Journal of immunological Methods, 148(1992) 101-114

101

© 1992ElsevierSciencePublishersB.V. All rights reserved 0022-1759/92/$05.00

JIM 06231

Quantitation of human IgG subclass antibodies to Haemophilus influenzae type b capsular polysaccharide Results of an international collaborative study using enzyme immunoassay methodology D.J. H e r r m a n n a, R.G. H a m i l t o n b, T. B a r i n g t o n c, C.E. Frasch d, G. A r a k e r e d, O. M~ikel~i c, L.A. Mitchell f, J. Nagel g, G.T. Rijkers h, B. Zegers h, B. D a n v e i, J.I. W a r d J a n d C.S. Brown a Connaught Laboratories, Inc., Swiftwater, PA, USA, b Johns Hopkins Unit'ersity, Baltimore, MD, USA, CRigshospitalet, Copenhagen, Denmark, a Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda. MD, USA, e University of Helsinki, Helsinki, Finland, f Vaccine El'aluation Center. B.C., Canada, R Rijksinstituut roor Volksgezondheid en Milieuhygier, e, Bilthot'~.n. Netherlands, h Wilhelmina Kinderziekenhuis, Utrecht, Netherlands, i Institut M#rieux, Lyons, France, j UCLA Medical Center, Torrance, CA, USA

(Received 19 August 1991,revisedreceived 12 November1991.accepted 13 November1991)

An international collaborative study was conducted at ten sites to examine the performance of enzyme immunoassays (EIAs) for the quantitation of lgG1, lgG2, IgG3, igG4 and total lgG anti-Haemophilus influenzae type b (Hib) capsular polysaccharide in human serum. All groups used the same reagents: microtiter plates coated with polyribosylribitol phosphate (PRP) conjugated to poly-L-lysine (PLL), reference, control and test human sera, biotin-conjugated International Union of Immunological Societies (IUIS)-documented monoclonal anti-human lgG1-4 and lgG Pan detection antibodies, avidin-peroxidase and TMB substrate. Initial mixing of soluble PRP antigen or an equal volume of buffer with the 20 test sera prior to analysis confirmed PRP antigen specificity in all five EIAs with > 80% competitive inhibition at most sites. Positive correlation between the total lgG anti-Hib and sum of I g G l - 4 anti-Hib was demonstrated (r 2 = 0.99, Y ffi 1.13X-0.15). Good agreement was shown between the total IgG anti-Hib as measured by EIA and the total Hib-specific antibodies measured by the current radiolabeled antigen binding assay (r 2 = 0.97, Y = 4 . 6 X - 5.8). Assay parallelism was demonstrated with an average interdilutional %CV of 22% and parallel dose-response curve slopes. The interdilutional %CVs were calculated as an average per sample of the variation of # g / m l (corrected for dilution) at different dilutions per laboratory for all participating sites. The interlaboratory variation was the only performance parameter studied that exceeded the target level of 35% CV in all lgG1-4 and total lgG anti-Hib assays. IgG subclass distributions in the test sera demonstrated a predominance of lgG1 anti-Hib in the

Correspondence to: D. Herrmann, Research Department, ConnaughtLaboratories, Inc., P.O. Box 187, Swiftwater, PA 18370, USA (Tel.: 717-839-4369;Fax: 717-839-0619). Abbreviations: Hib, Haemophilus influenzae type b; PRP, polyribosylribitolphosphate; EIA, enzyme immunoas.~ay;RIA, radioimmunoassay;MAb,monoclonalantibody;HSP3, humanserum pool #3; HSP4, humanserum pool #4; IgG, immunoglobulin G; CV, coefficientof variation;PLL, poly-L-lysine;HSA, humanserum albumin;ID, inter-dilutional;NP, 4-hydroxy-3-nitro-phenylaeetyl.

102 pediatric serum pools and IgG2 anti-Hib in the adult sera, with low but detectable levels of Ig(33 and IgG4 anti-Hib in each group.

Keywords: Haemophilusinfluenzae type b; EL1SA;Polyribosylribitolphosphate;Poly-L-lysine;Monoclonalantibody

Introduction

Haemophilus influenzae type b (Hib) is a gram negative rod that is responsible for recurrent upper respiratory tract infections and the majorit5, of bacterial meningitis-associated morbidity and morthlity in children (Fraser et al., 1974; CDC, 1979). Immunologicaily mature humans produce serum antibodies to the Hib capsular polysaccharide, polyribosylribitolphosphate (PRP), upon vaccination that have been associated with protection (Fraser et al., 1974; Robbins, 1978; Granoff and Munson, 1986). The humoral immune response to PRP, however, is age dependent with poor antigenicity in small children that may be attributed to the late appearance of thymus-independent B cells (Seppala et al., 1988). In recent years, protein-conjugated Hib capsular polysaccharide vaccines have become available which are immunogen[c in infants younger than 1 year of age (Robbins, 1978; Peltola et ai., 1984; Granoff and Munson, 1986). Serological evaluation of Hib specific humoral immune responses in children, however, requires a maximally sensitive, reproducible clinical laboratory procedure. Until recently, a standardized radioimmunoassay (RIA) has been routinely employed to measure total Hib specific antibody levels in individuals. Technical difficulties with the R1A and the concern over the sensitivity and variability of functional ba~,*ericidal assays have created interest in a non-isotopic immunoassay for the measurement of antibody responses to Hib ~:apsular polysaccharides. Moreover, well established functional differences between the subclasses of IgG antibodies have renewed questions about the subclass distribution of IgG antibodies to capsular polysaccharide antigens in adults and children. Of the studies reporting the subclass distribution of human l e g antibodies to PRP (P¢ltola et al., 1984; Ramadas et al., 1986; Makela et al., 1987; Shackelford et al., 1987; Barra et ai.,

1988; Claesson et al., 1988; t3ranoff et al., 1988; Musher et al., 1988), few have used the same reagents and methods and thus it has been difficult to compare reported IgG antibody levels and subclass patterns from these studies. In this study, we have performed an international collaboration to document the performance of an enzyme immunoassay that uses conjugated PRP on the solid phase and biotinylated IUIS-documentod monocional antibodies (MAbs) specific for the human IgG subclasses as detection antibodies. All 14 participating sites utilized a standard assay protocol, materials (microtiter plates) and reagents (coating antigen, soluble antigen for inhibition, reference, control and test sera, biotin-MAbs, avidin horseradish peroxidase, and TMB substrate) that had been optimized primarily for sensitivity. The goal of this study was to determine the level of interlaboratory performance (sensitivity, specificity, reproducibility, parallelism) that can he achieved using a standardized assay system.

Materials and methods

Reagents Hib polyribosylribitol phosphate (PRP, Connaught Laboratories, Swiftwater, PA (CLI); MW 250,000-1000,000) was conjugated to poly-L-lysine (PLL, Sigma #P2636, MW 55,090) by a modification of the method of Gray (1979). Uncoupled Hib-PRP was supplied at 34 mg/ml to all pa;ticipants for use in soluble antigen inhibition studies to test assay specificity. IUIS-documented murine monoclonal antibodies (Jefferis et ai., 1985) specific for human leG1 (HP6069), IgG2 (HP6002), IgG3 (HP6047), leG4 (HP6023) and IgG Pan Fc (HP6017) were obtained in a biotinylated form at 1 mg/ml from the Hybridoma Reagent Laboratory (Baltimore, MD). All chemicals unless otherwise specified were obtained from Sigma Chemi-

103 cal Company, St. Louis, MO. All buffers were prepared using endotoxin-free deionized water.

Human serum specimens Participants in the study each provided two 10 ml aliquots of individual or pooled sera from humans immunized with Hib vaccines. Each 10 ml sample was dialyzed against PBS overnight at 4°C to remove preservatives (e.g., azide), brought to a 15 ml final volume, assigned random coded numbers and frozen at -70"C until shipped to the collaborating laboratories. The reference (HSP4) and control (HSP3) serum pools were prepared from adult volunteers who had been previously immunized with PRP-D and PRP vaccine (CLI), respectively. The content of IgG1, IgG2, lgG3 and IgG4 anti-Hib in the HSP4 was determined by the Baltimore group using cross-calibration (Butler and Hamilton, 1991) with a previously reported method that utilizes human lgG1-4 anti-NP chimeric antibodies to generate heterologous dose response curves (Hamilton, 1991). A second calibration of the lgG1-4 anti-Hib content in the HSP4 was performed independently by the Finnish group, using a radioimmunoassay and a human serum standard that had been prepared by soluble HibPRP antigen immunoprecipitation and quantitation of the igG1-4 protein in total IgO subclass assays (Seppala et al., 1988). The FDA 1983 human anti-H, influenzae type b capsular polysaccharide serum was provided by Dr. C. Frasch (CBER, FDA, Bethesda).

lgG l - 4 anti-Hib enzyme immunoassay The strategy in the design of this assay was to optimize performance and not to miaimize the assay turnaround time. lmmulon II microtiter plates (Dynatech, Chantilly, VA) were coated overnight at 4°C with PRP-PLL (0.1 ml/well) that had been diluted in endotoxin-free PBS to a final concentration of 5 p.g/ml PRP and 1.25 t~g/ml PLL. Concurrently, reference, control and test sera were diluted in PBS containing 0.05% (v/v) Tween 20 and 2% (w/v) bovine serum albumin (PBS-T-BSA) and incubated overnight at 40C with equal volumes of buffer (control) or unconjugated PRP (0.34 mg/ml final concentration). Following adjustment to room temperature,

the plates were washed three times with PBS-T. Reference, positive control and test sera at eight, three and four two-fold dilutions, respectively, were pipetted into duplicate wells (0.1 ml/well) and incubated overnight at room temperature. The positive control serum and a PBS-T-BSA buffer blank were analyzed on every plate. Plates were washed four times with PBS-T and biotinylated monoclonal anti-human igGI-4 or lgGFcPan were added at 1 /~g/ml in PBS-T-BSA. Following an overnight incubation at room temperature, avidin-horseradish peroxidase (Extravidin-peroxidase, Sigma #A7030) was added at 0.1 ml/well (! /~g/ml in PBS-T-BSA) for 3 b at room temperature. After a final wash, equal volumes of 3,3',5,5'-tetramethylbenzidine and HzO 2 (TMB Peroxidase Substrate; Kirkegaard & Perry Labs, Inc., Gaithersburg, MD) were mixed and added to all wells at 2 rain intervals between plates. The optical density was read on a variety of microtiter plate readers at 30 rain at 2 rain intervals using a wavelength of either 630 nm (reaction not stopped) or 450 nm (stopped with 2 N HzSO4). Variations on this general assay format included the use of 50 versus 100 t~l sample volumes, and slight variations in serum and substrate incubation times.

Radioantigen binding assay All sera were analyzed in the radioantigen binding assay by a modification of the method of Anderson (1978). Intrinsically labeled Hib polysaccharide was used at 10 ng/ml in 0.1 M phosphate buffered saline, pH 7.4. Antibody values were determined by comparison to a calibrated CBER reference (lot 1983) which had been assigned a value of 70/zg/ml of Hib-specific antibody.

Computations igGI-4 and total igG anti-Hib levels in uninhibited and soluble antigen inhibited samples were interpolated from the HSP4 standard curve using either a cubic spline (Johns Hopkins University, Baltimore) or four-parameter logistic data processing program. The percent inhibition was calculated using the following formula: [(p,g/ml uninhibited - tzg/ml inhibited) + t~g/ml uninhibited] x 100. While 80% inhibition was the tar-

104 TABLE 1 Hib-SPECIFIC ENZYME IMMUNOASSAYPERFORMANCE Parameter

Serum dilution

lgGi Ref a Test b (ng/ml) 10 20 40 80 160 320 640 1,280 2,560 500 !,000 2,000 4,000 8,000 16,000 32,OO0 64,000 128,000 256,000

Detection antibody Working range (ng/ml) Sensitivity(ng/ml)

lgG2 Ref a Test b (ng/ml)

lgG3 Ref a Test b (ng/ml) PA

P

18.6 9.3 4.7 2.3 1.2 0.6 O.3

PA

HP6069 0.6

3.9 2 l 0.5 0.25

Total lgG Ref a Test b (ng/ml)

PA

P A 32.2 16.1 8 4 2 I 0.5 0.3

HP6002 9.0

0.3

86.6 43.3 21.7 10.8 5.4 2.7 1.4

38.4 19.2 9.6 4.8 2.4 1.2 0.6 0.3

lgG4 Ref a Test b (ng/ml)

2

HP6047 19

1.4

1

HP6023 20

0.3

0.5

A

HP6017 3

0.5

0.25

7 0.3

Ref = reference serum refers to HSP4. Hib specifictotal lgG and igGl-4 are presented in ng/ml for dilutions analyzed. b Test = test sera refers to pediatric (P) or adult (A) sera. Recommendedmean dilutions are denoted as P or A.

get value for acceptable soluble antigen inhibition, 50% inhibition was considered evidence for the presence of specific antibody. Mean dose estimates based on each replicate determination were analyzed at each serum dilution and used to compute a grand mean and an interdilutional (ID) coefficient of variation (CV = S D / m e a n ) to evaluate parallelism. ID%CVs < 20% were taken as evidence of assay parallelism. The grand mean dose estimate for each serum generated by each laboratory was used to calculate the interlaboratory variation (data point numbers -- 9, 10, 10, 10, and 9 for the IgG, IgG1, IgG2, lgG3 and IgG4 anti-Hib calculations, respectively). Interlaboratory %CVs < 35% were acceptable, considering intra-assay variation displayed typically 5 - 1 5 % CV (Butler and Hamilton, 1991). Interlaboratory and interdilutional CVs were p l o t t e d in separate precision profiles

(Ekins, 1981) as a function of their associated mean dose to study trends in variation over the working range of each lgG subclass assay.

Results

IgG1-4 anti-Hib EIA performance Quality of the current immunoassay was assessed in terms of its specificity, sensitivity, reproducibility and parallelism. Specificity. Both Hib-PRP antigen and human IgG subclass specificity of the E I A were investigated. Hib antigen specificity of each EIA was studied by each participating laboratory using soluble antigen inhibition. Pre-incubation of soluble P R P with the reference, control and test sera prior to immunoassay analysis demonstrated 80100% inhibition for the majority of sera in the

ll)5

t o t a l ( F i g . l A ) a n d I g G 1 - 4 ( Fig. 1 B ) E I A s . W h i l e select sera displayed an inability to be inhibited > 8 0 % w i t h s o l u b l e a n t i g e n , 9 5 % w e r e inh i b i t a b l e b y > 5 0 % , a v a l u e w h i c h is a b e n c h m a r k t a r g e t for c l i n i c a l a n t i b o d y assa' rs. T h e SeC-

ond specificity issue involved the lgG subclass restricted nature of each immunoa~say. This was achieved through the use of extensively documented murine monoclonal antib~lies (Jefferis et al., 1985). E a c h b i o t i n y l a t e d M A b w a s f u r t h e r

A 100.

•

i

|. m

_.

,r

•

•

•

,

8. 6o.

1

lo loG

[3

I1O.

11o j

loG1 eee

i: i

ale

•

10G2

maim

'!t

•

•

lOO

lpg/mll

° o

. "•. " - "".

""

-:

~o-I •

w-

IO-i

• ........

.ol

5o-I , ........ , .I 1 lOG1 l p g l m l )

....... lo

401 .I loG2 IgG4

IoG3

~

•

1oo-

(p0/ml) ms

• glee

w4 gill

|

m!

__ao ?o4

w4 so4

20.Ol

.i IgG3

(p01ml)

4oi

.o01

01 IgG4

lpg/ntt)

Fig. I. Specificity assessment of the total lgG (A) and IgG 1-4 (B) anti-Hib enzyme immunoassays. The percent inhibition of the binding of Hib antibody to solid phase PRP-PLL with soluble PRP is plotted as a function of the log content of IgGI-4 or total IgG anti-Hib in 21 test and control sera. 80% inhibition (solid line) was the target used in this study as a criterion for acceptable Hib specificity.

106 tested under conditions of these E I A by one group (Johns Hopkins University, Baltimore, MD) using a well defined panel of h u m a n lgG1, IgG2, IgG3 and IgG4 myeloma proteins and h u m a n I g G I - 4 chimeric antibodies and established methods (Hamilton, 1990a, 1991). Sensitivity and working ranges. The minimal detectable antibody level (sensitivity) for each assay was determined by dilutional analysis, corn-

A

':I

paring response levels obtained with the calibrated reference serum to those produced with buffer. The lowest m e a n levels of IgG anti-Hib considered detectable above the buffer blank were 0.3, 1.4, 0.3, 0.25 and 0.3 n g / m l in the IgG Pan, IgG1, IgG2, IgG3 and IgG4 EIAs, respectively (Table D. One aspect of IgG subclass assays is the wide range of serum dilutions required and the work-

o

solo

==-.o I I

o

~ot sot

o o o

I /

;~t,~ ........ O,,o .............................................. o,I %:5. . . .

B

100

0

.~-

9oi o

z _o

801

0

O0

0

40 ~

20

•

_

eoo

_e

~

0 .... :

lOO 0

9oi e0. o 70( 60

0

40~.. 30.t~;""e""

•

~ ' ...... (pg/ml) IDICV= • Inter-Lob ~

1'0 1

• 0 ........... i l ................. S . . . . .

8O 7O

80k s0~

0 o 0 0 0 ................................................ 0 .....

....

°o.o,

-co i~ 0

. . . . . . . 1903 o n t i - H I B

OT , , ~ , , I 0.4 1

~.~o (pg/ml)

'

~

,

, : :::::

: : 10 (pg/ml) IOIICV= • Inter-I,~b ~

IgO2 o n t i - H I B

U

= 0

30

10

0 a"'i~

lOO~

0

o o

50 40

u

0 0

• 1 IgG1 o n t i - H I B

o.1

20

0

.---

lO

~o

• C~

0

50

.....

--.-~--ii¢~"

1ozcv= • Inlet-Lab ~

0

60 o ~-,

O0

0

7O

0 o

so

~z

100t

90 1 80

o

30

IDSCV= • Inter-Lob IKCV- 0

0

70. o 60, ~-

.... ,'dot

....... ~b . . . . . . . UgGonti-HIB OJg/ml)

i

0 0 %

_ o e~

40L

o° 0

0 0 0.3 o,oo2

•

55

=C

o

3o~" : r ' GZ • ................ ~,,,a. . . . . .•.

10~-

:

o

0

•o

•

• o.o~o IgO4 o n t i - H I B

o.os (pg/ml)

Fig. 2. Precision profile (Hamilton, 1991) for the total IgG (A) and lgGi, lgG2, IgG3 and lgG4 (B) anti-Hib enzyme immunoassays.The interdilutional % coefficientof variation (CV) (closed circles; measure of parallelism)and interlaboratory %CV (open circles; measure of interlaboratoly variation) are plotted for the 21 test and control sera as a function of their mean antibody level. Target levels for interdilutional and interlaboratory CVs are 20% and 35%, respectively(dashed and dotted lines).

lift CORRELAIIONOF E!A TOTALVS SUM AND RIA

ing ranges that are displayed by the four individual assays. Table I summarizes the observed working ranges and recommended dilutions for analysis of pediatric (P) and adult (A) human sera in the l g G l - 4 and total IgG anti-Hib immunoassays. Interestingly, the working ranges of the total igG and lgG3 anti-Hibs are the widest (> 100fold), due in part to the higher affinity detection monoclonal antibodies used in these assays (Phillips et al., 1987). The working ranges of the lgGl, IgG2, and IgG4 anti-Hib assaysvaried from a 62to 16-fold serum dilution range. Reproducibility. An important goal of this study was the investigation of the magnitude of variation between laboratories in a multieenter study that all used the same reference serum, assay reagents and methods. Interlaboratory variation was assessed by computing coefficients of variation using final IgG1-4 and lgG Pan anti-Hib results in /~g/ml. Interlaboratory precision profiles are presented for total lgG (Fig. 2,4, open circles) and for each IgG subclass assay (Fig. 2B, open circles). At higher levels of IgG1-4 and total antibody (e.g., IgG anti-Hib > 5 /zg/ml, ]gGl anti-Hib > 3/zg/ml), most interlaboratory CVs were within the accepted target level of 35% or less. At lower/~g/ml concentrations of antibody, interassay CVs exceeded 35%, which probably reflects the fact that the enzyme immunoassays used in this study are multi-step procedures with many reagents and incubation times that are sensitive to minor variations, especially in the low working range. Assay variation may change under normal laboratory circumstances when individual laboratories become familiar with the assay and use their own plates and reagents. The observed level of variation in the current study is consistent with intcrlaboratory variation obtained in a recent lUIS-sponsored collaborative study of immunoassays for quantitation of human IgG subclass protein (Kcmeny, Hamilton et al., unpublished). The observed interlaboratory variation is partly attributed to intralaboratory, interassay variability that ranged from 15-30% CV for replicate measurements performed at several sites with total lgG anti-Hib levels > 1 / z g / m l (data not shown). Parallelism and internal agreement. Two hallmarks of a quantitative IgG subclass antibody

5o

ol

•

,

y- 1.1~(-o.ls

t9G1-4 onti-HIB TOTAL(Hq/ml) by E~

Fig. 3, Correlationplots of IgG anti-Hib levels(/~g/ml) as measuredin the total lgG anti-HibEIA (x axis)versusth(r-¢ measured by RIA (closedcircles, right axis) or obtained by summation of the IgGl, lgG2, IgG3 and lgG4 anti-Hib EIA results (open circles, left axis). There was a good correlation (r e = 0.97) but a systematicbias (Y = 4.6X -5.8) between the lgG anti-Hib levels measured in the EIA and RIA. Within the EIAs, there was good agreement between fie total IgG antiHib and summed lgGI-4 antibody levels (r 2 = 0.99, slope = I. 13, Y intercept = 0.15) which is an indicator of good IgGI-4 EIA performance.

immunoassay are parallelism, or the ability of ~he test and reference sera to dilute in parallel and internal agreement between the sum of I g G I - 4 and the total. In a parallel assay, antibody estimates performed with individual sera analyzed at multiple dilutions yield comparable results after interpolation and correction for dilution, in the current study, interdilutional coefficients of variation, an indicator of parallelism, are presented in Fig. 2 (closed circles) for the l g G l - 4 and total lgG anti-Hib assays. All assays demonstrated good parallelism with I D % C V s of approximately 20% over the entire working range of each subclass assay. Fig. 3 (open circles) presents the correlation between the sum of the individual IgG1, IgG2, lgG3 and lgG4 anti-Hib levels and the total lgG anti-Hib concentration as measured in separate enzyme immunoassays. The l g G l - 4 sum and total lgG antibody levels agree well ( Y = 1 . 1 3 X 0.15, r 2 = 0.99). As suggested by other lgG subclass studies (Hamilton et al., 1988, 1990b; Kemeny, Hamilton et al., unpublished), internal agreement between the sum of the four individual subclass measurements and the total by a

108 A.

10~

IgG

~ HSP4 y.3.509-0.974x R^2.0.992

o .1

.01

1000

B.

'

......................

'....

10000 100000 RECIPROCAL DILUTION

1000000

to IgG1

19G2

y132S9.1.177= R'2.0991

v.;e2s.o ol~:. R'2.0 8119

.1 .1 i

::,:::os. Ol lOOO

.Ol 100000 lOO lO ~ IgG4

_ , 10000

lOOO

lgG3

1

10000

100000 • y.O,831.o,71Ix R,2.o.osa

i-1~4 y.l.ool-o,744: .2.o.~19

|

1 ~ • FDAlg~ y.O.7SS-o.es4= nag.O.gg3 .01 ........ , ....... . ........ 1 lO lOO Iooo 10000 IO RECIPROCAL OILUTION

_

~

e

FDA1903

........

, . . too RECIPROCAL OILUTION

Fig. 4. DUutional analysis o f the F D A 1983 anti-Hib reference serum against the calib+'ated HSP4 in the total IgG ( A ) and l g G I - 4 ( B ) anti-Hib EJAs. The slopes o f the dilution curves were not different for the two sera, demonstrating assay parallelism. Direct interpolation produced the I g G l - 4 and total l g G anti-Hib estimates in the F D A 1983 that are summarized in Table IlL

separate assay is a mark of an IgG subclass immunoassay quality.

Agreement between methods The radiolabeled antigen binding assay is widely used for the measurement of h u m a n Hib-

specific antibody and uses the C B E R F D A lot 1983 h u m a n serum as a reference standard. Fig. 3 (closed circles) presents the correlation between E I A measurements of total IgG anti-Hib and R I A measurements of total Hib-specific antibodies. While the correlation between the two meth-

109 TABLE II TOTAL AND lgGI-4 ANTI-Hib LEVELS (/zg/ml) MEASURED BY EIA AND RIA IN THE TEST AND CONTROL SERA Serum group

Serum #

lgGI mean

Pediatric ( < 2 years)

PED I PED 2 PED 3 PED 4

1.22 0.77 2.25 0.36

Adult individual

ADULT 10 ADULT 16 ADULT 7 ADULT 14 ADULT 9 ADULT 15 ADULT 8 ADULT 17 ADULT ! I ADULT 12 ADULT 13

Adult pool

lgG2 mean

lgG3 mean

lgG4 mean

0.45 0.40 1.89 0.86

0.154 0.134 0.226 0.058

0.0033 0.0021 0.0175 0.0084

2.03 1.31 5.27 1.25

0.73 0.82 0.37 3.95 0.13 0.91 0.18 2.01 0.27 0.70 1.49

1.43 13.31 0.52 10.14 0.84 13.88 0.97 12.95 5.13 2.85 4.70

0.027 0.076 0.011 0.036 0.047 0.131 0.010 0.206 0.029 0.024 0.044

0.0042 0.0121 0.0025 0.0119 0.0041 0.0069 0.0027 0.0093 0.0053 0.0050 0.0059

ADULT 2 ADULT 3 ADULT 6 ADULT 1

0.40 0.64 4.32 0.12

!.81 4.31 53.39 1.18

0.032 0.045 0.225 0.018

0.0028 0.0046 0.0452 0.0027

Pool adult-child

ADULT 4

0.23

15.88

0.036

0.0104

HSP-3

ADULT 5

5.27

20.37

0.254

0.0158

o d s is t i g h t ( r 2 = 0.97), t h e r e a p p e a r s to b e a b i a s t o w a r d 4 - 5 - f o l d h i g h e r levels (Y= 4 . 6 X - 5 . 8 ) w i t h t h e R I A ( s e e T a b l e II).

Calibration and agreement between standards To date, there has been no internationally agreed upon human serum reference which has

lgG-Pan mean

Sum IgGl-lgG4

Ratio Pan/sum

Total RIA

1.829 1.307 4.390 1.293

I. I I 1.00 1.20 0.97

3.89 3.67 22.9 1.85

3.14 16.25 0.56 14.90 2.61 15.48 1.74 18.46 4.39 5.48 6.70

2.189 14.217 0.901 14.144 1.028 14.928 1.161 15.173 5.436 3.583 6.233

1.44 1.14 0.62 1.05 2.54 1.04 1.50 1.22 0.81 1.53 1.07

4.97 76.7 1.32 42.3 0.837 74 0.682 27.0 15.5 17.1 46.1

3.21 5.13 62.65 1.27

2.239 5.002 57.981 1.326

1.43 1.03 i.08 0.96

1.54 5.31 254 0.433

16.06

16.159

0.99

42.2.

24.54

25.914

0.95

66.6

b e e n c a l i b r a t e d in t e r m s o f its l g G I - 4 a n t i - H i b content. We employed a reported method of c r o s s - c a l i b r a t i o n ( H a m i l t o n , 1991) u s i n g c h i m e r i c a n t i b o d i e s to e s t i m a t e t h e c o n t e n t o f h u m a n l g G l - 4 a n t i - H i b in an a d u l t s e r u m p o o l (HSP4). A n t i b o d y e s t i m a t e s in t h e H S P 4 w e r e 9.7, 43.3, 0.8 a n d 0.04 / z g / m l o f I g G l , l g G 2 , l g G 3 a n d

TABLE !11 lgGI-4 ANTI-Hib ANTIBODY LEVELS IN HUMAN REFERENCE SERUM POOLS Method

Pool #

lgG-Pan /~g/ml

lgGl ~ g /ml

IgG2 /.tg/ml

IgG3 gg/ml

IgG4 /.tg/ml

Sum IgGI-4 ~g / ml

% Difference between sum and Pan

HET a Ppt b

4 4

64.3 70.0

9.7 15.0

43.3 40.0

0.8 1.5

0.04