Immunology Letters, 34 (1992) 213-220 0165 - 2478 / 92 / $ 5.00 © 1992 ElsevierSciencePublishers B.V. All rights reserved

IMLET01869

Rapid immunoassays for the measurement of immunoglobulin G subclass concentration in immunoglobulin preparations and human serum Safia Wasi, Susan A n n e Th6riault a n d Peter Gill The Canadian Red Cross Society, National Reference Laboratories, Protein Chemistry Laboratory, Ottawa, Ontario, Canada

(Received 26 June 1992;revisionreceivedand accepted 8 September 1992)

1.

Summary

Enzyme immunoassays (EIA) capable of determining total IgG1, IgGz, IgG3 and IgG4 subclass concentrations in human serum preparations have been developed. Subclass-specific monoclonal antibodies (mAbs) are bound to polyacrylamide bead-conjugated anti-mouse immunoglobulin antibodies. Bound immunoglobulins are detected with a peroxidase-conjugated anti-IgG antibody or a biotin-conjugated anti-IgG antibody followed by peroxidase streptavidin. The standard curves were found to be linear in the regions 16.0-2.0 ktg/ml for IgGl, 4.04).5 #g/ml for IgG2, 0.44).06 pg/ml for IgG3 and 0.25-0.05/~g/ml for IgG4. Coefficient of variation (CV) values range from 0.32-7.32% for IgG1, 0.66--4.85% for IgG2, 1.62-6.85% for IgG3 and 0.05-6.47% for IgG4 standard curves. The inter-assay variability for the control human serum samples was 9.6% for IgGl, 6.7% for IgG2, 9.5% for IgG3 and 6.8% for IgG4.

2.

Introduction

Human immunoglobulin G (IgG) consists of four subclasses, IgGl, IgG2, IgG3 and IgG4, Key words: Immunoassay;ImmunoglobulinG; Subclass Correspondence to: SafiaWasi,The CanadianRed Cross Society,National ReferenceLaboratories,ProteinChemistryLaboratory, Ottawa, Ontario,CanadaKIG 4J5.

which are distinguished by variations in their heavy chains [1,2]. Immunoglobulin G isotype constitutes approximately 75% of total serum immunoglobulin in normal human adults. The subclass profile of circulating IgG comprises approximately 70% IgG1, 20% IgG2, 7% IgGa and 3% IgG4 [3]. The four subclasses exhibit different biological properties, such as, complement fixation, ability to cross placenta, and binding to macrophages [4]. Certain disease states may cause changes in total IgG levels as well as subclass distribution. IgG-subclass deficiencies are frequent in obstructive lung disease patients [5]. In several immune diseases, such as X-linked agammaglobulinemia, decreased levels of all IgG subclasses have been reported [6,7]. Deficiencies of one or more subclasses are also associated with recurrent sino-pulmonary infections both in adults and pediatric patients [8-10]. Intravenous and specific immunoglobulin preparations are being used for the management of the above and a variety of primary immunodeficiencies such as congenital agammaglobulinemias, common variable immunodeficiency, Wiskott-Aldrich syndrome and severe combined immunodeficiencies [10-12]. The intravenous immunoglobulin preparations are also used prophylactically to raise platelet counts prior.to surgery in idiopathic thrombocytopenic purpura patients [13]. The ability of these preparations to opsonize and neutralize microbes and toxins makes them important in the management of primary and recurrent bacterial infections, as in hypogammaglobulinemia, B-cell chronic lymphocytic leukemia 213

and tetanus [14,15]. These preparations are also important in the management of a variety of viral mediated disease states such as varicella-zoster, hepatitis B, rabies and human cytomegalovirus (HCMV). In most immunoglobulin preparations the IgG subclass concentrations are proportionally the same as those found in normal human serum. For anti-viral and anti-bacterial treatment the efficacy of hyperimmune globulin preparations should depend on the concentrations of neutralizing antibodies. Furthermore, it has been shown that the profiles of reactivity vary with respect to immunoglobulin subclasses for different infections [16]. For example the major IgG subclass responses to HCMV are elicited by IgG1 and IgG3 [17,18]. The ability of IgG~ and IgG3 to fix complement and to bind to phagocytic cells suggests that these subclasses may play an important role in controlling the virus and ameliorating the infection [19,20]. The neutralizing capacity of anti-HCMV IgG3 was found to be ten-fold higher than that of anti-HCMV IgGl [20,21]. This may suggest that anti-HCMV immunoglobulin preparations would be more efficacious if high proportions of neutralizing anti-HCMV-IgGl and anti-HCMV-IgG3 subclasses were present. Highly sensitive assays are required to allow more precise evaluation of intravenous immunoglobulin preparations which may enhance the efficiency in the management of IgG deficiency states. IgG subclass assays, currently available, include radial immunodiffusion (RID), enzymelinked immunosorbent assay (ELISA) and radioimmunoassay (RIA) which require long incubation periods or the use of radioisotopes [23,24]. It has been observed that the RID assay was not sufficiently sensitive for IgG subclass quantification [25]. Therefore rapid and sensitive assays were required to quantify the IgG subclass concentrations in the therapeutic immunoglobulin preparations and in sera from patients with abnormal levels of IgG subclass concentrations. This paper describes enzyme immunoassays which measure IgG1, IgG2, IgG3 or IgG4 in normal sera, subclass-deficient sera, or immunoglobulin preparations utilizing subclass-specific monoclonal antibodies (mAbs) linked to rabbit anti-mouse immunoglobulin-coupled Immunobeads. These assays eliminate the use of radioiso214

topes and offer a simple, rapid, and sensitive method for measuring IgG subclass concentrations. A sensitivity 100-1000-fold higher than RID method can be attained. Furthermore, the assays are more economical than commercial EIA kits. Sigmoidal standard curves ranging from 100 #g/ml to 0.1 #g/ml (for IgGl and IgG2) and 10 #g/ml to 0.01 #g/ml (for IgG3 and IgG4) were obtained. 3. 3.1.

Materials and Methods Materials

Rabbit anti-mouse immunoglobulin antibodies covalently linked to polyacrylamide Immunobeads were obtained from Bio-Rad Laboratories (Canada) (Mississauga, Ont.). Monoclonal antiIgG1 antibodies were obtained from Zymed Laboratories Inc. (Dimension Labs Inc., Mississauga, Ont.) and AMAC, Inc. (Bio/Can Scientific, Mississauga, Ont.), mAb to IgG2 from Nordic Immunologicals (Cedarlane Labs, Hornby, Ont.), mAb to IgG3 from AMAC, Inc. (Bio/Can Scientific, Mississauga, Ont.) and mAb to IgG4 from Pharmingen (Cedarlane Labs, Mississauga, Ont.). Horseradish peroxidase-conjugated goat F(ab')2 anti-human IgG antibody was from Jackson Immunoresearch Labs (Bio/Can Scientific, Mississauga, Ont.). Biotin-conjugated goat anti-human IgG was from Cappel (Organon Teknika Inc., Scarborough, Ont.). Peroxidase-conjugated streptavidin was from Calbiochem Corporation (Terochem Laboratories Ltd., Edmonton, Alta.). Quality Control Serum (lot CS012.C) for subclass control reference and IgG subclass standards were purchased from The Binding Site (Oxoid Canada Inc., Nepean, Ont.). CAP Reference Preparation for Serum Proteins (lot 4) was obtained from the College of American Pathologists (Skokie, IL). Normal rabbit serum (NRS), RIA grade bovine serum albumin (BSA) and ophenylenediamine dihydrochloride (OPD) were from Sigma Chemicals (St. Louis, MO). Dynatech Immulon 96-well microtiter plates were obtained from Dynatech Laboratories (Fisher Scientific, Ottawa, Ont.). All other reagents were of analytical grade.

3.2. IgG subclass enzyme immunoassays These assays measure the amount of IgGl IgG2, IgG3 and IgG4 that binds to subclass-specific mAbs linked to rabbit anti-mouse immunoglobulin Immunobeads. The bead matrix was first prepared by incubating 4 #g of either anti-IgG2, anti-IgG3 or anti-IgG4 monoclonal antibodies per mg of beads (bead suspension, 1 mg/ml in phosphate-buffered gelatin (PBG); 150 mM phosphate, 1 mM EDTA, 0.1% gelatin, pH 7.4), in a 10 x 100 mm glass tube for 1 h at 37°C. For IgG1 a mixture of two different mAbs was used, each at a concentration of 4 /~g mAb/ml beads. The beads were then centrifuged at 1500 x g for 10 min, the buffer decanted and the beads resuspended in 10 ml PBG. This wash cycle was repeated twice, before the bead suspension was adjusted to a working concentration of 1 mg/ml in PBG. The standards (from normal human serum) were prepared in PBG and serially diluted (× ½) as follows: 64.0-0.125/~g/ml for IgG1, 80.04).156 #g/ml for IgG2, 4.0-0.016 #g/ml for IgG3 and 8.00.016 #g/ml for IgG4. PBG (0.200 ml), mAb beads (0.100 ml) and standard or control solution (0.100 ml) were mixed in 12 x 75 mm glass tubes, in quadruplicate. The tubes were incubated for 1 h at ambient temperature followed by two 3-ml washes of PBG, as described above. Two ml of peroxidase-conjugated goat anti-human IgG (F(ab')2 fragment (diluted 1/20000 in PBG, 1% BSA, 1/ 100 normal rabbit serum (NRS)) were added to the IgG3 assay tubes. Two ml of biotinylated goat anti-human IgG (heavy and light chain specific; diluted 1/2000 in PBG, 1% BSA, 1/100 NRS) were added to the IgG1 and IgG2 assay tubes, respectively. IgG4 assay tubes were incubated as for IgG1 and IgG2 except the NRS was diluted 1/75. The beads were re-suspended by vortexing and incubated for 1 h at ambient temperature prior to washing two times as above, followed by the addition of 2 ml of peroxidase-conjugated streptavidin (diluted 1/2000, 1% BSA in PBG) to IgG1, IgG2 and IgG4 assay tubes. The beads were resuspended by vortexing and incubated for 45 min at ambient temperature. This step was not performed for the IgG3 assay. The beads were washed twice and briefly drained onto a

stack of filter paper. One ml of substrate solution, OPD in 200 mM citrate-phosphate buffer, pH 5.5, 0.025 ml 30% hydrogen peroxide (0.042 mg/ml for IgG1 and IgG4, 0.033 mg/ml for IgG2 or 0.050 mg/ml for IgG3) was added to each tube and the colour was allowed to develop. The reaction was stopped after 4.5 min by adding 0.100 ml 6 N HCI to each tube. The beads were spun down at 1500 x g for 10 min and aliquots (0.200 ml) from each tube were transferred to blank microtiter plates to facilitate rapid spectrophotometric quantification at 492 nm on a Biomek 1000 Automated Workstation (Beckman Instruments, Mississauga, Ont.). Data analysis was performed with Immunofit EIA/RIA software (Beckman Instruments, Mississauga, Ont.).

3.3. Radial immunodiffusion assay Immunoglobulin subclass concentrations of sera from patients with sinopulmonary infections, were determined using commercially prepared RID plates from The Binding Site Inc. (San Diego, USA). 4.

Results

4.1. Evaluation of monoclonal antibodies For each of the assays, a series of mAbs was screened to select one that showed a high degree of reactivity with the corresponding immunoglobulin subclass antibody standards. The selection was based on a satisfactory absorbance (A 492 nm) signal over a wide range of immunoglobulin subclass concentrations. Those mAbs which gave a wide spectrum of absorbances were further evaluated by experimenting with a variety of anti-human IgG antibody conjugates to determine whether the signals could be augmented to produce a useful concentration range. The A ranges for the various mAbs used were: 0.200--0.880 (blank 0.290) for IgGl, 0.360-0.920 (blank 0.260) for IgG2, 0.290-0.650 (blank 0.100) for IgG3 and 0.2504).720 (blank 0.150) for IgG4. Two mAbs were used in conjunction to measure IgG1 concentrations since no single mAb that was tested gave a usable A range. For the other subclasses, only a single mAb was required. 215

a)

1.o-

0,8oJ 0 . 6 o~ O

(b) lO.

y

12-

O.4

0.8

0.6

a~ 0

0.4

O2

0.2 .o..I-

0

O0.1

1.0

100

pqlrnl

0.1

1000

1.0

100 pg/ml

IgG I

C ) 1.8"

(d) 1.2.

1.5-

1.0

I.?

100D

IgO 2

0.8. E

rE 09: ,7, o,¢ O 0.6

06 0

0,4

0.2

0.3 O ~

001

O

03

~b I~/ml

I00

0.01

0.1

1.0 pglml

IgG 3

IOQ

IgG4

Fig. 1. (a) IgGl standard curve. Concentration of IgG1 subclass standard versus absorbance at 492 nm. Each point is a mean of quadruplicate values and the blank (,4 = 0.290) has been subtracted. (b) IgG2 standard curve. Concentration of IgG2 subclass standard versus absorbance at 492 nm. Each point is a mean of quadruplicate values and the blank (A = 0.260) has been subtracted. (c) IgG3 standard curve. Concentration of IgG3 subclass standard versus absorbance at 492 nm. Each point is a mean of quadruplicate values and the blank (A = 0.100) has been subtracted. (d) IgG4 standard curve. Concentration of IgG4 subclass standard versus absorbance at 492 nm. Each point is a mean of quadruplicate values and the blank (A = 0.150) has been subtracted.

4.2.

Standardization

T h e e v a l u a t e d c o n c e n t r a t i o n ranges o f m A b s were used to generate s t a n d a r d curves. U s i n g a s t a n d a r d i z e d p r e p a r a t i o n o f n o r m a l h u m a n serum, I g G l , IgG2, IgG3 a n d IgG4 s t a n d a r d curves were elicited for each subclass. G r a p h s o f I g G subclass c o n c e n t r a t i o n versus a b s o r b a n c e at 492 n m p r o d u c e d sigmoidal curves in the regions between 32.0-0.25 #g/ml for IgG1 (Fig. la), 40.0-0.31 #g/ ml for IgG2 (Fig. lb), 4.0--0.16 #g/ml for IgG3 (Fig. lc) a n d 4.0-0.16 #g/ml for IgG4 (Fig. ld). These curves were linear between 16.0-2.0 # g / m l for I g G l , 4.0-0.5 # g / m l for IgG2, 0.44).06 # g / m l for IgG3 a n d 0.254).05 # g / m l for IgG4. Coefficient o f v a r i a t i o n values r a n g e d f r o m 0.32-7.34% 216

for IgG1, 0 . 6 6 - 4 . 8 5 % for IgG2, 1.62-6.85% for IgG3 a n d 0 . 0 5 - 6 . 4 7 % for IgG4 s t a n d a r d curves.

4.3.

Controls

U s i n g the a b o v e assays quality c o n t r o l h u m a n reference serum f r o m T h e B i n d i n g Site a n d hum a n c o n t r o l s e r u m f r o m T h e College o f A m e r i can P a t h o l o g i s t s ( C A P ) were a n a l y z e d for I g G b IgG2, IgG3 a n d IgG4 c o n t e n t a n d the values obt a i n e d were all within 10% o f the m a n u f a c t u r e r ' s stated values (1.8% for I g G b 0.3% for IgG2, 6.7% for IgG3 a n d 0.6% for IgG4). I n t e r - a s s a y v a r i a b i l i t y for the c o n t r o l serum samples was 9.6% for IgGa, 6.7% for IgG2, 9.5% for IgG3 a n d 6.8% for IgG4.

TABLE 1 Comparison of RID and EIA techniques for the determination of IgG subclass concentrations in deficient sera. RID, radial immunodiffusion; EIA, immunobead enzymeimmunoassay; ND, not detectable. RID data were kindly provided by Dr. C. Roifmann. Sample Technique used

Concentration (#g/ml) IgGi

lgG2

IgG3

IgG4

R.L.

RID EIA

4680.0 4762.8

540.00 954.97

210.00 215.43

39.00 81.50

H.H.

RID EIA

4600.0 4231.5

980.00 1375.40

300.00 306.98

110.00 116.94

K.F.

RID EIA

6390.0 5785.2

570.00 859.68

4 2 0 . 0 0 450.00 5 9 4 . 7 8 101.36

R.C.

RID EIA

4640.0 4634.5

700.00 401.57

240.00 218.25

ND 22.32

L.S.

RID EIA

7770.0 6861.6

780.00 1112.28

520.00 530.80

170.00 153.78

M.V.

RID EIA

6710.0 6832.2

9 9 0 . 0 0 960.00 1047.60 1214.70

215.00 236.36

4.4.

Assay evaluation utilizing abnormal patient's sera

Sera from six patients (ages 6-17) with sinopulmonary infections were tested for IgG subclass levels (generously provided by Dr. C. Roifman, Division of Immunology and Allergy, Hospital for Sick Children, Toronto). The immunobead technique was evaluated using the above sera and results were compared with RID method (Table 1). IgG2 subclass concentrations in all six sera were found to be below the expected normal ranges for this age group [26]. The RID technique was unable to detect a measurable quantity of IgG4 concentration in one of the specimens; whereas the immunobead assay could detect measurable quantities of this subclass in all samples. Moreover, the detection limits (0.1 pg for IgG1 and IgG2; 0.01 #g for IgG3 and IgG4) of the assays described here were several orders of magnitude higher than that of RID method (10 #g for

all subclasses; data not shown). The immunobead inter-assay variability for the patient sera tested was 9.5% for IgGb 16.0% for IgG2, 9.7% for IgG3, and 6.0% for IgG4. 5.

Discussion

The assays described above offer a fast, sensitive and inexpensive alternative to commercially available ELISA, RIA and RID assays. The bead assay format has an added advantage over other procedures, for example, polyacrylamide beads provide a three-dimensional surface for binding and antibody presentation as compared with the two-dimensional surface microtiter wells. Other drawbacks of microtiter wells are weak interactions between proteins, peptides and polystyrene surfaces as well as significant losses in washing steps. In previous studies we have used polyacrylamide beads to quantitate IgA subclasses and vitronectin [27-29]. In this study we have modified and optimized the polyacrylamide bead assays to measure IgG subclass concentrations in normal human serum, patient sera with sino-pulmonary infections, and immunoglobulin preparations. The assays have good sensitivity, coefficient of variation, and inter-assay variability as compared with commercial assays. The assays described utilize commercially available beads to which rabbit anti-mouse immunoglobulin antibodies have been covalently linked. A schematic diagram of the general assay set-up, shown in Fig. 2, depicts the two layers of capture antibodies, followed by the sample and then the colorimetric detection system. This allows any mAb preparation to be bound with relative ease, requiring only a 1 h incubation period. Monoclonal antibodies from different commercial sources were compared with respect to their reactivities with the IgG subclass antibody standards. This was subjectively tested first on the basis of signat strength with a range of standard concentrations. A suitable mAb was selected to optimize conditions for a variety of anti-IgG antibody conjugates. The evaluation performed in this study reveals that IgG subclass specific mAbs can vary widely in their antigen affinity (data not shown). This has previously been reported in other studies [30]. The mAbs which displayed a 217

CONTROL

HORSERADISH or

POLYACRYLAMIDE BEADS

PEROXIOASE

STANDARD

(1 - 10 U i )

oI

IgG

I "

492nm

C O N T R U L Or STANDARD or

Fig. 2. Schematic diagram outlining the EIA set-up for measuring IgG~, IgG2, IgG3 and IgG4 subclasses. significantly higher avidity for the standards were studied further to confirm the subclass specificity. The assays described for measuring I g G l , IgG2, IgG3 and IgG4 can be completed in 5 to 7 hours and are easily adaptable to a u t o m a t e d E I A workstations. Efforts were m a d e to utilize as m a n y commercially available reagents as possible. The m e t h o d o l o g y is currently being utilized to examine the levels o f I g G subclasses in experimental and commercially available HCMV-specific immunoglobulins and intravenous i m m u n o g l o b u lins.

Acknowledgements This w o r k was supported by a grant f r o m Miles-Cutter/The C a n a d i a n Red Cross Society Research and D e v e l o p m e n t Fund. We would like to t h a n k Dr. Leslie A. Mitchell (Departments o f P a t h o l o g y and Pediatrics, University o f British Columbia) for helpful discussions, and Dr. C. Roifman, Head, Division o f I m m u n o l o g y , Hospital for Sick Children, T o r o n t o for subclass deficient sera samples.

218

References [1] World Health Organization (1966) WHO Technical Series. Report 35, 953. 12] Goodman, J.W. (1987) Immunoglobulins 1: structure and function. In: Basic and Clinical Immunology. (P.D. Stites, J.D. Stobo, H.H. Fudenberg and J.V. Wells, eds.) p. 27. Lange, Los Altos, CA. [3] Benjamini, E. and Leskowitz, S. (1988) Immunology. A Short Course, p. 64. Alan R. Liss, New York. [4] Jefferis, R. (1990) in: The Human IgG Subclasses: Molecular Analysis of Structure, Function and Regulation (F. Shakib, ed.) pp. 93-108. Pergamon Press, Toronto. [5] Eibl, M.M., Pum, M., Bernatowska, E. and Liebl, H. (1990) P/idiatrie und P~dologie 25, 231. [6] Wilson, N.W., Daaboul, J. and Bastian, J.F. (1990) J. Clin. Immunol. 10, 330. [7] Robbins, J.B., Skinner, R.G. and Pearson, H. (1979) N. Engl. J. Med. 280, 70. [8] Klaustermeyer, W.B., Wong, S.C., Schoettler, J.J., Gianos, M.E. and Heiner, D.C. (1989) Ann. Allergy 63, 327. [9] Sehoettler, J.J., Schleissner, L.A. and Heiner, D.C. (1989) Chest 96, 516. [10] Silk, H.J., Ambrasino, D. and Geha, R.S. (1990) Ann. Allergy 64, 21. l 11] Stiehm, E.R. (1979) Pediatrics 63, 301. [12] Buckley, R.H. (1979) in: Immunoglobulins: Characteristics and Use of Intravenous Preparations (B.M. Alving, J.S. Finlayson, eds.) pp. 3-8. Department of Health and Human Services, Washington, D.C.

[13] Bussel, J.B., Kimberly, R.P., Inman R.D., Schulman, I., Cunningham-Rundles, C., Cheung, N., Smithwick, E.W., O'Malley, J., Barandun, S. and Hilgartner, M.W. (1983) Blood 62, 480. [14] Bunch, C., Chapel H.M., Rai, K., the CLL Clinical Working Group and Gale, R.P. (1987) Blood 70, Suppl. 1, 753, 224a. [15] Cooperative Group for the Study of Immunoglobulin in Chronic Lymphacytic Leukemia (1988) N. Engl. J. Med. 319, 902. [16] Hamilton, R.G. (1989) in: The Human IgG Subclasses, Table 6, p. 37. Calbiochem Corporation, San Diego, CA. [17] Linde, G.A., Hammarstrom, L., Persson, M.A.A., Smith, C.I.E., Sundqvist, V-A. and Wahren, B. (1983) Infect. Immun. 42, 237. [18] Gilljam, G., Sundqvist, V-A., Linde, A., Pihlstedt, P., Eklund, A.E. and Wahren, B. (1985) J. Virol. Met. 10, 203. [19] Spiegelberg, H. (1974) Adv. lmmunol. 19, 259. [20] Meyers, J.D. (1991) Annu. Rev. Med. 42, 179. [21] Wahren, B., Linde, A., Sundqvist, V-A., Ljungman, P., Lonnqvist, B. and Ringden, O. (1984) Transplantation 38,

479. [22] Gilljam, G. and Wahren, B. (1989) J. Virol. Met. 25, 139. [23] Hayashi, K., Miyasaka, H. and Tagawa, M. (1990) Anal. Lett. 23, 1017. [24] Tijhuis, G.J., Klaassen, R.J.L., Modderman, P.W., Ouwehand, W.H. and Kr. von dem Borne, A.E.G. (1991) Br. J. Haematol. 77, 93. [25] Meissner, C., Reimer, C.B., Black, C., Broome, C., Rabson, A., Siber, G.R., Delaney, N., Connors, M. and Ambrosino, D.M. (1990)J. Pediatr. 117, 726. [26] Oxelius, V.A. (1979) Acta Paediatr. Scand. 68, 23. [27] Haun, M. and Wasi, S. (1989) J. Immunol. Met. 121, 151. [28] Haun, M. and Wasi, S. (1990) Immunol. Lett. 26, 37. [29] Haun, M. and Wasi, S. (1990) Anal. Biochem. 191,337. [30] Jeferis, R., Reimer, C.B., Skvaril, F., de Lange, G., Ling, N.R., Lowe, J., Walker, M.R., Phillips, D.J., Aloiso, C.H., Wells, T.W., Vaerman, J.P., Magnusson, C.G., Kubagawa, H., Cooper, M., Vartdal, F., Vandvik, B., Haaijnan, J.J., Makela, O., Sarnesto, A., Lando, Z., Gergely, J., Rajnav61gyi, E., L/tszl6, G., Radl, J. and Molinaro, G.A. (1985) Immunol. Lett. 10, 223.

219