Variable region expression in the antibody responses of infants vaccinated with Haemophilus influenzae type b polysaccharide protein conjugates Description of a new lambda light chain associated idiotype and the relation between idiotype expression, avid

Variable region expression in the antibody

responses of infants vaccinated with

Haemophilus influenzae type b

polysaccharide-protein conjugates. Description of a new

lambda light chain-associated idiotype and the

relation between idiotype expression, avidity,

and vaccine formulation. The Collaborative

Vaccine Study Group.

D M Granoff, … , S J Holmes, A H Lucas

J Clin Invest.

1993;

91(3)

:788-796.

https://doi.org/10.1172/JCI116298

.

Haemophilus influenzae b polysaccharide (Hib PS)-protein conjugate vaccines differ

chemically and immunologically. To determine whether anti-Hib PS variable region

expression might differ according to vaccine formulation, infants were vaccinated at 2, 4,

and 6 mo of age with Hib PS coupled to either meningococcal outer membrane protein

complex (Hib PS-OMPC) or tetanus toxoid (Hib PS-T), or Hib PS oligomers coupled to a

mutant diphtheria toxin (Oligo-CRM). Two anti-Hib PS idiotypes were measured in sera

obtained after the third injection: HibId-1, expressed by anti-Hib PS antibodies having the

kappa II-A2 variable region, and HibId-2, a newly defined cross-reactive idiotype associated

with a subset of anti-Hib PS antibodies having lambda VII variable regions. HibId-1 was

present in 33, 68, and 64% of infants given either Hib PS-OMPC, Oligo-CRM, or Hib PS-T,

respectively (P < 0.001). The respective values for HibId-2 were 47, 18, and 10% (P =

0.001). Subjects who were vaccinated with Hib PS-OMPC or Hib PS-T and who produced

detectable HibId-1-positive antibody, had significantly higher mean antibody avidity than

subjects who did not produce HibId-1 positive antibodies. In contrast, Oligo-CRM evoked

high avidity anti-Hib PS antibodies, irrespective of the idiotypic profile. These findings

Research Article

Variable Region Expression in the Antibody Responses of Infants Vaccinated

with Haemophilus

influenzae Type b Polysaccharide-Protein Conjugates

Description of a New

X Light Chain-associated Idiotype and the Relation

between

_Idiotype

_{Expression, Avidity,}

and Vaccine

Formulation

Dan M.Granoff,* Penelope G.Shackelford,* Sandra J. Holmes,* TheCollaborative Vaccine Study Group, and Alexander H.Lucast

*Division ofInfectiousDiseases, Edward

Mallinckrodt

Department ofPediatrics, Washington University School ofMedicine, St. Louis Children'sHospital,St. Louis, Missouri 63110;and

*Children's

Hospital Oakland Research Institute, Oakland, California 94609

Abstract

Introduction

Haemophilus

influenzae b

polysaccharide (Hib PS)-protein

conjugate vaccines differ chemically andimmunologically.To

determine whether anti-Hib PS variable region expression

might

differ

accordingtovaccineformulation, infantswere

vac-cinatedat2, 4,and6moof age with HibPScoupled to either

meningococcal outer membrane protein complex (Hib

PS-OMPC)or tetanus toxoid (Hib PS-T),orHib PSoligomers coupled to a mutantdiphtheriatoxin(Oligo-CRM).Two

anti-HibPSidiotypeswere measured in seraobtainedafter the third

injection:Hibld-1,expressed byanti-HibPSantibodies having theKII-A2variableregion,and

Hibld-2,

anewlydefined

cross-reactive idiotype associated witha subsetofHib PS

anti-bodieshaving XVII variable regions. HibId-1waspresentin33,

68,

and 64% ofinfants

given

either Hib PS-OMPC, Oligo-CRM,orHib

PS-T,

respectively (P<0.001). The respective

valuesfor

Hibld-2

were47, 18,and10%(P=0.001). Subjects who were vaccinated with Hib PS-OMPC or Hib

PS-T

and who produceddetectable

Hibld-1-positive

antibody,had

signifi-cantly

higher

meanantibody aviditythansubjects who didnot produce

Hibld-1

positive antibodies. In contrast,Oligo-CRM

evokedhigh avidity anti-Hib PS antibodies, irrespective ofthe

idiotypic profile. These findings indicate fundamental

differ-ences inbothvariableregioncontentand

antibody

quality elic-ited by different HibPSconjugate vaccines.

(J.

Clin.Invest.

1993.91:788-796.) Key words:

antibody

functional

activity

cross-reactive

idiotype

*

antibody

repertoire

*vaccine

The Collaborative Vaccine Study Group consisted of the following individuals who enrolledchildren in theprospectiveimmunogenecity trials andprovidedserumsamples forourstudy:T. V.Murphy, Univer-sityofTexas, Southwestern, Dallas;M. T.Osterholm and J.McHugh, Minnesota Health Department, Minneapolis; E. L. Anderson and

R. B.Belshe, St. LouisUniversity School ofMedicine;J.Esbenshade, University of Utah, Salt LakeCity; M.Blatter, Upper St.Claire, PA; andH.Keyserling,Atlanta, GA. K.P. Guito and C.K.Meschievitz, ConnaughtLaboratories, Inc.,Swiftwater, PA, organizedand super-vised oneof the multicenter studies.

AddresscorrespondencetoAlexanderH.Lucas,Children's

Hospi-talOakland ResearchInstitute,74752ndStreet, Oakland,CA94609. Receivedfor publication 12December 1991 andinrevisedform8 October1992.

Successful induction ofantibodyresponses tothe Haemophi-lus

influenzae

typebcapsularpolysaccharide (Hib PS)'in

in-fants requirestheconjugation ofproteintothe HibPS ( 1-3).A

varietyof conjugate vaccine formulations havebeen prepared

that differinpolysaccharide size, protein carrier, and covalent linkage (3). These structural differences result in vaccines with

differentimmunological properties. Forexample,theHib PS-outer membraneprotein complex (Hib PS-OMPC) vaccine, whichconsists of high molecular weight Hib PS covalently

cou-pled toNeisseria

meningitidis

outermembrane

protein

(4), stimulatesserumanti-Hib PSantibody productionaftera sin-gle injection in infants as young as 2 mo of age, but stimulates

onlymodest increases in serumantibodyuponsubsequent

im-munizations at 4 and 6 mo

of

age (2, 5-7). In contrast, a

vaccine consisting of Hib PS oligomers coupled to nontoxic

mutantdiphtheria toxin (Oligo-CRM), doesnotelicit

signifi-cantHib

PS-specific antibody

aftera

single

injection

at2moof

age (7), but primes foran anamnestic response as antibody production is stimulatedupon twofurther immunizationsat

intervalsof1-2 mo

(6-8).

Another

striking

differencebetween

thesetwoconjugatevaccines is therequirementfor carrier

im-munization. WhenOligo-CRMis administeredtoinfant rhe-sus

monkeys

without additional diphtheria toxoid

(DT)

carrier

vaccination,

anti-Hib PS

antibody

responses are poor, even

aftertwo orthree

injections (9).

Similarresults have been

re-ported in human infants

(10). However,

immunization with Hib

PS-OMPC,

which hasacarrier

protein

unrelatedtoDT, induces

high antibody

levelsinthe absence ofDT vaccination

(9, 10). Furthermore, arecentstudy shows that

Oligo-CRM

and Hib PS-tetanus toxoid

(Hib PS-T) conjugate

vaccines

elicit antibodiesof

higher

mean

avidity

than does the Hib

PS-OMPC

conjugate ( 11).

Taken

together,

these

findings

suggest that these different Hib PS

conjugate

vaccines elicit

specific

antibody responsesby different mechanisms.

Previous data fromanumberof laboratoriesindicate that

the human

antibody repertoire

toHib PS is of limited

struc-tural

diversity. IgG

antibodiesare restrictedtothe

IgG

1 and

IgG2subclasses(12,

13),

Kchain antibodies

predominate

over

1.Abbreviations usedinthis paper: CRI, cross-reactive idiotype; DT, diphtheriatoxin; H, heavy; Hib PS, Haemophilus influenzae type b capsular polysaccharide; Hib PS-OMPC, Hib PS-outer membrane complex; HibPS-T, Hib PS-tetanus toxoid; Id, idiotype; L, light; Oligo-CRM, Hib PS oligomers coupled to nontoxic mutant DT; RABA, radioantigenbinding assay; V, variable.

J.Clin.Invest.

© The AmericanSocietyforClinicalInvestigation,Inc.

0021-9738/93/03/07881/09 $2.00

X chainantibodies (14, 15), and a limited number of serum anti-Hib PS clonotypes are resolvable by isoelectric focusing (12, 16). Anti-Hib PS antibody heavy (H) chain variable(V) regions are usually oftheVH3 subgroup and are encoded by the VH 9.1 or VH 26 genes, orclosely relatedgenes(17-19).

Anti-Hib PS VL regions may be KI,KII,KIII,KIV orX,but the mostcommonly expressed VL region is derived from the VKII gene, A2 (20-22). Further, a cross-reactive idiotype (CRI), designatedHibld- 1, is a marker for anti-Hib PS antibodies

hav-ing theVKIIA2light(L) chain (22), and HibId- 1-positive anti-bodies comprise on the average more than half of the

vaccine-induced serum anti-Hib PS induced in older children and

adults(22-23).

With the exception ofa few infants vaccinated with Hib PS-OMPC (23), no

information

isavailableontheexpressed idiotypic/Vregion

antipolysaccharide repertoire

ininfants less than7 moofage.Infants ofthisage areofparticular interestfor

threereasons.First,theyaretheprincipaltarget

population

for conferring protection by vaccination with bacterial

polysaccha-rides. Second, theyare

unresponsive

to vaccination with

un-conjugated polysaccharides and the mechanism(s) of this an-ergy is unknown.Third, because of theiryoung ages, the

anti-body repertoire expressed after vaccination of infants is

unlikelyto have been affected by prior

environmental

expo-sure toHib PS.

The

availability

of Hib PS

conjugate vaccines for

use in

infants 2-6 mo ofage allows usto access the

early

antibody repertoire

specific for

a clinically relevant polysaccharide. In

the present study, we have examined

conjugate

vaccine-in-duced antibodies of infants for

expression

oftwo idiotypes: HibId- 1, andanewly definedCRI,

Hibld-2, which

serves as a

serological marker fora subset of

_{V,-containing}

anti-Hib PS

antibodies. Our

findings

demonstratethat therelative

expres-sionof thesetwoCRIsis dependentupon the

conjugate

vac-cine formulation. In

addition,

we show that the anti-Hib PS

antibody avidity differences

elicited by vaccination withtwoof the

conjugate

vaccines correlate with the relative

expression

of thesetwo

idiotypes.

Methods

Subjects andvaccines.Serafrom 189healthyinfants,whowere vacci-natedat2,4, and 6 moofagewitheitherHib PS-OMPC (PedVAX-HibR; MerckSharp and Dohme, West Point, PA) (n =43),

Oligo-CRM (HIBTiterR; Praxis Biologics, Rochester, NY) (n=66),orHib PScoupled totetanustoxoid (ActHIBh; Merieux Institute,Lyon, France) (n = 80),wereavailable for this study. Bloodwasobtained immediatelybefore and 1moafterthe third dose of vaccineaspart of ongoing comparativeimmunogenicity trials of various conjugate

vac-cines(7, 7a), and serumanti-Hib PS antibody concentrationswere

determined. However, because of the small volumes of pre-third dose vaccinationseraavailable, the additional studies described belowwere

limitedtothe post-third dose vaccination samples. Selection of these

sera wasbasedonavailability of sufficient quantity, and the presence of

>0.8

_Ag/ml

of anti-Hib PS antibody, the minimal concentration neededtoperform the multiple immunologic assaysonthe small

vol-umesofseraavailable.

The characteristics of the subjects andserumantibody

concentra-tionsaresummarized in Table I. Group 1 received a standard dose consisting of 15 ,g of Hib PS and 256 Mg ofprotein, usingasingle lot of HibPS-OMPC (lot 238 1 R). Group 2 receivedastandard dose of multi-ple commercial lots of Oligo-CRM, containing 10

_,g

of Hib PS and 25

tgofprotein.

Group

3receivedoneof three

investigational

lots of Hib PS-T (lots S2141, S2181, and S2189), usingadose of 10,ugofHib PS and 20tigofprotein. All vaccines were stored at 2°-8° C° and 0.5 ml doses were administered intramuscularly in the thigh. Coded serawere

stored at-20°C in multiple vials. All assayswereperformed blindly without knowledge of the vaccine assignment.

Measurementofanti-Hib PSantibody.Totalantibodyto Hib PS was measuredin aradioantigenbinding assay (RABA) using '25I-la-beledHib PSasdescribedpreviously (24). Antibodyconcentrations were determined by a standard curve generated using dilutions of the Office of Biologicsreferenceserumpool. Anti-Hib PSantibodyavidity wasmeasuredusing a modified RABA, as described previously (25). Briefly,theavidityconstantisdetermined by measuring the Hib PS binding capacity ofa serum usingtwoconcentrationsof'25I-Hib PS, 0.0025nM and0.1 nM.Theavidityconstantisgiven by the relation-ship,K=

PI

{ ( 1-P)X[Ab] }, whereKistheavidityconstant,[Ab]is the concentration of anti-Hib PS antibody molecules (mol wt

=_{150,000) determined from the RABAusing}0.1 nM'25I-HibPS, and Pis the fraction ofbound Hib PS molecules determined from the RABAusing0.0025 nM'25I-HibPS.Kis calculatedfromseveralserum

dilutionshaving P valuesbetween 20and80%,and thefinal avidity value,in nM-',wasassigned based on the average.

Humanmonoclonal anti-Hib PSantibodies.Thefollowinghuman monoclonal anti-HibPS antibodieswereused in this study: LSF-2 (IgA,, X), ED8.4 (IgM, K), RAY4 (IgA2, X), SB5/D6(IgA1,X), ED6. 1 (IgG2, K), JB21 (IgG2, X), JB32 (IgA,, X), ANN2 (IgA, K), ANN23 (IgM, K), CB6 (IgA2, X), CBl (IgA, K), and CA4 (IgG2, K).Withthe exception of CA-4, ANN2, ANN23, CB6, and CB 1, the isolation, char-acterization,and sequence analysesofthese mAbs are reported else-where (18, 26).

TableI.Characteristicsofthe Subjects

Serum anti-Hib PSantibody* geometricmean

Ageatthird

Subject No. vaccination Pre-third Post-third

group Vaccine tested mean±SD White Male vaccination vaccination

months % % '.g/ml

1 Hib PS-OMPC 43 6.3±0.3 98 47 8.2 10.3

2 Oligo-CRM 66 6.3±0.5 97 59 0.9 10.8

3 Hib PS-T 80 6.4±0.5 99 44 2.1 12.6

ProductionofmAbantiidiotypic antibody(LuC10)to mAb LSF-2. BALB/c mice were immunized in several subcutaneous sites with hu-man mAb LSF-2emulsifiedin CFA. A single dose consisted of 20

,g

of purified antibody.3 wk later themice wereimmunizedsubcutaneously with the same dose of LSF-2 in incomplete adjuvant. 5 wk later a third injection of 20

_jig

of LSF-2 was administered intraperitoneally, and the spleen was taken 3 d later and used for hybridoma production. Hybrid-omas were preparedusing establishedmethods(27),byfusingspleen cells to the X63-Ag8.653 cell line (28) with 50% polyethylene glycol. A solid phaseradioimmuneassaywas used to testhybridoma superna-tantsforantiidiotypic activity. This assay wasessentiallythe sameas that described in reference (22), except that microtiter wells were coatedwithpurified LSF-2 mAb insteadof purifiedserumanti-HibPS antibody.Culture supernatants were screened foranti-LSF-2idiotypic activity bymeasuringtheability ofthe supernatant toinhibitthe bind-ingof

'251I-Hib

PSto the solid phase adsorbed LSF-2 antibody.After subcloning bylimiting dilution,ahybridomaclone wasisolatedthat secreted a mAb (IgG1,K),designatedLuC10,thatcompletelyinhibited thebinding ofHib PS to mAb LSF-2. MAbLuC1Owasisolated from ascitesfluidby ammoniumsulfate fractionationandion exchange chro-matography. Alkaline phosphatase was coupled toLuCOusing glutar-aldehyde.

Characterization of mAbLuC10. A previously described ELISA was used to assess the binding ofLuC1Oto human Ig (23). Briefly, wellsofmicrotiterplates werecoated witheitherhumananti-Hib PS mAbs, human myelomaproteins(providedby Dr. HansSpiegelberg, UniversityofCalifornia,SanDiego),orHib PS-absorbedpooled intra-venous gammaglobulin. Wells were blocked with albumin, washed, and then reactedwithvarying concentrations ofLuC10-alkaline phos-phatase. Afterseveral hoursofincubation,boundLuC 10 was detected with p-nitrophenyl phosphatesolution.

Theabilityof Hib PS to inhibit thereactionbetweenLuC1Oand LSF-2 was measured in ELISA(23).Solidphaseadsorbed LSF-2 was reacted withvarying concentrations of HibPS, followed byaddition of LuC10-alkalinephosphatase. Bound LuC0 was detectedusing appro-priatesubstrate, andoptical densitydetermined after - 30 min. The percent inhibition ofLuC1O bindingbyHib PSwascalculatedas previ-ouslydescribed(22).

Thereaction between mAb LuC 10 and human Ig was also exam-inedbyELISA. Microtiter wells were coatedovernightwith LuC0 (10

qg/ml

PBS), blockedwithalbumin,and then reacted with postvaccin-ationserumsamplesthathad been absorbed with either Hib PS-Seph-arose orunmodified Sepharose. Thewellswerewashed andalkaline phosphataseconjugated antiisotypic antibodies were addedfor 3 h. Afterwashing,substrate solution was added and absorbanceat405nm

was

measured

after - 30 min.

100

._.

iE 80-fA

._

_ 60'

In

0

o

40-._

* 20

LUC10

u

:2CD

._

0._

:2

-._

.1

Assaysof anti-Hib PS Hibld-J andHibld-2expressionby LuC9 and LuC10 inhibition ofHibPSbinding. Hibld-Iand Hibld-2 expres-sion by total serum anti-Hib PSantibodieswasquantifiedby measur-ing thedegree to which the respective monoclonal antiidiotypic anti-body inhibited the binding of1251I-HibPS inthe RABA. This assay has beendescribed previously in detail (22, 23). Briefly, a dilution of serum oranti-Hib PS mAb,equivalentto ananti-HibPSconcentration between0.8 and 2.0

gg/ml,

wasmixed withanequal volumeof either LuC9 (60

_,g/ml),

LuC1O (60 ug/ml), or diluent. After incubation, theHib PSbinding activitywasdeterminedbyadditionof - 0.1nM

'251I-HibPS,followed by addition of 50% saturated ammonium sulfate. Inhibition of'251I-HibPS bindingby the anti-Id was determined as described previously (23). An ELISA was used to measure CRI expres-sion by K and X anti-Hib PS antibodies. Postvaccination sera were screened forK and X anti-Hib PSantibodies usingmicrotiterplates coated with Hib PS-poly-L-lysine andalkaline phosphatase coupled goatantibodiesspecific for human K or Xchains(Tago Inc., Burlin-game,CA). 46 sera were selected for idiotypicanalysisbasedon suffi-cient quantity,andthepresenceof22.0Mg/mlof K- and/or X-specific anti-Hib PSantibody. 19 of these sera hadsufficientK and Aanti-Hib PSantibodyto measure CRIexpression in relation to each L chain; an additional 16 had sufficient Kantibodybut notA, and 11 hadsufficient A but not Kantibody.This assay,which has been described previously for measurement ofIgGIandIgG2-specificCRI (23), entails incubat-ing dilutions ofhumanserawith afixed concentration of LuC9 or LuC10,and thentransferringthese sera tomicrotiterwellscoated with Hib PS-poly-L-lysine. Bound antibody was detectedwithalkaline phos-phatasecoupled goat antibodiesspecific foreither human K or A chains (Tago,Inc.),andinhibition of HibPSbindingby the anti-Id was calcu-lated aspreviously described (23).

Results

mAb LuC10reacts witha

combining

site-associated

idiotypic

determinant

of

humanmAb LSF-2. Asdescribedin

Methods,

a murine mAb designated LuC10 (IgG1, K) was isolated that inhibited the binding of human mAb LSF-2

_(IgAj,

X)to 125I-Hib PSinasolidphase radioimmune assay. Fig. 1 showsthe dose-dependent inhibitionof

1251-Hib

PSbinding byLuC10in a conventional liquid phase RABA. For comparison, previ-ously publishedinhibition data with mAb LuC9 are also shown

(22). In thisassay a fixed concentration(0.2

,g/ml)

of anti-Hib PS mAbs LSF-2orED8.4,wasreacted withvarying con-centrations of LuCI0orLuC9,followedbyaddition of

radiola-LUC9

.01 .1 1

mAb (ug/ml)

Figure1.mAbLuC1Oand mAb LuC9 inhibition of'25I-HibPSbinding byhumanmAbsLSF-2 (o) and ED8.4(A)asdetermined in the

radioan-___0_0 tigenbindingassay.Afixed concentration(0.2

1g/ml)of anti-Hib PSmAbs,LSF-2orED8.4,was 10 100 _{reacted with varying concentrations of anti-Ids,}

LuclOorLuc9(x-axis).

beled Hib PS and determination of binding activity. LuC 10

inhibited 100%of LSF-2 Hib PS binding activity at

concentra-tionsas low as 0.1

tg/ml.

Incontrast, no inhibition of ED8.4 (IgM, K) Hib PSbindingwasobserved.ED8.4 is a human mAb

specificfor Hib PS and expresses the CRI, Hibld- 1, associated

withK L chains having the A2 V region (22). With LuC9, the reverse pattern wasobserved: >90% inhibition ofED8.4 Hib PS binding but no inhibition of LSF-2 Hib PS binding.

Consid-erably higherconcentrations of LuC9 were needed to inhibit

ED8.4comparedtothatof LuC IO needed to inhibit LSF-2.

The specificity ofLuC 10 was examinedbydirectELISA,

where LuC10-alkaline phosphatase was incubated in

micro-titer wells thathad been coated with eitherLSF-2, ED8.4, hu-man myeloma

proteins (representative

of the

IgM,

_IgG1,2,3,4,

IgA1,

and IgDHchainisotypes,and K and X L chaintypes), or

pooled normal y-globulin absorbed with Hib PS. LuC 10

bound onlyto wells coated with LSF-2(datanotshown).Fig.2

showsthat thebindingof LuC10tosolidphaseadsorbed LSF-2 wasinhibited 90% by 10

,g/ml

HibPS, and that 50%

inhibi-tionwasachievedbetween .01 and 0.1 og/mlof Hib PS.

Collectively, these data indicate that mAb LuC10 reacts

withanidiotope (Id)on humanmAb LSF-2 that is associated with the Hib PS-binding site, and that is distinct from the

previously described major CRI, Hibld-1. Hereafter we will

refertothisnewIdasHibld-2.

HibId-2

is a cross-reactive

idiotype

associated with

anti-Hib

PS

antibodies

having

X L

chains.

Hibld-2 isnot a

private

Id restrictedtomAbLSF-2, but is

expressed by

heterologous

anti-Hib PS antibodies having only X Lchains. Thisfinding was

demonstrated inanELISAwhere LuC10wastested for its

abil-ityto

inhibit

the

binding

of eitherKorXserumantibodies to

Hib PS evoked by vaccination of infantswith different

conju-gatevaccines. Inthisassay, >20% inhibition is consideredto

be

positive

for

idiotype expression (>

3 SD above

background

inhibition).

Asshown in

Fig.

3A, LuC10 variablyinhibited

onlyXanti-HibPS antibodies. This resultimpliesthatasubset

of

X-bearing

serumanti-HibPS antibodies express Hibld-2. In contrast,aspreviouslydescribed(22), Hibld-l (asdefinedby LuC9

inhibition of

Hib PS

binding)

wasconfinedto asubset of

anti-Hib

PSserumantibodies

having

KLchains

(Fig.

3

B).

Further

evidence

that Hibld-2 wasrestrictedto

X-bearing

anti-Hib PS antibodieswas obtained

using

an ELISA where

wells coated with mAb LuC10werereactedwith six

postvac-cinationsera thathad been absorbedwitheither Hib

PS-Seph-arose or as a

negative control,

unmodified

Sepharose.

The

100 us 80 ._-c 60 40 0-= 20 c 68

Figure 2. Hib PS inhibi-tion of the reacinhibi-tion

be-tweenmAbLuClOand humanmAbLSF-2. Microtiter wells,coated

withmAbLSF-2,were

reacted withalkaline phosphatase-coupled LuClOin theabsence orpresenceof various

dosesof Hib PS, andthe percentinhibitionwas 10 determined as described

(22). 100 -00 0. o 8 20 .= 60-m o 40-c

* 20

-eS .0 @0 00 -~~ 00 ...

*0

r

.

fHi

A 100-D 80--0

.0 60

-m 00 0. 40 I. c 402-c 0 @ 20 B 00 . .

*-u00

0*0

.01

.0 *. %*e akA&

X _K x K

Figure 3.Light chainrestrictionofcross-reactive idiotypes. Percent inhibitionofKor Xanti-Hib PSbindingby LuC 0 (A)orLuC9 (B) wasdeterminedby ELISA as described in Methods.Postvaccination

serafrominfantsimmunized withdifferentHib PSconjugates (Table I)were selected based on the presence of 2 2.0

kog/ml

ofKand/or X-specificanti-Hib PSantibodyasdetermined in ELISA.

LuC 10-reactiveIg was detected with enzyme-conjugated

anti-bodies

specific for

K or X L chainsorIgG, IgA, or IgM. The

results obtained withtwo serum samples are shown in Fig.4.

MAbLuC10 reacted exclusively withX Ig and thisreactivity

wasremoved by absorptionwith Hib PS-Sepharose. In

addi-tion,

theLuC 10-reactive Ig of these

infant

sera wasoftheIgG andIgM classes, withnodetectable IgA. This isotypepattern

reflects the isotype distribution oftotalanti-HibPSantibodies

inthese samples asdeterminedby ELISA using Hib PS coated wells,i.e., predominantly IgGwith someIgMand little or no

IgA (datanotshown). The same patternof idiotype and iso-type

reactivity

wasobservedwith the other fourserumsamples.

Thus, Hibld-2 is restrictedto Xanti-HibPSantibodies,and is

expressed by IgGor IgMantibodies.

mAb LuC10 reactivity with sequence-defined human mAbs

indicates that HibId-2 is a marker for a

VX

VII subset

ofanti-Hib

PS

antibodies.

ThisLchain reactivitypattern was also seen

withLuC 10inhibition of Hib PS binding by X chain bearing

human mAbs.InadditiontoLSF-2,theHib PS-specific mAbs

SB5 / D6 (IgA1, X)and RAY 4(IgA2,X) wereinhibited65 and

100%, respectively,byLuC10when tested inthe RABA

inhibi-tionassay. However,the Hib PS binding ofthree other mAbs

also

having

X Lchains, JB21,JB32,andCB6,was notinhibited

byLuc10. Also, in keeping with thelackofLuC10inhibition of serum Kantibodies inELISA (Fig. 3) and the lack of inhibition

of mAb ED8.4 (Fig. 1),LucIO didnot inhibit Hib PS binding

of five other human K chain containingmAbs: ED6. 1, CB 1,

ANN23, ANN2, and CA4,representative of antibodies having KI, KII, or KIII Lchains(reference 26, and authors' unpublished

observations). Thus, consistent with the L chain restriction observedwiththe serumantibodies, these inhibition data with the human mAbsindicatethatHibld-2is aCRI expressed by a subsetofXanti-Hib PS antibodies.

Thenucleotidesequences of the VH and VL regions have beendetermined for five ofthesixanti-Hib PS mAbs having X

chains, includingthethreeexpressing Hibld-2(18, 26). All five

antibodies useVHsegments that are 93-98% homologous to thecandidate germlinegene 9.1(18, and

authors'

unpublished

.001 .01 .1 Hib PS (ug/ml)

A- lI-I

-H

L-0

0

I

-AL&db--Serum 29039 Sepharose-Absorbed

F

0

1

6

ni

-

-100 1000

Serum Dilution

10

Hib PS-Sepharose Absorbed

.1000

Serum Dilution

Serum 29790

Sepharose-Absorbed

2-E

0

(6

Hib PS-SepharoseAbsorbed

--O_ -_~ !=

100 1000

Serum Dilution

100 1000 Serum Dilution

Figure 4. MAb LuClOreactswith IgG and IgM anti-Hib

PS antibodies havingXLchains. Twosera(29039 and

29790)obtained 1moafter the third vaccination with

HibPS-OMPC,wereabsorbed with either Hib

PS-Seph-aroseorunmodified Sepharose. Serial dilutions of the absorbedserawerethen incubated in microtiter wells coated withLuCI0.LuC10-reactive Igwasdetected using enzyme-conjugated antibodies specific for IgG (o), IgM

(i), IgA(n),K(A),orX(.).

observations). One mAb, JB2 1, is encoded by a VXII gene

family member, and this antibodywasnegative for Hibld-2.

The remaining fourantibodies use VXVII segments thatare

96-98% homologous to each other and that may originate fromasingle germlinegene(26).Three of thesefour antibod-ies, LSF2, RAY4, andSB5/D6,werepositive for Hibld-2, and the fourth antibody JB32, was negative. The VXamino acid

sequencesofRAY4 and SB5/D6are100 and 98%homologous

in the framework regions with LSF-2 (theimmunogenusedto

prepare the anti-Id), and they each differed from LSF-2 by

only three amino acid residues in thecomplementarity deter-mining residues (CDRs) (26). For comparison, JB32, the

HibId-2-negative mAb, is 98% homologous to LSF-2 in the framework regions, and differs from LSF-2 in eight CDR amino acid residues (26). Whether these amino acid

differ-encesaccountfor the differences in Hibld-2 expression hasyet tobedetermined. Nonetheless, thedata indicate that HibId-2

serves as aserologic marker forasubset of VXVII anti-Hib PS

antibodies.

Different

Hib PS conjugate vaccines induce

different

idio-typic expression after vaccinationat2, 4,and6moofage.To

examine the effect of vaccineformulationonidiotypic

expres-sion, infants were immunized with either Hib PS-OMPC, Oligo-CRM,orHib PS-Tat2, 4, and 6moofage,andHibId-1 andHibId-2weremeasured inseraobtained moafter the

thirdinjection (Table II).Aserumwasconsideredpositive for idiotype if> 20%of

1251-Hib

PSbindingwasinhibitedby the

respective mAbanti-Id. Ahighly significantdifferencewas ob-served in the frequency of expression of the two idiotypes

among,the recipients of the different vaccines. Hibld- was

detected in 33% ofthe HibPS-OMPCgroup,ascompared with 68 and 64% of infants immunized with either Oligo-CRMor

Hib PS-T, respectively (P=0.001). Thereversepattern was seenwithHibld-2;47% of the HibPS-OMPC-immunized sub-jects expressed Hibld-2,whereas 18 and 10% ofsubjects vacci-nated witheither Oligo-CRM orHib PS-Twere positive for Hib Id-2, respectively (P = 0.001). Thus, significant

differ-ences areobserved in the frequency of Hibld- and HibId-2

expression in infants immunized withdifferent vaccine formu-lations at 2, 4, and 6 mo ofage. However, amongsubjects positiveforHibId- orHibld-2, therespective meanidiotype levelswerenotsignificantlydifferentamongthe different

vac-cinegroups,andonaverage,Hibld- orHibld-2accountedfor

- 50%of the total serumanti-Hib PSantibody.

0

6

in

ao

0

6T

Table II.Effect of Vaccine FormulationonCRIExpressionafter Vaccination ofInfantsat2, 4, and 6 Mo ofAge

Serum cross-reactiveidiotype expression*

Hibld- 1 Hibld-2

Percent inhibition Percentinhibition

Subject by LuC9 byLuC10

group Vaccine No.positivet/total mean±SDO No.positivet/total mean±SDO

1 Hib PS-OMPC 14/43

(33)"

63±22 20/43

(47)'

47±17 2 Oligo-CRM 45/66

(68)"

55±22 12/66(18)' 44±22

3 Hib PS-T 51/80

(64)"

53±21 8/80(10)' 43±12

* Serumwasobtained I moafter the thirdinjection.

Hibld-I

andHibId-2weremeasuredusingRABAinhibitionassay with mAbs LuC9 and LuC10, respectively. $A serum wasconsideredpositive for Hibld-IorHibld-2if>20%oftheHibPSbinding activitywasinhibitedbythe respective mAb. §Of subjectspositive forHibId-1 orHibld-2. 11

_X2

= _15;df=2;P<0.001. I

_X2

_{=22; df= 2;P<0.001.}

Irrespective ofvaccine

formulation,

bothidiotypeswere in-frequently expressed in thesameindividual (Table III). Thus,

the effectof vaccineformulationonHibld-2expression isseen

principally

in subjects who didnot expressHibId-1. Among the

Hibld-1

negative subjects,

Hibld-2

was detected in 66% of thosegiven Hib PS-OMPC vaccine (Group 1) as compared with 29% and 17%of children given Oligo-CRMorHib PS-T

(Groups

2 and

3),

respectively

(P<0.001 ).

Associationsbetween idiotypeexpressionand anti-HibPS

antibody avidity.

Recentlywe reported that the

different

Hib PSconjugatevaccineselicitserumantibodies of different aver-ageavidities ( 11). To determine whether idiotype expression might be correlated with these differences in avidity,we strati-fied the

previously determined antibody avidity

values by

vac-cinegroupand

idiotypic profile (Table

IV). By

measuring

two

idiotypes,

we can

stratify

the vaccinees

into four

categories based upon

HibId-l

/Hibld-2 expression: +/+,

+/-,

-/+,

and -/-.

Since

overall only 5%

of subjects

were

positive for

both

idiotypes

(+ / +), therewere toofew

subjects

inthisgroup for statisticalanalysis.Wethereforerestrictedour comparisons

Table III. RelationBetween

Hibld-J

and HibId-2Expression inInfants ImmunizedwithDifferentHib PSConjugate

VaccineFormulations

No.of Percent

subjects inhibition

Subject No.of positive by LuC10

group Vaccine HibId- * subjects forHibld-2 mean±SEY

1 HibPS-OMPC + 14 1(7) 46 - 29 19(66)§ 47±17

2 Oligo-CRM + 45 6 (13) 36±19 - 21 6(29)§ 51±25 3 Hib PS-T + 51 3 (6) 35±10

- 29 5(17)§ 48±11

Subjectswerevaccinatedat2,4,and6moof age. HibId-1 and HibId-2weremeasuredinseraobtained 1 mo after the third injec-tion. *CriterionforHibld- 1 andHibId-2 positivity is indicatedin

legendtoTableII. tOf subjectspositive forHibld-2. §

_X-

= _15;df

=_{2; P<0.001.}

to the other three idiotypic categories. Among infants given

three injections of Hib PS-OMPC vaccine, those whowere

Hib-Id-1+_{/HibId-2-hadsignificantly highermeanantibody}

avid-ity than those whowereeither HibId- -_{/Hibld-2+ or}

Hibld-1-_/Hibld-2-_{(P < 0.001).In contrast, when the same analysis}

wasperformed for the infants given three injectionsof Oligo-CRM vaccine,there were nosignificantdifferences in the re-spective mean avidities of the groups with different idiotypic profiles (P > 0.27). Further, in all three idiotypic categories analyzed, infants givenOligo-CRMshowedhighermean

avid-ity thanchildren with the respective

idiotypic

categories given

Hib PS-OMPC (P < 0.05). Previously we found that when

analyzed as a group, infants given Oligo-CRM had higher

meanaviditythan those vaccinated with either Hib PS-OMPC orHib PS-T( 11).The presentanalysisindicates that this dif-ference between the mean avidity of antibodies elicited by

Oligo-CRMand Hib PS-Tcanbeaccounted forbythe lower

avidityantibodies in the Hib PS-T-vaccinated children who were

negative

forHibld-l (P<0.001).In contrast,the major-ityof infants respondedtoHib PS-T vaccination by producing Hibld-1 antibodies, and these subjects had nearly identical mean antibody avidity as the corresponding infants given Oligo-CRM (2.55nM-' vs.2.73nM', P=0.6).

Discussion

Molecular characterization of clonally purifiedserumanti-Hib

PSantibodies( 17,20,21 ), andnucleic acidsequenceanalyses

(18, 26) indicate that the antibody response to the Hib PS

antigen

is quite restricted withrespectto Vregiongene utiliza-tion. TheHchains generally use geneshomologous to oneof

two

germline

sequencesfrom theVHIII family(18-21),and

theL

chains

are

characterized

byselective expression ofa lim-itednumberofKand X V region genes ( 19-22). Sequencing of

antibody Vgenesprovides importantinformation on the po-tentialgenetic repertoire ofanimmuneresponse, and on the

possible

relationship

betweenantibody functionand structure. However, this approach is not practical for investigation of factors such asvaccine formulation, that may influence the

expressed immune repertoire of populations. Toaddressthis

question,

wehaveemployed anti-Idreagents ofdefined specific-itytoextend ourpreviousanalysis oftheidiotypeexpression of

Previ-Table IV.Anti-Hib PS Idiotype and Avidity after Immunization with Conjugate Vaccine

HibPS-OMPC Oligo-CRM Hib PS-T

Avidity Avidity Avidity

mean±s.e., mean±s.e., mean±s.e.,

Hibld-1/Hibld-2 No. nM No. nM- No.

nM-+/- 10 1.76±0.16a* 20 2.73±0.31d 22 2.55±0.249

-/+ 16 0.84±0.12 4 1.89±0.69e 1 1.07

-/- 10 0.94±0.16c 9

2.79±O.46'

13 1.26±0.31

+/+ 1 2.23 1 1.71 2 1.52±0.14

Infantswereimmunizedat2,4, and 6 moofage.Avidityandidiotypeexpressionwere measured in seraobtained -1 moafter the third injec-tion.Ofthe 189infantsinthe presentstudy, avidityhad beenpreviouslymeasuredinsera on 109ofthesesubjectsaspartofaseparate study

(11). Becauseof insufficient quantities ofsera, the avidity andidiotypeanalyses were limited to these samples. *Significancetesting by analysis ofvariance: a vs.b,P<0.001;avs.c,P<0.001;g vs. h, P=0.002. a vs.d, P= 0.05;b vs. e, P=0.003; c vs.f,P<0.001;f vs. h, P <0.001.

ously,

we

found

that

antibodies

to

Hib PS frequently

express

Hibld-1,

aCRI which is localizedto the VKIIA2 light chain (22). Indeed,

Hibld-

l isahighly prevalent CRI,andisdetected

in thesera

of

- 80%

of subjects

vaccinatedat1.7to57 yofage

with Hib PSorHib PS coupledtodiphtheria toxoid. HibId-1

comprises

ontheaverage morethan 50%

of

theserumantibody

of children

and adults

(22,

23). However,most

of

thesubjects in these studies had anti-Hib PS antibody in prevaccination

sera,

presumably

as aresult

of

naturalexposure toHiborother

bacteria

that elaborate

polysaccharides

cross-reactive with Hib PS (29). Therefore, the antibody repertoireexpressed after

vac-cination

may

have

been

affected

by

prior

environmental

expo-sures.

Inthepresentstudy,we

examined

the

idiotypic expression

of

the

antibody

response to

Hib

conjugate vaccination of

in-fants

2-6mo

of

agein whom the

effect

of prior environmental

exposures to

cross-reactive

polysaccharides

wouldbe

expected

tobe

minimized

because

of

theiryoung ages. We also

exam-ined

the

idiotypic repertoire

induced in response to several

structurally distinctive Hib conjugate vaccines

thatappear to

elicit

antibody

responses

by different immunologic

mecha-nisms

(30-32). Finally,

we

described

anew

CRI,

HibId-2,

that allowedus toextendour

analysis

to asubset

of

non-HibId-

1 V

regions expressed after vaccination. This

new

CRI,

Hibld-2,

is

defined by

mAbLuC10.

This anti-Id

reacted

specifically

with

anti-Hib PS

antibodies

having

X L

chains. Further,

comparison

between

HibId-2

expression

andVXnucleotidesequence

sug-gested

that

Hibld-2

was

confined

to asubset

of

VXVIIanti-Hib PS

antibodies.

Thetwo

anti-Id

reagents wereusedto

investigate

the

effect

of

conjugate vaccine

formulation on V

region expression

in

postvaccination

sera.Afterthree

injections beginning

at2mo

ofage,Hib PS-OMPC evokedverydifferentpatternsof

idioty-pic

expression

than that observed after three

injections

of

Oligo-CRM

andHib PS-T

given

atthesameages. With both

Oligo-CRM

and Hib

PS-T,

HibId- 1 was more

commonly

ex-pressed,and

Hibld-2

was lesscommon,thanafter vaccination with Hib PS-OMPC vaccineatthisage.

Hibld-2

tended to be more

frequently

expressed after

Oligo-CRM

thanHib PS-Tbut

this difference was not

significant (18

vs.

10%,

P .

_0.15).

These dataindicate thattheanti-Hib PSVLregions elicited by

Oligo-CRM

orHib PS-T

vaccination,

appearsimilartoeach other, but different from those elicited

by

HibPS-OMPC

vacci-nation.

Of

thethree

Hib conjugate

vaccinesusedinourstudies, Hib

PS-OMPC is unique inthatit isable to elicit a primary

anti-bodyresponsein 2-mo-old infants (2,5-7), whereasbeginning

at

this

age,the othertwoconjugates requiretwo or three injec-tionstoelicit

antibody

(6-8). This difference in immunogenic-ity between thethreeconjugate vaccines isshown by the

signifi-cantly

higher antibody

levels before the thirddose, inthe

in-fants

who hadreceived the OMPC conjugatethan those who

received the othertwoconjugates(Table

I).

Recently we ob-served that

Hibld-2 expression

followingasingle injection of PRP-OMPC tendstodiminish withage (33). Thus, the higher usage

of

Hibld-2

after

three

injections of

the Hib PS-OMPC

conjugate

observed in the present study, may relate to the

unique ability ofthis conjugate

toinduce

Hibld-2

antibodies in 2-mo-old

infants

and thereby

fix

theidiotypic repertoireatthis age. In contrast,the

antibody

response tothe othertwo

conju-gatesoccurredat 4-6mo

of

age, by which time the potential

idiotypic repertoire

mayhaveshifted.

The

mechanism(s)

by which the Hib PS-OMPC vaccine is ableto

induce

a

primary antibody

response in 2-mo-old

in-fants,

aswellasinduce such adistinctive idiotypic profile, is unknown. This

conjugate

uses an outer membrane complex from Neisseria

meningiditis,

and this

carrier

protein may

uniquely

modulate the

infant's immune

response. Recently,

this

conjugate vaccine

was

found

tobe

mitogenic for murine

B cells

(30-32) and,

thus, in addition to

having thymic

depen-dent

properties,

theHib PS-OMPC

conjugate

mayhave

proper-ties of

a

thymic-independent

type 1

antigen. Conceivably,

the

differences

inV

region expression

evoked

by

the

different Hib

conjugate vaccines

may

reflect differences

in the relative

con-tributions

ofthymic-dependent

and

thymic-independent

mech-anisms

operative

in the

antibody

responsesto

Oligo-CRM

or

Hib PS-Tcomparedtothose

operating

with the Hib PS-OMPC

conjugate.

Evidence inamurine

antihapten antibody

response suggeststhat

different

Bcell

subpopulations

arestimulated

by

TIand TDforms ofhapten

(34). Therefore,

different HibPS

conjugate

vaccinesmaystimulatedistinctiveBcell

subsets,

and these subsetscould differ in their

potential

toexpressHibld-1

and/or

Hibld-2.

The

conjugate

vaccines may also differ in

their

capacity

topromotesomatic mutation and this property could affectV

repertoire

content. For

example,

the lower

fre-quency of

expression

of

Hibld-2

antibodies after vaccination with eitherthe tetanusor

Oligo-CRM conjugates

as

compared

that extinguish the idiotope. In this case, lack ofHibld-2

ex-pression would not reflect lack of utilization but rather somatic

modification of the VXVII region. Consistent with this inter-pretation are the data with the anti-Hib PSmAbs, showingthat Hibld-2 is expressed by a subset (three offour)ofantibodies

usingVXVII.

Recently we compared the avidities of serum anti-Hib PS

antibodies evoked by vaccination with Oligo-CRM, Hib

PS-OMPC,orHibPS-Tconjugate at2,4, and 6 mo of age

( 1).

Oligo-CRMstimulated antibodies ofhighermeanaviditythan those elicited by Hib PS-OMPC vaccination,while themean avidity of antibodies evoked by Hib PS-T was intermediate, being

significantly

higherthan that ofthe Hib PS-OMPC group but lowerthan that of the groupgiven Oligo-CRM. Further,

there was an inverse correlation between the serum

antibody

concentrationrequiredtoactivatecomplement-mediated bac-teriolysis of Hib, and theavidityof theanti-Hib PSantibody,a finding that suggeststhatdifferencesinantibody aviditymay play an important role in

conferring

protection against Hib disease. Thesedifferencesinantibody aviditywerelikelya re-sult of Vregiondifferences andnotisotypic

differences,

since on average > 90% of the anti-Hib PS antibody responses to each of theconjugates wasIgGand there was no correlation between themagnitudeof theIgMresponse andavidity (11).

Inthe present study,weexamined the

relationship

between average serum anti-Hib PS

antibody avidity

andCRI expres-sion. Thedata indicate that both

immunogen

(vaccine formu-lation) and CRIcontentcorrelate with

antibody avidity.

Sub-jects who wereimmunizedwitheither HibPS-OMPCor Hib

PS-Tvaccines and who

produced

detectable Hibld-1-positive

antibody, hadsignificantly highermeanaviditythan subjects whodidnotproduceHibId-

1-positive antibodies.

Althoughwe

havenotmeasured

avidity

ofclonally purified anti-Hib PS

anti-bodies, this correlation suggests that antibodies

expressing

Hib-Id-l,i.e., having the unmutated VKII-A2region (21, 22, 26), havehigher

avidity

for Hib PS than the non-Hibld-1

popula-tion induced by Hib PS-T or Hib PS-OMPC. Thisputative

higher avidity of VKIIA2 antibodiesmayconferaselective ad-vantagein protection against Hib disease and thereby maintain

the

nonpolymorphic

A2 gene inthepopulation.

However, thepatternof average antibody avidity andCRI

expression

seen after

Oligo-CRM vaccination

suggests that

non-Hibld-1

antibodies

are not

necessarily

of lower avidity.

After three injections of Oligo-CRM, themeanantibody

avid-ity insera notcontaining HibId-1/VKIIA2wassimilarto that

ofsera

having

Hibld-1

antibodies.

In addition, vaccination withOligo-CRM appearedtoinduce uniformlyhigher avidity

anti-Hib PS

antibodies

thaneither Hib PS-OMPCorHibPS-T,

irrespective of the idiotypic profile.

The induction of higher avidity antibodies by the Oligo-CRM vaccine mayresult from the presence of HibPS oligo-mersin the Oligo-CRM conjugate ratherthan the higher molec-ularweight form ofHibPS that is present in thetetanusand

OMPC conjugates. Theoligomeric form ofHib PS may

selec-tively

activate B cell clones having highaffinity receptors. In

addition,

theOligo-CRM conjugate may elicit greater T cell

help comparedwith the other twoconjugates,which may favor greateraffinity maturation of Hibld-I-negative antibody

popu-lationsbypromotingmoreextensive somatic mutation during

thesecondary immuneresponse. TheHibPS conjugate

vac-cines differ with respect tothe

type

of carrier

_molecule,

the

natureofthechemical

_linkage

between the

_{polysaccharide}

and

the

protein,

andmolecularsize. These factorscould affect the

immunological properties

of the

vaccines,

suchas

clearance,

processing,

and retention in reticuloendothelial tissues and, thus,

possibly

contribute to both theobserved

avidity

differ-encesand the distinct

idiotype

profiles induced

by

thedifferent

conjugate

vaccines.

In

previous studies,

differencesinmeanavidities of two-or

threefoldwereassociated with

greater

thansixfold differences

in the

ability

ofanti-Hib PS antibodies to activate

comple-ment-mediated bactericidal

activity (

11

)

andopsonic

activity

(35). Thus,

vaccineformulationsthat evokehigher avidity

an-tibody

may offer greater protection against invasiveHib

dis-ease. Therearecleardifferencesin V

region

usage and in the meanaviditiesof theanti-HibPS antibodies inseraofinfants vaccinated with different conjugate vaccines. These factors

may be

important

in the

design

offuturevaccines aimed at

optimizing protective immunity.

Acknowledgments

Dr. Y.Schlesinger performedsome of the measurements ofantibody avidity.Venita Boelloeni, Ann Grace, Anthony Quinn, and Marina Kitamuraprovided experttechnical assistance. Drs. Elisabeth Adder-son and William Carrollperformedthe sequenceanalysisofthe

hybrid-omas.

Thiswork wassupported bygrantsAl- 17962, AI-2 1842, Al- 19350, andAl 25008 from the National Institute ofAllergyand Infectious Diseases, National Institutes of Health.

References

1.Insel, R. A., and P. W. Anderson. 1986.Oligosaccharide-proteinconjugate

vaccines induce andprimeforoligoclonal IgG antibodyresponsestothe Hae-mophilus influenzaeb capsular polysaccharide in human infants. J. Exp. Med. 163:262-269.

2.Einhorn, M. S., G. A. Weinberg, E. L. Anderson, P. D. Granoff, and D. M. Granoff. 1986. Immunogenicityin infantsofHaemophilus influenzaetype b polysaccharide in a conjugate vaccine with Neisseriameningitidis outer-mem-braneprotein.Lancet. ii:299-302.

3.Weinberg,G.A., and D. M. Granoff. 1988.Polysaccharide-protein

conju-gatevaccines for the prevention ofHaemophilusinfluenzae type b disease.J. Pediatr. 113:621-631.

4.Marburg,S., D. Jorn, R. L.Tolman,B.Arison,J.McCauley,P. J. Knis-kern, A. Hagopian, and P. P. Vella. 1986. Biomolecular chemistry of macromole-cules:synthesisofbacterialpolysaccharide conjugatewithNeisseriameningitidis

membraneprotein.J. Am. Chem. Soc. 108:5282-5287.

5.Lenoir, A. A., P. D. Granoff, and D. M. Granoff. 1987. Immunogenicity of Haemophilus influenzaetype b polysaccharide-Neisseria meningitidis outer

membrane protein conjugate vaccine in 2- to 6-month old infants. Pediatrics. 80:283-287.

6.Decker, M. D., K. M. Edwards, R. Bradley, and P. Palmer. 1992.

Compara-tive trial in infants of fourconjugate Haemophilusinfluenzaetype bvaccines.J. Pediatr. 120:184-189.

7. Granoff,D. M., E. L.Anderson,M. T. Osterholm, S. J. Holmes, J. E. McHugh, R. B. Belshe, F. Medley, and T. V. Murphy. 1992. Differences in the immunogenicityofthreeHaemophilus influenzaetype bconjugate vaccines in infants. J. Pediatr.121:187-194.

7a.Holmes, S. J., B. Fritzell, K. P. Guito, J. Esbenshade, M. M.Blatter,K. S. Reisinger, H. Keyserling,E. P. Rothstein, H. H. Bernstein, S. Feldman, and D. M. Granoff. 1993. TheimmunogenicityofHaemophilusinfluenzaetype b polysaccharide-tetanustoxoidconjugatevaccine in infants. Am. J. Dis. Child. In Press.

8. Madore,D. V., C. L.Johnson,D.C. Phipps, Pennridge Pediatric Asso-ciates, L. A. Popjoy, R. Eby, and D. H. Smith. 1990. Safety and immunologic response toHaemophilus influenzae boligosaccharide-CRM197conjugate vac-cinein 1- to 6-month-old infants. Pediatrics. 85:331-337.

9.Vella,P.A.,and R. W. Ellis. 1991. Immunogenicity of Haemophilus

in-fluenzaetype bconjugatevaccines in infant rhesus monkeys. Pediatr. Res. 29:10-13.

10.Granoff,D.M.,M.H.Rathore,S.J.Holmes,P.D.Granoff,andA.H. Lucas. 1993. Effect ofimmunitytothe carrierproteinonantibodyresponsesto

11.Schlesinger,Y., D. M. Granoff,and theVaccineStudyGroup. 1992. Avidityandbactericidalactivityofanti-capsular antibodies elicited bydifferent Haemophilus influenzae type b conjugate vaccines. J. Am. Med. Assoc. 267:1489-1494.

12.Shackelford,P.G.,D. M.Granoff,S. J.Nelson,M.G.Scott,D.S.Smith, and M. H. Nahm. 1987. Subclass distribution ofhumanantibodies to Haemophi-lusinfluenzae b polysaccharide. J.Immunol. 138:587-592.

13. Makeld, O., P. Mattila,N. Rautonene, I. Sepptla,J. Eskola,and H. KAyhty. 1987. Isotype concentrations of human antibodies toHaemophilus in-fluenzae type bpolysaccharidein young adults immunized with the polysaccha-ride as such orconjugated to a protein (Diphtheria toxoid). J. Immunol.

139:1999-2004.

14. Insel, R. A., P.Anderson,M. E.Pichichero,S.Schuster,M. D. Amstey,G. Ekborg, and D. H.Smith. 1982.Anticapsular antibodytoHaemophilus

influen-zaetype b. InHaemophilus influenzae, Epidemiology,Immunology,and Preven-tion of Disease. S. H. Sell and P. E.Wright,editors. Elsevier SciencePublishing Co., Inc.,NewYork.155-168.

15.Ambrosino,D.M., V.Greif, C. Thompson,and G. R.Siber. 1990.Kand X lightchaincompositionofantibodytothecapsularpolysaccharide ofHaemophi-lusinfluenzaetype b. J.Infect.Dis. 161:922-925.

16. Insel, R. A., A.Kittelberger,and P.Anderson.1985.Isoelectric focusingof humanantibodytotheHaemophilusinfluenzaebcapsularpolysaccharide: re-stricted andidenticalspectrotypes inadults.J.Immunol. 135:2810-2816.

17.Scott,M.G.,J.J.Tarrand,D. L.Crimmins,D. W.McCourt,N. R.Siegel,

C. E.Smith,and M. H. Nahm. 1989. Clonalcharacterization ofthe humanIgG

antibody repertoiretoHaemophilus influenzaetype bpolysaccharide. II.IgG antibodies containVH genesfromasingle familyand VL genesfromatleastfour VLfamilies.J.Immunol. 143:293-298.

18.Adderson,E. A., P.G.Shackelford,A.Quinn,andW. L.Carroll. 1991. RestrictedIg HchainVgene usage in the humanantibodyresponseto

Haemoph-ilusinfluenzaetype bcapsularpolysaccharide.J.Immunol. 147:1667-1674. 19.Silverman,G.J.,and A. H. Lucas. 1991.Variableregiondiversityin humancirculatingantibodiesspecificforthecapsular polysaccharideof Hae-mophilus influenzaetype b:preferentialusage oftwotypes of VH3heavychains. J. Clin.Invest.88:911-920.

20.Scott,M.G.,D.L.Crimmins,D. W.McCourt,G.Chung,K.F.Schable,

R.Thiebe,E.-M.Quenzel,H.G.Zachau,and M. H. Nahm. 1991.Clonal charac-terization ofthe humanIgGantibodyrepertoiretoHaemophilus influenzaetype bpolysaccharideIV. The lessfrequently expressedVLareheterogeneous. J. Immunol. 147:4007-4013.

21. Scott,M.G.,J.J.Tarrand,D.L.Crimmins,D. W.McCourt,I.Zocher,R. Thiebe,H.G. Zachau,and M. H. Nahm.1989. Clonal characterization ofthe humanantibody repertoiretoHaemophilusinfluenzaebpolysaccharide.III.A single VkIIgene andoneofseveral Jk genesarejoined byaninvariantarginineto

formthe mostcommonLchainVregion.J. Immunol. 143:4110-4116. 22.Lucas, A.H.,R.J.Langley,D. M.Granoff,M. H.Nahm,M. Y.

Kita-mura, and M.G. Scott. 1991. An idiotypic marker associated with a germ-line encoded K lightchainvariable regionthat predominates the vaccine-induced human antibody response to theHaemophilusinfluenzae b polysaccharide. J. Clin. Invest.88:1811-1818.

23.Lucas, A. H., and D. M. Granoff. 1990. A major crossreactive idiotype associated with human antibodies to the Haemophilusinfluenzaeb polysaccha-ride: expressioninrelationtoage andimmunoglobulinG subclass. J. Clin. In-vest.85:1158-1166.

24.Granoff,D.M., P.G. Shackelford, J. P. Pandey, and E. G. Boies. 1986. Antibodyresponses toHaemophilusinfluenzaetype bpolysaccharide vaccine in relation to Km(l) and G2m(23) immunoglobulin allotypes. J. Infect. Dis. 154:257-264.

25. Griswold, W. R., A. H. Lucas, J. F. Bastian, and G. Garcia. 1989. Func-tionalaffinityofantibodytothe Haemophilusinfluenzaebpolysaccharide. J. Infect.Dis. 159:1083-1087.

26. Adderson, E. A., P. G.Shackelford,R. A.Insel, A. Quinn, P. M. Wilson, and W. L. Carroll. 1992.Immunoglobulinlightchain variable region gene se-quences for human antibodies toHaemophilusinfluenzaetypeb capsular polysac-charide are dominated by a limited number ofVKand VX segments and VJ combinations. J. Clin. Invest.89:729-738.

27.Fazekas de St. Groth, S., and D.Scheidegger. 1980.Production of mono-clonal antibodies: strategies and tactics. J.Immunol. Methods.47:129-144.

28.Kearney, J. F., A. Radbruch, B. Liesengang, and K. Rajewsky. 1979. A newmouse myeloma cell line that has lost Ig expression but permits construction ofantibody-secreting hybridcelllines.J.Immunol. 123:1548-1550.

29.Robbins,J. B., R. Schneerson, M. P. Glode, W. Vann, M. S. Shiffer, T.-Y. Liu, J. C. Parke, and C. Huntley. 1975.Cross-reactiveantigens and immunity to diseasescausedbyencapsulatedbacteria. J.Allergy Clin. Immunol.56:141-15 1. 30.Donnelly, J. J., R. R. Deck, and M. A. Liu. 1990.Immunogenicityof Haemophilusinfluenzaeb polysaccharide-Neisseriameningitidisouter mem-brane proteincomplexconjugatevaccine. J.Immunol. 145:3071-3079.

31.Stein,K.1992.Thymus-independentandthymus-dependentresponses to polysaccharide antigens. J.Infect.Dis. 165:S49-S52.

32.Liu, M. A., A.Friedman,A. I.Oliff,J.Thai, D. Martinez, R. R. Deck, J. C. T.Shieh,T. D.Jenkins,J.J.Donnelly,and L. A. Hawe. 1992. Avaccine carrier derived from Neisseriameningitidis with mitogenic activity for lympho-cytes. Proc. Natl. Acad. Sci. USA. 89:4633-4637.

33.Lucas,A.H.,F. H.Azmi, C.M.Mink,and D. M.Granoff.1993. Age-de-pendent variableregion expression inthe humanantibodyresponse to the Hae-mophilusinfluenzaetype bpolysaccharide.J.Immunol.InPress.

34.Tittle, T.V.,and M. B.Rittenberg.1978.DistinctsubpopulationsofIgG memory B cellsrespondtodifferentmolecularforms ofthe same hapten. J. Immunol. 121:936-941.