Quantitative Determination of the Human Immune

Response to Immunization with Meningococcal Vaccines

EMIL C. GOTSCHLICH, MicHEL REY, RENETRiAU, and KENNmrmJ. SPARKSFrom The Rockefeller University, NewYork 10021, Faculte Mixte de Medicineet de Pharmacie, Dakar, Senegal, and Institut Merieux, Lyon, France

A B S T R A C T Radioactive antigen binding tests havebeen developed to measure quantitatively the antibodyresponse of 167 adults, 84 children, and 51 infants toseveral different preparations of group A and group Cmeningococcal polysaccharides. Almost all the adultsinjected responded and the geometric mean responseswere approximately 15 /Ag/ml of antibody protein inindividuals vaccinated subcutaneously with two prepara-tions of group A vaccine. The geometric mean antibodyconcentration after immunization with two preparationsof group C vaccine was approximately 35 Ag/ml. Mostchildren aged 7 yr responded to immunization with twogroup A vaccines, and their mean response was onlyslightly less than that seen in adults. There was no dif-ference between the subcutaneous and the intradermalroute if both were given with jet gun. The majorityof infants aged 6-13 months responded to a preparationof group A vaccine and the geometric mean titer wasapproximately 1.2 ug/ml. Adults, children, and infantsresponded significantly less to a preparation of groupA polysaccharide which was of low molceular weight.

INTRODUCTIONCurrently there are under development two meningo-coccal vaccines based upon the use of high molecularweight group-specific polysaccharides. Immunization ofmilitary recruits with the group C polysaccharide hasbeen shown to be effective in preventing meningococcaldisease (1), and in lowering the rate of acquisition ofthe nasopharyngeal carrier stage by group C meningo-cocci (2). Such information has not yet been obtainedfor the group A vaccine, in part because these strainsare exceedingly rare in the U. S. military population.The efficacy of this vaccine will probably need to betested in the African "Meningitis belt" where diseaseis caused primarily by group A organisms (3).

Received for publication 12 July 1971.

Administration of either the group A or the groupC vaccine to adults induces the formation of antibodiesbelonging to the three major immunoglobulin classeswhich have both bactericidal (4) and opsonic proper-ties (5). To date quantitative data is available on theresponse of children.' There is no published data on theimmune response of infants to these vaccines. Inasmuchas this age group stands to benefit most from meningo-coccal immunoprophylaxis (6), quantitative data oftheir serological response is vital.

The present report will describe the immune responseof adults, children, and infants to immunization withseveral preparations of meningococcal polysaccharidesmeasured by means of quantitative radioactive antigenbinding tests (7).

METHODSAntigens. Group A meningococcal vaccines lots A-5,

A-7, V-1, and group C meningococcal vaccines lots C-6 andC-7 were prepared as described by Gotschlich, Liu, andArtenstein (8). The group A polysaccharide contained inlots V-4 and V-5 was prepared by a modified procedure,employing cold phenol extraction to remove protein con-tamination.2 A-5 and C-6 were prepared at the Walter ReedArmy Institute of Research; lots A-7 and C-7 by E. R.Squibb & Sons, New York; and lots V-1, V-4, and V-5by Institut Merieux in Lyons, France. The molecular sizeof the different preparations was assayed by gel filtrationon Sepharose 4B (Pharmacia Fine Chemicals, Inc., Piscata-way, N. J.). The void volume of the column was deter-mined with Blue Dextran (Pharmacia Fine Chemicals, Inc.,Piscataway, Nr J.) and the total volume with tritiatedwater (New England Nuclear Corp., Boston, Mass.). Thedistribution coefficient (Kd8) was calculated from the peakelution volume of the major. peak of the polysaccharide (9).The column was monitored either by determination ofphosphorus by the method of Chen, Toribara, and Warner(10) in the case of the group A polysaccharide, or by theresorcinol method of Svennerholm (11) in the case of the

'Manuscript in preparation.2 Manuscript in preparation.'Abbreviation used in this paper: Kd, distribution coef-

ficient.

The Journal of Clinical Investigation Volume 51 1972 89

,, 2000

0

_ Ii oo - t

0 1 2 4 8 169g of antigen

FIGURE 1 Quantitative precipitin curves obtained with hu-man sera drawn 2 wk after immunization with meningococ-cal group-specific polysaccharide. The upper curve depictsreactions with group C antigen; the lower two curves withgroup A antigen.

group C antigen. The column was calibrated by determiningthe Kd of two preparations of group A polysaccharide. Themolecular weights of these two preparations had been de-termined by ultracentrifugal methpds and were respectively170,000 and 25,000.' The two distribution coefficients wereplotted against the logarithm of their molecular weightand the resulting line was used to estimate the averagemolecular weight of the vaccine preparations (12). Thissame procedure was used to obtain an estimate of the molec-ular size of group C vaccine preparations.

Quantitative precipitation analyses. Five sera obtained2 or 3 wk after immunization with group A antigen andsix sera obtained 2 wk after immunization with group Cantigen were centrifuged at 70,000 g for 60 min to floatthe lipids, and centrifuged at 35,000 g for 30 min to removeall insoluble protein. A precipitin curve was constructed byadding to 1 ml portions of serum 100 tul of antigen solutiondissolved in 0.2 M sodium EDTA pH 7.0 and containing1,000 U penicillin and 1 mg of streptomycin. The quantityof antigen added consisted of 1, 2, 4, 8, and 16 Ag. Thetest was carried out in duplicate in 3-ml conical centrifugetubes which were sealed with parafilm and stored for 1 wkat 4°C. The tubes were gently shaken daily. The precipi-tates were sedimented by centrifugation for 30 min at 2,000g at 40C and washed twice with 2-ml volumes of cold saline.The precipitates were dissolved in 500 Al of 0.1 N NaOHand the protein concentration determined by the Folin re-action employing an autoanalyzer (13) (Technicon Co.,Inc., Tarrytown, N. Y.). The test was quantitated by refer-ence to a standard curve determined simultaneously usingchromatographically purified human gamma globulin ob-tained from Pentex Biochemical (Kankakee, Ill.).

Radioactive antigen binding test zeith intrinsically labeledantigens. Radioactive group A polysaccharide was preparedby incubating group A meningococci, strain A-1, in Frantzmedium (14) supplemented with 1 mc of "4C-labeled sodium

'Manuscript in preparation.

acetate (New England Nuclear Corp., Boston, Mass.).Group C meningococci of strain C-11 were grown in thesame medium supplemented with 25 mc of tritium labeledsodium acetate (New England Nuclear Corp., Boston,Mass.). The polysaccharides were purified by the usualmethod (8). The recovered group A polysaccharide had aspecific activity of approximately 2,800 cpm/,ug and thegroup C had approximately 10,000 cpm/,ug.

The radioactive antigen binding test was performed byadding 100 ul of radioactive antigen (0.17 ,ug of group Apolysaccharide or 0.5 ,ug of group C polysaccharide) to 100,ul of serum, and after overnight refrigeration the globu-lins were precipitated by the addition of 200 ul of 80%saturated (room temperature) ammonium sulfate. After theprecipitate had been sedimented by centrifugation for 60min at 2,000 g, 100 ,ul of the supernate was dissolved in 10ml of a liquid scintillation fluor consisting of Liquifluor(Pilot Chemicals, Inc., Watertown, Mass.): Soluene (Pack-ard Instrument Co., Inc., Downers Grove, Ill.) : toluene (42:100: 858). The radioactivity was measured in a Beckmanliquid scintillation counter, Model LS 133 (Beckman In-struments, Inc., Fullerton, Calif.).

Radioactive antigen binding test with group A polysac-charide labeled with ...I. In order to be able to iodinategroup A polysaccharide, phenolic groups were inserted byactivating the polysaccharide with cyanogen bromide (15)and letting it react with tyramine.' Such polysaccharidecould be readily iodinated by the method of Hunter, Green-wood, and Glover (16) and polysaccharide with a specificactivity of approximately 1,000 cpm/ng was obtained. Adouble label technique was used in order to be able to dothe binding tests on very small quantities of sera. The prin-ciple of this method was to add 1 ,uc of 'Na per milliliterof antigen solution as a volume marker. To 10 jul of serumwas added 10 IA of antigen solution and after overnightequilibration, 20 ,l of 80% saturated (room temperature)ammonium sulfate was added and the precipitate sedimentedby centrifugation. An arbitrary portion of the supernatewas drawn off and discarded, taking care not to disturb theprecipitates. This precipitate with a variable volume ofsupernate overlying it was counted in a Model 4233, twochannel gamma spectrometer (Nuclear Chicago, Des Plaines,Ill.) with one channel set to count the sodium and the otherchannel the iodine. The iodine count represented the totalamount of antigen left in the tube, both that bound andthat present in the remaining supernate; the sodium countindicated the volume of supernate left. With these data itis possible to calculate the per cent of added antigen boundto antibody (17).

Three antigen concentrations containing approximately4.0, 0.4, and 0.04 ,ug of radioactive antigen per milliliterwere routinely used. All sera were heat inactivated for 30min at 56°C and, when called for, were diluted in heat in-activated fetal calf serum. All volumetric measurementswere done with Eppendorf pipets (Brinkmann Instruments,Inc., Westbury, N. Y.) and the test was carried out in dis-posable Microfuge tubes, and centrifuged in a Microfuge(Beckman Instruments, Inc., Fullerton, Calif.).

Statistical methods. Geometric mean antibody concentra-tions were calculated by taking an arithmetic average of theper cent binding. Inasmuch as the degree of binding islinearly related to the logarithm of the antibody concen-tration, this yields a geometric average of the antibody con-centrations. The significance of differences in mean antibodyconcentration were tested by the Student t test.

6 Manuscript in preparation.

90 E. C. Gotschlich, M. Rey, R. Triau, and K. J. Sparks

Human sera. Dr. M. S. Artenstein kindly provided thesera of military recruits vaccinated with preparation A-7,C-6, and C-7. The source of the sera on which quantitativeprecipitation studies were performed, as well as the seraobtained from recruits injected with A-5, has been describedbefore (2, 4). All other sera were obtained by personnel ofthe Department of Infectious Diseases of the School ofMedicine of Dakar, Senegal. All subjects were volunteers,or informed consent was obtained from their legal guard-ians. They were observed clinically for- 48 hr and rectaltemperatures were obtained before, and at 6, 24, and 48 hrafter vaccination on all individuals except the military re-cruits. No significant pyrexic responses were seen and onlyminor local reactions lasting for 24 hour were noted. APed-o-jet gun (Ver.nitron Corporation, Farmingdale, N. Y.)was used in some of the studies and was adjusted to injecta volume of 0.5 ml subcutaneously, or 0.25 ml intradermally.Serum was obtained from bleedings performed 2 wk afterimmunization. Both the infants and the children were ingood health. The ages of the infants ranged from 6-13months with more than half of them aged 9 months or less.

RESULTSQuantitative determination of antibody by precipita-

tion and by radioactive antigen binding capacity. Oneof the simplest methods for the quantitative determina-tion of antibodies is the radioactive antigen bindingtest developed by Farr (7). Inasmuch as it measuresthe primary interaction of antibody with antigen, it isaffected only by the concentration of antibody and bythe average affinity of this antibody. The results of theradioactive antigen binding test have usually been ex-pressed as the dilution of a serum which will bind acertain amount of antigen (antigen binding capacity,ABC) (7). However, because of the data available inthe literature on the human immune response to thepneumococcal polysaccharides (19) in terms of micro-grams of precipitating antibody, it was thought desirableto express the results obtained with the radioactive anti-gen binding test in terms of antibody concentration. Torelate antigen binding capacity to antibody concentration,

TABLE IPrecipitating Antibody in Sera from Volunteers Immunized

with Meningococcal Polysaccharides

Precipitating antibody tomeningococcal polysaccharide

ImmunizedSerum with Group A Group C

#g/ml*W. C. B. C Not done 199I. G. A and C 38 38J. S. A and C 16.5 43.5M. S. A. A and C 47 401E.C.G. AandC 140 127J. W. A and C 85 36

*,ug/ml of antibody protein.

E 1.800

c

o l.400-

C

c 1.0000~01-

00

0

0'

0 0.200J~

F0

I..

x

Fx

0

-x /ra

a

l0 30 50 70 90% Antigen binding

FIGURE 2 The antigen binding capacity of dilutions of fivesera obtained from volunteers immunized with group Ameningococcal polysaccharide. The per cent of added in-trinsically labeled (14C) group A polysaccharide bound byantibody is plotted against the logarithm of the antibodyconcentration as determined by quantitative precipitation.Serum: 0, I. G.; E, E. C. G.; X, M. A. S.; 0, J. W.;A, J. S.

the radioactive antigen binding capacity of several serawith known content of precipitating antibody was deter-mined. Sera were obtained 2 or 3 wk after intradermalimmunization of six volunteers. One was immunizedwith group C vaccine only (W. C. B.) and the otherswith both group A and group C polysaccharide. Theirimmune response has been documented previously (4)by means of passive hemagglutination, bactericidal ac-tivity, and immunofluorscence. The concentration ofantibody was determined by quantitative precipitationThree representative precipitin curves indicating thehighest, the lowest, and an intermediate response aredepicted in Fig. 1. The quantity of precipitating anti-body found in each serum is summarized in Table I.

The capacity of these sera and dilutions thereof tobind intrinsically labeled group A and group C antigenwas measured by the method of Farr as indicated inthe Methods section and it was found that there was alinear relationship between the per cent of antigen boundand the logarithm of the antibody concentration. Thislinear relationship appeared to be valid between thelimits of 10 and 90% antigen binding. The results areset forth in Figs. 2 and 3. The lines were drawn by themethod of least squares, and the correlation coefficientswere 0.94 in both instances.

The immune response of adults. The antigen bindingcapacity of sera obtained from 167 adults immunized withseveral different lots of meningococcal vaccines weremeasured and the quantity of antibody determined byreference to the calibration curves obtained with the

Immune Response to Meningococcal Vaccines 91

L

E 1.800

CP

cxo 1.4000

0)D 1.000

c

0U

V0.2

g' 0.200-J

_

+ 0

I-a

0a

oIF+ 0

IF

0 30 50 70 90

% Antigen binding

FIGURE 3 The antigen binding capacity of dilutions of sixsera obtained from volunteers immunized with group Cmeningococcal polysaccharide. The per cent of added in-trinsically labeled (8H) group C polysaccharide bound byantibody is plotted against the logarithm of the antibody-concentration as determined by quantitative precipitation.Serum: 0, J. W.; EO, W. C. B.; o, I. G.; A, E. C. G.;+, J. S.; X, M. A. S.

standard antisera (Figs. 2 and 3). The results are setforth in the scattergrams (Fig. 4) and summarized inTable II.

>8'84

C-6 S.C. C-7Before After Before

0 T~:5?o

70r_ ._

as60

>~ 50-oc 400

0

g' 30

a

-o

0

. 0

10

5

TABLE I IThe Antibody Responses of Adults Injected with Various

Preparations of Group A and Group CMeningococcal Polysaccharide

Geometric meanantibody

No., No., concentrationRoute of of of

administra- sub- fail- Pre-Vaccine Kd* tion j ects ures immune Immune

pg/miC-6 0.30 s. c. (needle) 22 0 6.2 33.5C-7 0.27 s. c. (needle) 19 0

S.C. V-4 I.D.After Before After

* *..

V-5 S.C. V-5Before After Before

-if

>3.0E 3.0

0.,* 2

0

.0a)

20

-0

*0

E 1.0- 0* 0.5a

.. 0

seemed the most suitable because of its simplicity andits versatility in terms of sensitivity. The quantity ofantigen bound by a serum is a function of the con-centration of antibody present and the average affinityof this antibody. In this study the radioactive antigenbinding technique was standardized by reference toantisera with a known content of precipitating antibodyto be able to convert the observed degree of antigen bind-ing to antibody concentration. Strictly, this is correctonly in the instance where the average association con-stant of the unknown serum and of the standard serumare identical. This criticism, of course, also applies to agreater or less extent to other immunological techniquessuch as quantitative precipitation. To explore the mag-nitude of the error introduced by ignoring differencesin affinity of different antibodies, the antigen bindingof dilutions of a number of sera with a known contentof precipitating antibody was determined and the resultsare summarized in Figs. 2 and 3. All the points tend tofall rather closely on the line derived by least squaresas is evidenced by the high correlation coefficients (r=0.94) obtained in both cases indicating that for thesesera the antigen binding capacity was closely correlatedwith the amount of antibody present as determined byquantitative precipitation. This suggests that little erroris introduced by this treatment and that the associationconstant of these adult sera do not differ widely enoughto be noticeable in this system.

The conversion of antigen binding capacity to anti-body concentration, aside from the potential errors dis-cussed above, has two advantages. It allows the resultsobtained in this study to be compared with the serologicaldata obtained by quantitative precipitation on volun-teers immunized with several pneumococcal polysac-charides. The other advantage is that the test is stan-dardized by reference to antisera whose potency remainsstable and which can readily be exchanged between dif-ferent laboratories.

The response of 167 adults to vaccination with groupA or group C meningococcal vaccines was measured. Itcan be concluded that the vast majority of adults in-jected with group C vaccine or with group A vaccine ofhigh molecular weight had immune responses whichquantitatively were comparable to those seen in indi-viduals immunized with various pneumococcal polysac-charides (18). The responses to the group A vaccineapparently are lower than those obtained with group Cvaccine. One preparation of group A vaccine, lot V-1,because of improper storage before packaging, was ofconsiderably lower molecular weight than the otherpreparations tested. The immune response to this vac-cine was significantly less, not only in adults but also inchildren and in infants. This is in accord with data

obtained with various preparations of dextran of differ-ent molecular weight (19).

The response of children to two lots of group A vac-cine of very high molecular weight, V-4 and V-5, werestudied. The vaccine was administered by jet gun eitherintradermally or subcutaneously. In both instances the re-sponses were comparable to those observed in adultsinjected subcutaneously by needle with vaccine lotV-4, and there was no advantage to intradermal in-jection. These data do not exclude the possibility thatintradermal injection by needle may give a higherresponse than subcutaneous injection as suggested bythe results obtained with adults, because there is con-siderable leakage from an intradermal injection sitewhen performed with a jet gun.

Infants were injected with lot V-1 and V-4 and theresponse to the larger molecular weight vaccine was sig-nificantly higher. The mean response of these infants,who ranged in age from 6 to 13 months, was approxi-mately tenfold lower than those observed in adultsand children. The reasons for this lower response are atpresent unknown. The simplest possibility is that theinfants are having a primary immune response, whereasthe children are having an anamnestic response. It istrue that the preimmune antibody levels of the infantsare much lower than those seen in children and thatalmost half had no detectable antibodies (less than 0.1/g/ml). This thesis can easily be tested in future stud-ies by reimmunization of infants and observing whetherthey then produce a response akin to those seen inchildren or adults.

Controls consisting of infants not immunized a sec-ond time but studied over a longer period of time wouldresolve the question whether infants need more than2 wk to mount their maximal immune response.

The crucial question is whether the immune responsethat the infants did mount is adequate to protect themfrom meningococcal disease. This will be answered de-finitively only by a large field trial. Nevertheless, it isa fact that agammaglobulinemic children are protectedagainst meningococcal disease by passive immunopro-phylaxis. It can be calculated that with the usual dose(0.7 mVkg body weight) of concentrated gammaglobu-lin, and taking into account that the average concentra-tion of antibody in adult sera is approximately 5 tg/ml,the concentration of antipolysaccharide antibody in thepatient's blood stream would be less than 1.0 /g/ml.This consideration suggests that the low responses ofthe infants may be sufficient to be protective.

ACKNOWLEDGMENTSWe would like to thank Dr. Branko Cvjetanovic of theWorld Health Organization, for making this study pos-sible and for his help in the preparation of this report.

Immune Response to Meningococcal Vaccines 95

Supported in part by the Commission on Acute Respira-tory Diseases of the Armed Forces Epidemiological Boardunder Contract No. DADA 17-70-C-0027 and by a grantfrom The World Health Organization.

REFERENCES

1. Artenstein, M. S., R. G. Gold, J. C. Zimmerly, F. A.Wyle, H. Schneider, and C. Harkins. 1970. Preventionof meningococcal disease by Group C polysaccharidevaccine. N. Engl. J. Med. 282: 417.

2. Gotschlich, E. C., I. Goldschneider, and M. S. Arten-stein. 1969. Human immunity to the meningococcus. V.The effect of immunization with meningococcal GroupC polysaccharide on the carrier state. J. Exp. Med.129: 1385.

3. Lapeyssonie, L. 1963. La meningite cerebro-spinale enAfrique. Bull. World Health Organ. 28(Suppl. 1): 1.

4. Gotschlich, E. C., I. Goldschneider, and M. S. Arten-stein. 1969. Human immunity to the meningococcus.IV. Immunogenicity of Group A and Group C meningo-coccal polysaccharides in human volunteers. J. Exp.Med. 129: 1367.

5. Roberts, R. B. 1970. The relationship between Group Aand Group C meningococcal polysaccharides and serumopsonins in man. J. Exp. Med. 131: 499.

6. Goldschneider, I., E. C. Gotschlich, and M. S. Arten-stein. 1969. Human immunity to the meningococcus. I.The role of humoral antibodies. J. Exp. Med. 129:1307.

7. Farr, R. S. 1958. A quantitative immunochemical mea-sure of the primary interaction between I*BSA andantibody. J. Infec. Dis. 103: 239.

8. Gotschlich, E. C., T. Y. Liu, and M. S. Artenstein.1969. Human immunity to the meningococcus. III.Preparation and immunochemical properties of the

Group A, Group B, and Group C meningococcal poly-saccharides. J. Exp. Med. 129: 1349.

9. Ackers, G. K. 1964. Molecular exclusion and restricteddiffusion processes in molecular-sieve chromatography.Biochemistry. 3: 723.

10. Chen, P. S., Jr., T. Y. Toribara, and H. Warner. 1956.Microdetermination of phosphorus. Anal. Chem. 28:1756.

11. Svennerholm, L. 1957. Quantitative estimation of sialicacids. II. A colorimetric resorcinol-hydrochloric acidmethod. Biochim. Biophys. Acta. 241: 604.

12. Andrews, P. 1964. Estimation of the molecular weightsof proteins by Sephadex gel filtration. Biochem. J. 91:222.

13. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.Randall. 1951. Protein measurement with the folinphenol reagent. J. Biol. Chem. 193: 265.

14. Frantz, I. D. 1942. Growth requirements of the menin-gococcus. J. Bacteriol. 43: 757.

15. Axen, R., J. Porath, and S. Ernback. 1967. Chemicalcoupling of peptides and proteins to polysaccharides bymeans of cyanogen halides. Nature (London). 214:1302.

16. Greenwood, F. C., W. M. Hunter, and J. S. Glover.1963. The preparation of 'I-labeled human growth hor-mone of high specific radioactivity. Biochem. J. 89: 114.

17. Gotschlich, E. C. 1971. A simplification of the radio-active antigen-binding test by a double label technique.J. Immunol. 107: 910.

18. Heidelberger, M. 1953. Persistence of intibodies in manafter immunization. In The Nature and Significance ofthe Antibody Response. A. M. Pappenheimer, Jr., edi-tor. Columbia University Press, New York. 90.

19. Kabat, E. A., and A. E. Bezer. 1958. The effect ofvariation in molecular weight on the antigenicity ofdextran in man. Arch. Biochem. Biophys. 78: 306.

96 E. C. Gotschlich, M. Rey, R. Triau, anid K. J. Sparks