 531932; InfectiousDiseases Section, Department of Medicine, Veterans Administration Medical Center and Creighton University School of

Medicine, Omaha, Nebraska 681O54

Received 26 June 1985/Accepted 30 September 1985

In the absence of specific antibody, opsonization of Streptococcus pneumoniae may be mediated by thealternative complement pathway (AP) or by C-reactive protein (CRP) via Cl binding. To determine the roleof these mechanisms in pneumococcal (PNC) disease, we studied 19 patients with differing severities of PNCinfection. C4 and CRP levels and zymosan-induced consumption of 50% hemolytic complement (CH50) weremeasured in specimens obtained acutely and then weekly. In patients with complicated illness, the modifiedmean CH50 in acute sera was 178 + 57 U/ml, significantly lower than the mean CH50 of 331 + 80 U/ml inpatients with uncomplicated illness (P < 0.05). The values of the two groups on a given day approximated eachother on days 7, 14, and 23. Consumption of complement by zymosan was also lower in acute sera of patientswith complicated illness, with a mean value of 19 ± 18 U/ml compared with 58 ± 30 U/mI in those withuncomplicated illness (P < 0.05). This difference was also seen on day 7 (P < 0.05). Disease involvinglower-numbered PNC serotypes (less than 10) correlated with reduced availability of AP factors in acute sera,independent of illness severity. Mean CRP levels were inversely related to zymosan-induced complementactivation in patients with complicated illness. These data suggest that in vivo depletion of AP factors issignificantly greater in patients with complicated illness and is associated with high CRP levels. CRP mayenhance AP activation via C3 convertase generation and function with it as a preantibody host defensemechanism.

Recovery from bacterial infections depends in large parton the ingestion and killing of invading microorganisms bythe phagocytes of the host. The encapsulated pneumococcusresists phagocytosis unless serum opsonins are available. Inthe absence of specific antibody, the alternative complementpathway (AP) can be activated via the cell wall teichoic acid(TA) (12, 31). C3b thus generated fixes to TA containing cellwall polymers (26, 30, 31), allowing granulocytes to adhereand phagocytosis to ensue. This form of opsonization pro-ceeds without capsular material or C4 but may be facilitatedby cell wall binding of nonimmunospecific immunoglobulinG (22, 28, 29) as well as by specific antibody (18). The lattergenerates C3 convertase via classical pathway activation,amplifying the AP C3b feedback cycle. In addition, there isin vitro evidence for variation among pneumococcal (PNC)serotypes in their ability to activate the AP (7, 9, 22).

Streptococcus pneumoniae may be opsonized via interac-tion with C-reactive protein (CRP). CRP has specific bindingaffinity for phosphocholine moieties (4, 14, 24); in thepresence of calcium ions, binding of CRP to PNC cell wallconstituents activates the classical complement pathway.These constituents are probably the choline-TA-containingpolymers. Mold and Gewurz showed in vitro that CRPinhibits PNC-induced AP activation in a dose-dependentfashion (16).

This study was done to determine the relationship be-tween severity of PNC infection, complement activation,and CRP levels. We measured the functional activity of the

* Corresponding author.t Present address: 691 Murphy Rd., Suite 115, Medford, OR

97504.

AP, using zymosan (Z)-induced consumption of 50% hemo-lytic complement (CHs5)), in serial serum specimens frompatients with PNC disease. Concomitant C4 and CRP levelswere also measured. We expected that in vivo consumptionof AP factors would reduce the factors available for in vitroZ activation; that is, the difference between CH50 content ofnormal saline (NS)- versus Z-treated sera would be smallerthan that found in a healthy subject with undiminished initialAP content. Our observations suggest that in vivo depletionof AP factors is significantly greater in severely ill patientsand is associated with high CRP levels. Neither positiveblood culture nor PNC serotype subgroups shows a consis-tent relationship to AP activation.

MATERIALS AND METHODS

Patients. Hospitalized patients with clinical diagnoses ofPNC pneumonia, meningitis, or bacteremic otitis media wereevaluated. A total of 19 patients was selected, based on oneof the following criteria: (i) growth of S. pneumoniae fromblood or cerebrospinal fluid cultures or (ii) a diagnosis ofpneumonia and characteristic sputum Gram stain or culturefor pneumococci, in conjunction with counterimmuno-electrophoresis detection of type-specific PNC capsularpolysaccharide in serum, urine, or sputum. Patients wereexcluded if neutropenia, hemoglobinopathy, complementdeficiency, or connective tissue disorder was present.Hospital records were reviewed to determine patient age,sex, presence of underlying diseases, duration and severity ofillness, microbial culture results, and PNC serotype. Illnesseswere categorized as complicated if meningitis, septic shock,or disseminated intravascular coagulation was present, threeor more pulmonary lobes were involved, or prolonged fever

56

on March 4, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

COMPLEMENT ACTIVITY IN PNEUMOCOCCAL INFECTION

(greater than 7 days), hospitalizatioh greater than 21 days, or

death due to PNC infection occurred. Ten normal healthysubjects were studied to establish ranges and means of normalcomplement activity for the assays performed. The serum ofone subject, whose hemolytic complement titer on repeatedassay consistently approximated these means, was chosen tobe an assayed control. Patient blood was obtained acutely(within 72 h of hospitalization) and then weekly for up to 3weeks. Serum was recovered, frozen, and stored at -70°Cuntil serial specimens from each patient could be simulta-neously assayed.

Serotype subgroups. On the basis of prior in vitro studies(7, 9, 22), we classified serotypes 14 and 19 as "goodactivators" of the AP (6 patients) and serotypes 3, 4, 7, 9,and 18 as "poor activators" (12 patients) of the AP.

Reagents. Zymosan A (no. Z-4250; Sigma Chemical Co.,St. Louis, Mo.) was boiled for 2 h in NS, washed three timesin NS, and suspended to a final concentration of 2 mg/ml. Astock buffer with final molarity of 0.1 M EGTA [ethyleneglycol-bis(,3-aminoethyl ether)-N,N,N',N'-tetraacetic acid]and 0.02 M MgCl2 was prepared by heating EGTA andMgCI2 in NS. NaOH was added dropwise until the EGTAdissolved; the solution was titrated back to pH 7.45 and thenbrought to a final volume with NS.Z consumption test. The functional activity of the AP was

determined by comparing CH50 values of parallel serum

specimens that were exposed to either Z or NS. A modifi-cation of the method of Fine et al. was used (8). Eachspecimen was run in duplicate, and the assayed control was

included. Two 0.2-ml samples of freshly thawed serum inEppendorf microcentrifuge tubes were hand mixed with 20,ul of EGTA-Mg2' buffer. This chelates and blocks activationof the classical complement pathway but not the AP, therebypermitting measurement of AP function alone. The mixturewas incubated in a 37°C water bath for 5 min. Z or NS at 0.2ml was added to each sample and then incubated for 20 min,which resulted in a maximal rate of Z activation of comple-ment. Specimens were quickly cooled on ice and Z wasremoved by centrifugation at 756 x g for 10 min at 4°C. Thesupernatant was recovered, and 20 ,u1 of 0.1 M CaC12 was

added to each sample. The CH50 titers were determined, andthe number of CH50 units consumed by Z (CH50-difference)was derived from the difference in CH50 content between theNS-containing sample (CH50-NS) and the Z-treated sample(CH50-Z).

CH50. Serum CH50 titers were measured with Veronal-buffered saline containing 0.1% gelatin, 0.005 M Ca2 and0.00015 M Mg2+ (GVB2+). Sheep erythrocytes (GIBCODiagnostics, Madison, Wis.) sensitized with anti-sheeperythrocyte hemolysin were used at a concentration of 108cells per ml. A 1:100 dilution of each test serum was

prepared in GVB2 . Serial 0.1-ml dilutions were incubatedwith buffer and 0.5 ml of sensitized sheep erythrocytes (1.5ml total volume) for 60 min at 37°C in a shaking water bath.The reaction was terminated by the addition of 4 ml of coldNS. After centrifugation the optical density of releasedhemoglobin was read at 414 nm, and the CH50 titers were

calculated. Results were expressed as CH50 units per milli-liter and were accepted if duplicate values were within 10%and the assayed control was in range.C4 and C-reactive protein. Values were quantitated by

using the technique of single radial immunodiffusion(Calbiochem-Behring Corp., La Jolla, Calif.).Data analysis. Study variates (CH50-NS, CH50-Z, the dif-

ference between these, i.e., CH50-difference, and CRP) weresummarized by the mean and standard deviation at each time

point. Data of control specimens were summarized in thesame fashion. Student's t test was used to examine theeffects of patient classification factors (complicated versus

uncomplicated, bacteremia versus abacteremia, serotypenumerically higher versus lower than 10, and good versus

poor activators of the AP by serotype); means of patient sera

were compared with the means of control sera by both thepaired and unpaired Student's t tests. These tests were

performed at each study time point. A probability level of0.05 or less indicated statistical significance.

RESULTS

Patients. Clinical parameters of patients studied, groupedaccording to severity of illness, are shown in Table 1. Therewere no differences with respect to age, sex, bacteremia, or

serotype number higher or lower than 10. Except for sple-nectomy, underlying diseases were equally distributed.AP activity. The functional integrity of the AP was deter-

mined by measuring Z-induced consumption of hemolyticcomplement in patient sera. Despite an acute-phase reactanteffect on the CH50 titers of patients, the functional activity ofthe AP, expressed as percent consumption, was significantlydepressed in the acute phase as well as at weeks 1 and 2. Byweek 3, mean patient values approached the mean of thecontrol (Table 2).The functional activity of the AP in patients with compli-

cated and uncomplicated disease was compared (Fig. 1).Acutely, depression of functional activity of the AP was seen

only in the complicated subgroup, whether expressed as themean of CH50-NS, CH50-Z, or CH50-difference (P < 0.05).The depression of CH50-difference in patients with compli-cated disease compared to those with uncomplicated diseaseremained significant until week 2. Thus, increased severityof illness was associated with reduced availability of APfactors for in vitro activation. Because of the few patients

TABLE 1. Clinical data of study patientsNo. of patients No. of patients

Parameter with complicated with uncomplicateddisease disease

Patients studied 11 8Age (mean [yr]) 55 52Age (range [yr]) 28-72 25-81Male 7 6Female 4 2Underlying diseaseAlcohol abuse 3 4Splenectomy 3 0Hematologic malignancy 2 1Chronic renal insufficiency 1 1Chronic sinusitis 1 0Chronic obstructivepulmonary disease 1 0

Cerebrospinal fluid leak 1 0None 0 2

PNC diseaseMeningitis 5 0Pneumonia 6 7Otitis media 0 1

Positive blood cultures 9 5Negative blood cultures 2 3Serotype <10 4 3Serotype >10 5 3Untypeable isolate 2 2Serotype, good activators 4 1Serotype, poor activators 5 5

VOL. 23, 1986 57

on March 4, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

58 RABINOVITCH ET AL.

TABLE 2. AP activation in patients with PNC disease versus control subjects

patient type" Day 1-3 (mean 2) Week 1 (n = 17) Week 2 (n = 12) Week 3 (n = 6)

CH5o-NSPt 247.5 ± 102.6 274.2 ± 106.7 271.3 ± 102.5 227.7 ± 68.5Co 219.6 ± 14.9 223.9 ± 29.2 222.1 ± 28.5 226.2 ± 9.7

CH14,-ZPt 210.9 ± 84.6 248.0 ± 95.0' 231.6 ± 73.9' 174.7 ± 74.8Co 168.2 ± 9.7 172.2 ± 24.2 173.2 ± 28.3 170.7 ± 12.0

CH5O-DifferencePt 36.6 ± 27.7 26.2 ± 25.0' 39.7 ± 45.9 48.0 ± 32.0Co 51.5 ± 19.7 51.8 ± 17.7 53.5 ± 18.5 55.5 ± 9.7

% ConsumptiondPt 13.2 ± 8.2' 9.1 ± 7.2' 13.2 ± 12.2' 22.1 ± 13.6Co 23.2 ± 7.2 22.9 ± 6.3 23.5 ± 6.5 24.5 ± 4.1

"Pt, Mean CH5(, of all patients at each time period ± standard deviation; Co, mean CH,(, for control subject at each time period ± standard deviation.bNumber of patients.`P < 0.05 versus control by the paired t test (paired with the concurrent determination on control sera) and the unpaired t test.d % Consumption = CH5(-difference/CH4(0-NS x 100%, calculated for each patient.

(three) with splenectomy, we were unable to demonstrate anindependent effect of this factor on AP function.CH50 titers in patients with disease involving serotypes

numerically lower or higher than 10 were compared (Fig. 2).The lower-than-10 patients showed significant depression ofthe mean CH50-difference acutely when compared to thehigher-than-10 subgroup (15.3 ± 12.7 versus 46.5 ± 18.4U/ml). This contrasted with our results when patients weregrouped as good versus poor activators of the AP (seeabove). There were no significant differences between thesesubgroups, although the good activators achieved greatermean CH50-difference titers (Fig. 3).

Studies of the functional activity of the AP for patientswith bacteremia versus abacteremia failed to yield significantdifferences between subgroups.CRP. Patients with complicated and uncomplicated dis-

ease had elevated mean CRP levels acutely which gradually

500

400

E 300-

(nz

2 200-

100-

400A

D 300c

~200I

E

100-

declined over the convalescent period (Fig. 4). The differ-ence between subgroups was not significant. Sera frompatients with complicated disease had an inverse relation-ship between mean CRP levels and availability of AP factorsfor in vitro Z activation. This was not seen in the sera fromthe subgroup with uncomplicated disease.Complement levels. C4 levels occasionally showed eleva-

tion or depression beyond the normal range in a givenpatient; however, mean C4 levels for all subgroups werenormal, and no statistically significant differences wereobserved.

DISCUSSIONThe relationships between severity of PNC infection, AP

activation, and CRP levels have not been well studied.Because in vitro models may not reflect in vivo activation,we undertook a prospective study in patients with PNC

200-

E

, 150-u

a10'2IU

50oj -,-,

4 8 12 16 20 24TIME (Doy)

4 8 12 16 20 24TIME (Days)

FIG. 1. Mean CH50 titers of sera from patients with complicated disease versus that from patients with uncomplicated disease over time;Symbols: ------, uncomplicated; , complicated; *, statistical significance of difference between groups at P < 0.05 level.

2 4 6 8 10 14 l8 2224TIME (Doys)

J. CLIN. MICROBIOL.

-04kb

on March 4, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

COMPLEMENT ACTIVITY IN PNEUMOCOCCAL INFECTION

100'

90'

80'

- 70'

8 60'a0 50'

1 40'0

o 30'

20-

10.

25-

20

Control

5SeroI}i,es >10

'4

Sotps</0

:5m

E15-

C

._

0

1=10-0

5.

4 8 12 16 20 24TIME (Doys)

FIG. 2. Mean CH50-difference titers in patients with illness in-volving serotypes lower than 10 versus those higher than 10. 0,statistical significance between subgroups at P - 0.05 level.

Complicated

l0

4 8 12 16TIME (Days)

20 24

FIG. 4. Mean CRP levels in sera from patients in the subgroupsof complicated and uncomplicated disease.

infection. We chose the Z consumption assay, rather thanimmunochemical or immunoelectrophoretic analyses, asmore accurately reflecting the in vivo activity. Although theother assays may be more sensitive in detecting smallchanges in component levels, our assay normalizes the largechanges in baseline complement component levels due to theacute-phase response. By running each serum specimen induplicate (with and without Z activation), each specimenserves as its own control. We found that the functionalactivity of the AP, as measured by Z consumption, wassignificantly depressed in acute sera of patients with PNCdisease when compared to normal sera. This is consistentwith the work of Coonrod and Rylko-Bauer which compareda population of patients with PNC pneumonia to a popula-tion of healthy controls (6). To determine whether the

.I& - ,

"Good" Activators*

_

-_I

,,~~~~%1% Jf~/./%

7 14

"Poor" Activators

21

TIME (Days)FIG. 3. Mean CH50-difference titers in patients with PNC

serotypes categorized as good versus poor activators of the AP.

population of patients with PNC disease was homogeneouswith respect to AP activation, we analyzed different patientsubsets. Our studies indicated that patients with more severeor complicated illness had acute sera with reduced availabil-ity of AP factors for in vitro Z activation, compared to thatfrom uncomplicated patients. This agrees with the data ofothers showing greater AP depression in patients with severebacteremic infections (6).Reduced availability of AP factors may reflect greater in

vivo activation and consumption of the AP in the patientgroup with complicated disease. The resulting complementdepletion may impair opsonic activity, thus compromisinghost defense. There is a dramatic defect in clearance ofpneumococci in complement-depleted animals (2, 10). Inaddition, AP activation can generate activated complementcomponents (11, 13, 21) which may produce tissue injury andlead to more severe illness. The extent of in vivo activationof complement may depend on the concentration of availableantigen. In an in vitro model, >105 pneumococci per ml arerequired to induce measurable amounts of complementconsumption (7). Such concentrations may be achieved inacutely infected lung or meninges and may contribute tomore efficient AP activation.

Severely ill patients could suffer from decreased synthesisof AP factors or from inhibition in the functional interactionof these factors with Z. Some patients with complementdeficiencies, such as that seen in sickle cell anemia, arepredisposed to more frequent and severe PNC infections.While it is possible that similar preexisting mild deficienciesof the AP predisposed the complicated patients to moreserious illness, this is unlikely because the functional activityof AP in convalescent sera increased to within the normalrange. It is unknown which, if any, of these mechanisms areoperative in human disease. No clear-cut hypothesis can beinvoked to explain the observed association.Lower-numbered serotypes have an association with more

virulent infections (1, 23). In analyzing patients with disease

70

_ 60ED 50

c400

10u200

10

VOL. 23, 1986 59

i

-% elw 11----*

on March 4, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

60 RABINOVITCH ET AL.

caused by different serotypes of pneumococci, we chosesubgroups with serotypes numerically higher or lower than10. We found that sera from patients with lower serotypeinfections had reduced availability of AP factors for in vitroZ activation when compared to the higher-than-10 group.This was independent of illness severity. PNC serotypesdiffer in their in vitro ability to activate the AP and to fix C3band initiate opsonization (7, 9, 22). We stratified patients byserotype into good or poor AP activators according to thesestudies. In contrast to what one would predict, the subgroupof patients with good activator serotypes had greater avail-ability of AP factors for Z activation. The basis for thedifferent abilities of PNC serotypes to activate the AP isunknown. One contributing factor may be the varying me-chanical barrier imposed by the particular configuration ofcapsular material. The cell wall binding sites for C3b,predominantly choline-TA (12, 26, 31), lie deep to thecapsular polysaccharides that are ineffective in activating AP(27). The varying ability of PNC serotypes to activeate APmay be a reflection of different exposure of the binding sitesto C3b.The third patient classification factor we examined was

bacteremia versus abacteremia. Our studies demonstratedthat bacteremia was not independently associated with avail-ability of AP factors for Z activation. This indicated that thebacteremic patients were a subgroup different from thepatients with complicated disease. Coonrod and Rylko-Bauer found greater depression of the AP in bacteremicpatients (6). However, those patients also had more severeand complicated infections. Bacteremia may not be the mostcritical element in determining AP activation. There areseldom more than several hundred pneumococci per ml inthe circulating blood of patients with PNC pneumonia (23).With respect to AP activation, the quantity of bacteria in theblood may be inconsequential compared to the quantity ofantigen released into the bloodstream or present in lung ormeninges. Bukantz et al. showed that mortality due to PNCdisease correlates more highly with circulating capsularpolysaccharide than with bacteremia (3). C-substance (TA),a potent AP activator, as well as capsular polysaccharidemay diffuse into blood independent of the presence of viableorganisms in the circulation (5, 15).

Significant morbidity and mortality from PNC diseasecontinue to occur despite rapid killing of PNC by antibiotics(17). An inflammatory process ensues, with systemic dis-semination of inflammatory substances in severe PNC infec-tion (11, 13, 21). This process may involve CRP. In thepreantibody phase of PNC infection, binding of the anti-body-like CRP to cell wall choline-TA activates the classicalcomplement pathway (4, 14, 24). Specific antibody inducesclassical pathway activation (18), resulting in generation ofC3b and amplifying the AP autocatalytic sequence (25). CRPmay similarly result in augmented AP activity. We observeda marked increase in concentration of CRP in acute sera ofpatients with PNC disease. In patients with complicatedillness, there was an inverse relationship between CRPlevels and availability of AP factors for Z activation. Thismay simply be independent consequences of greater severityof disease or numbers of microorganisms. We found normalC4 levels in patients, confirming the findings of others ofnormal early classical complement pathway components inPNC infection (6, 19, 20). Prellner noted that despite normalCl levels, there are elevated levels of activated Cl com-plexes in patients with PNC pneumonia or meningitis, andshe concluded that consumption of early components maybe masked by an acute-phase response (19). Similarly,

consumption of CRP may be masked by an acute-phaseresponse. Measureable CRP levels may also be influencedby competition with AP factors for common cell wall (TA)binding sites. Mold and Gewurz examined the in vitrointeraction of CRP with PNC-induced AP activation (16).Using C2 deficient sera, they found a dose-dependent inhi-bition of C3 conversion by PNC in the presence of CRP.However, in vivo, AP factors may have a temporal advan-tage in the competition for TA binding, permitting APactivation while blocking CRP binding. This may furtheraugment the levels of CRP seen in the presence of APactivation. Such dynamics may reconcile the differencebetween activation of the AP in the in vitro model (16) andthat proposed by us to occur in vivo.

ACKNOWLEDGMENT

This work was supported by an American Lung AssociationResearch Fellowship.

LITERATURE CITED1. Austrian, R., and J. Gold. 1964. Pneumococcal bacteremia with

especial reference to bacteremic pneumonia. Ann. Intern. Med.60:759-776.

2. Bakker-Woudenberg, I. A. J. M., J. Y. T. dejong-Hoenderop,and M. F. Michel. 1979. Efficacy of antimicrobial therapy inexperimental rat pneumonia: effects of impaired phagocytosis.Infect. Immun. 25:366-375.

3. Bukantz, S. C., P. F. DeGara, and J. G. M. Bullowa. 1942.Capsular polysaccharide in the blood of patients with pneu-mococcic pneumonia. Arch. Intern. Med. 69:191-212.

4. Claus, D. R., J. Siegel, K. Petras, A. P. Osmand, and H. Gewurz.1977. Interactions of C-reactive protein with the first componentof human complement. J. Immunol. 119:187-192.

5. Coonrod, J. D., and R. P. Leach. 1976. Antigenemia in fulminentpneumococcemia. Ann. Intern. Med. 84:561-563.

6. Coonrod, J. D., and B. Rylko-Bauer. 1977. Complement levels inpneumococcal pneumonia. Infect. Immun. 18:14-22.

7. Fine, D. P. 1975. Pneumococcal type-associated variability inalternative complement pathway activation. Infect. Immun.12:772-778.

8. Fine, D. P., S. R. Marney, Jr., D. G. Coiley, J. S. Sergent, andR. M. Des Prez. 1972. C3 shunt activation in human serachelated with EGTA. J. Immunol. 109:807-809.

9. Giebink, G. S., J. Verhoef, P. K. Peterson, and P. G. Quie. 1977.Opsonic requirements for phagocytosis of Streptococcus pneu-moniae types VI, XVIII, XXIII, and XXV. Infect. Immun.18:291-297.

10. Gross, G. N., S. R. Rehm, and A. K. Pierce. 1978. The effect ofcomplement depletion on lung clearance of bacteria. J. Clin.Invest. 62:373-378.

11. Henson, P. M., K. McCarthy, G. L. Larsen, R. 0. Webster,P. C. Giclas, R. B. Dreisen, T. E. King, and J. 0. Shaw. 1979.Complement fragments, alveolar macrophages, and alveolitis.Am. J. Pathol. 97:93-110.

12. Hummell, D. S., R. W. Berninger, A. Tomasz, and J. A.Winkelstein. 1981. The fixation of C3b to pneumococcal cell wallpolymers as a result of activation of the alternative complementpathway. J. Immunol. 127:1287-1289.

13. Johnston, R. B., Jr. 1981. The host response to invasion by S.pneumoniae: protection and the pathogenesis of tissue damage.Rev. Infect. Dis. 3:282-288.

14. Kaplan, M. H., and J. E. Volanakis. 1974. Interaction ofC-reactive protein complexes with the complement system. J.Immunol. 112:2135-2147.

15. Kenny, G. E., B. B. Wentworth, R. P. Beasley, and H. M. Foy.1972. Correlation of circulating capsular polysaccharide withbacteremia in pneumococcal pneumonia. Infect. Immun.6:431-437.

16. Mold, C., and H. Gewurz. 1981. Inhibitory effect of C-reactiveprotein on alternative C pathway activation by liposomes and

J. CLIN. MICROBIOL.

on March 4, 2020 by guest

http://jcm.asm

.org/D

ownloaded from

COMPLEMENT ACTIVITY IN PNEUMOCOCCAL INFECTION

Streptococcus pneumoniae. J. Immunol. 127:2089-2092.17. Mufson, M. A., D. M. Kruss, R. E. Wasil, and W. I. Metzger.

1974. Capsular types and outcome of bacteremic pneumococcaldisease in the antibiotic era. Arch. Intern. Med. 134:505-510.

18. Polhill, R. B., Jr., S. L. Newman, K. M. Pruitt, and R. B.Johnston, Jr. 1978. Kinetic assessment of alternative comple-ment pathway activity in a hemolytic system. II. Influence ofantibody on alternative pathway activation. J. Immunol. 121:371-376.

19. Preliner, K. 1981. Complement in pneumococcal infections withvarying degrees of severity. Scand. J. Infect. Dis. 13:263-268.

20. Reed, W. P., M. S. Davidson, and R. C. Williams, Jr. 1976.Complement system in pneumococcal infections. Infect. Im-mun. 13:1120-1125.

21. Rytel, M. W., T. H. Dee, J. E. Ferstenfeld, and G. T. Hensley.1974. Possible pathogenetic role of capsular antigens in fulmin-ant pneumococcal disease with disseminated intravascular co-agulation (DIC). Am. J. Med. 57:889-896.

22. Stevens, C. G., R. C. Williams, Jr., and W. P. Reed. 1977.Classical and alternative complement pathway activation bypneumococci. Infect. Immun. 17:296-302.

23. Tilghman, R. C., and M. Finland. 1937. Clinical significance ofbacteremia in pneumococcic pneumonia. Arch. Intern. Med.59:602-619.

24. Volanakis, J. E., and M. H. Kaplan. 1974. Interaction ofC-reactive protein complexes with the complement system. J.

Immunol. 113:9-17.25. Winkelstein, J. A. 1981. The role of complement in the host's

defense against Streptococcus pneumoniae. Rev. Infect. Dis.3:289-298.

26. Winkelstein, J. A., A. S. Abramovitz, and A. Tomasz. 1980.Activation of C3 via the alternative complement pathway resultsin fixation of C3b to the pneumococcal cell wall. J. Immunol.124:2502-2506.

27. Winkelstein, J. A., J. A. Bocchini, Jr., and G. Schiffman. 1976.The role of the capsular polysaccharide in the activation of thealternative pathway by the pneumococcus. J. Immunol. 116:367-370.

28. Winkelstein, J. A., and H. S. Shin. 1974. The role of immuno-globulin in the interaction of pneumococci and the properdinpathway: evidence for its specificity and lack of requirement forthe Fc portion of the molecule. J. Immunol. 112:1635-1641.

29. Winkelstein, J. A., H. S. Shin, and W. B. Wood. 1972. Heatlabile opsonins to pneumococcus. III. The participation ofimmunoglobulin and of the alternative pathway of C3 activation.J. Immunol. 108:1681-1689.

30. Winkelstein, J. A., and A. Tomasz. 1977. Activation of thealternative pathway by pneumococcal cell walls. J. Immunol.118:451-454.

31. Winkelstein, J. A., and A. Tomasz. 1978. Activation of thealternative complement pathway by pneumococcal cell wallteichoic acid. J. Immunol. 120:174-178.

VOL. 23, 1986 61

on March 4, 2020 by guest

http://jcm.asm

.org/D

ownloaded from