Characterization of monoclonal antibodies physically adsorbed onto polystyrene latex particles

Characterization of monoclonal antibodies physically adsorbed onto polystyrene latex particles

Journal of hn m unologic a l Methods, 152 (1992) 19 I- 199 19 I © 1992 Elxevier Science Publishers B.V. All rights reserved 0022-!759/92/$05.0{] JI...

589KB Sizes 0 Downloads 44 Views

Journal of hn m unologic a l Methods, 152 (1992) 19 I- 199

19 I

© 1992 Elxevier Science Publishers B.V. All rights reserved 0022-!759/92/$05.0{]

JIM []6371

Characterization of monoclonal antibodies physically adsorbed onto polystyrene latex particles R. v a n E r p , Y . E . M . L i n d e r s , A . P . G . v a n S o m m e r e n a n d T.C.J. G r i b n a u Clinical Lab Systems Research Unit, Organon Teknika B.V., Boxtel. Netherlands

(Received 22 October 1991,revised received 7 February 1992,accepted 12 March 1992)

The physicochemical and immunochemical properties of monoclonal antibodies (MAbs), adsorbed onto polystyrene latex particles, have been investigated. Both native and pH 2 pretreated MAbs were compared before and after immobilization. It was found that the antigen binding capacity of the immobilized, acidic pretreated MAbs was significantly higher than for the immobilized, native lgG molecules. This enhanced antigen binding capacity appeared to be due to an improved molecular orientation following adsorption of the monomeric, pH 2 treated IgG fraction. Additionally, experiments using F ( a b ' ) 2 fragments demonstrated that the Fc portion of the MAb molecule is of major importance for achieving the enhanced binding capacity. Binding studies showed that the (apparent) affinity of native and pH 2 pretreated MAbs were similar; the K~ values of the immobilized MAbs were higher than those of MAbs in solution. Key words: Monoclonalantibody; Polystyrenelatex particle; Adsorption; Low pH treatment

Introduction The immobilization of monoclonal antibodies onto solid phase supports has been extensively used in diagnostic test systems such as ELISA,

Correspondence to: R. van Erp, SSRU, Organon Teknika B.V., P.O. Box 84, 5280 AB Boxtel, Netherlands. Abbreciations: A2so, absorbance at 280 nm; Ag, antigen; BSA, bovine serum albumin; DASP, double antibody solid phase; ELISA, en~me-linked immunosorbent assay; GaM, goat anti-mouse; hCG, human chorionic gonadotrophin; HPSEC, high performance size exclusion chromatography; HRP, horseradish peroxidase; lgG, immunoglobulinG; IU. international units; MAb. monoclonal antibody; PBS, phosphate-buffered saline; PEG, polyethylene glycol; RIA, radioimmunoassay; RT, room temperature; SaM, sheep antimouse; SPIA, sol particle immunoassay;UV, ultraviolet.

SPIA, RIA and latex agglutination assays (Parson, 1981; Tijssen, 1985; Bernard et al., 1986; Kakabakos et al., 1990). One of the most commonly used solid supports in immunoassays is polystyrene and methods for immobilization are predominantly based on physical adsorption. However, physical adsorption of MAbs onto solid supports can lead to partial or complete loss of the antigen binding activity as a result of either (a) steric hindrance caused by the attachment of the MAb to the solid support at a region close to its active site or (b) conformational changes or (c) restriction of the intramolecular flexibility imposed by multi-site attachment to the solid support. Nevertheless an improved binding activity of the physically adsorbed MAbs can be obtained by exposing the MAbs to a low pH environment prior to immobilization. Comparable results have

192 been described for polyclonal antibodies (Ishikawa et al., 1981; Conradie et al., 1983). Relatively few studies have addressed the molecular basis of this interestin? phenomenon which appeared to affect antigen binding capacity and test performance. Recently, a more detailed explanation has been published (Lin et al., 1989) regarding the influence of adsorption onto silica surfaces for native and 'pH modified' polyclonal antibodies. The main objective of this study was to gain more insight into the properties of physically adsorbed MAbs onto a polystyrene surface. Particular interest was paid to the effect of a pH 2 pretreatment by comparing the physicochemical and immunochemical properties of both native and pH 2 pretreated MAbs before and after immobilization. For this purpose, colloidal polystyrene latex particles were chosen because of the rapid reaction of the particle bound MAb with its antigen and easy experimental handling.

Organon Teknika (Boxtel, Netherlands). Sephacryl S-200 HR, S-300 HR, Protein A-Sepharose CL-4B and molecular weight standards for HPSEC (thyroglobulin, 669,000; ferritin, 440,000; IgG, 150,000; BSA, 67,000; ovalbumin, 43,000 and chymotrypsinogen-A, 25,000) were purchased from Pharmacia (Uppsala, Sweden). All other chemicals were of analytical reagent grade quality.

Protein determinations Protein concentrations were measured by UV spectroscopy at 280 nm, using the extinction coefficients of 0.39 (cm. mg/ml)- l for highly purified hCG and 1.45 (cm. mg/mi)- i for lgG and F(ab') 2 fragments. The extinction coefficient of 1.45 (cm. mg/ml) -I was also used for the pH 2 treated IgG materials since r~o significant changes occurred in absorbance for native and pH 2 treated MAbs at a concentration of 0.5 mg/ml. The molecular mass values of hCG, lgG and F(ab') z fragments were assumed to be 38 kDa (Moyle et al., 1983), 150 kDa and 100 kDa, respectively.

Materials and methods

pH 2 treatment of lgG Materials The mouse monoclonal antibodies used in this study were directed against hCG and of isotype IgG 1. The MAbs with different isoelectric points (OT-9A: 6.4-7.2; 7B, 5.9-6.6; 1C, 6.2-6.9; 4D, 6.5-7.0; 4E, 7.0-7.4) were arbitrarily chosen from a large panel of anti-hCG MAbs and the production in serum-free medium using hollow fibre bioreactors has been described elsewhere (Sch6nherr et al., 1987). The MAbs were purified by either sodium sulphate precipitation (20%, w/v) followed by Sephacryl S-200 HR gel filtration or protein A affinity chromatography (Van Erp, 1991, Van Erp et al., 1991a). Highly purified hCG (12,100 IU/mg)was obtained from Diosynth (Oss, Netherlands). The (hydrophobic) negatively charged and red colored polystyrene latex was supplied by AKZO-CR (Arnhem, Netherlands). The diameter of the latex was 797 nm as determined by quasi-elastic light scattering and the latex concentrations were determined using the dry weight method. The affinity purified goat antibodies specific for Fc and F(ab') 2 of mouse IgG, and (BSA) Boseral R were obtained from

Purified MAb solutions were diluted with 0.05 M glycine/HCI buffer pH 2.0 to a final concentration of 0.5 mg/ml. After incubation of the MAb solution at 0-4°C for 1 h the pH was adjusted to 7.0 by the addition of 0.1 M NaOH.

Adsorption of MAbs onto the latex particles The MAbs were physically adsorbed onto the polystyrene latex particles by a modification of the method of Fritz and Rivers (1972). A constant amount of native or pH 2 treated lgG (0.33 mg in 1.6 ml of a 5 mM sodium phosphate buffer pH 7.0 containing 0.1 g/I Thiomersal) was added dropwise and under continuous stirring to a polycarbonate tube containing 1.6 ml latex dispersion in the same phosphate buffer representing a surface area of 0.15 m 2 (1.33% w/w). The samples were incubated for 1 h at 30°C under continuous stirring, and were subsequently centrifuged for 15 min (104 N/kg). A sample was taken from the supernatants to determine the MAb concentration, while the sediment was resuspended in the 5 mM sodium phosphate buffer supplemented with 2 g/l BSA and incubated for another 30 min at

193 30°C. After a second centrifugation step, the pellet was resuspended in 5 mM phosphate buffer pH 8.0 containing 0.1 g / I Thiomersal and I g / I BSA to a latex concentration of (I.85% (w/w). The concentration of free lgG in the supernatants after adsorption (coating) was determined by sol particle immunoassay (SPIA) for mouse lgG (Leuvering et al., 1983, Van Erp et al., 1991b). The amount of MAb/latex surface ( m g / m 2) was calculated using these values and the original MAb concentrations.

Quasi-elastic light scattering Quasi-elastic light scattering has been performed using a Coulter Nanosizer (Coulter Electronics, Mijdrecht, Netherlands) to determine the hydrodynamic diameter of colloidal latex particles. Before measuring, the samples were diluted by a factor of 200 with 5 mM sodium phosphate buffer, pH 8.0.

Preparation of F(ab'), fragments The F(ab') 2 fragments of IgG were prepared as described elsewhcre (Van Erp, 1991; Van Erp et al., 1991a).

Coating of microtitration plates Goat anti-mouse lgG was dis~lved in the 'coat buffer' (5 mM NaH2PO 4, 0.14 M NaCI, 10 g / ! PEG 8(H)tl and 60 g / I saccharose, pH 7.4), at a concentration of 5.4 p,g/ml. 135 p,! of this .solution were pipetted into each well of a microtitration plate (Greincr, Aiphen a / d Rijn, Netherlands) and incubated overnight at 4°C. After aspiration, the wells were incubated for 2 h at room temperature with coat buffer (3(~} p.I/well) supplemented with 0.1% (w/v) BSA. Sub~quently, the wells were washed twice with coat buffer. Using this procedure both typc I, goat anti-(mouse l g G / F c ) , and type II goat anti-(mouse IgG/F(ab')2), coated microtitration plates were prepared.

Indirect determination of lgG orientation High performance size exclusion chromatography HPSEC was performed on a Zorbax GF-250 column (Dupont Nemours, Den Bosch, Netherlands) using a 0.2 M N a H 2 P O 4 / K 2 H P O 4 buffer pH 7.0 and UV monitoring at 206 nm as has been described previously (Van Erp, 1991; Van Erp et al., 1991b).

Purification of lgG aggregates by gel filtration The pH of the acidified MAb solutions (pH 2.0) was adjusted to pH 7.0 (total volume: 15 ml). These solutions were applied to a Sephacryl S-300 HR column (bed volume, 480 ml) which was equilibrated in 0.05 M sodium phosphate buffer pH 7.0 containing 0.15 M NaCI and 0.1 g / I sodium azide. The flow rate used was 0.8 m l / m i n and the effluent was monitored at 280 nm. The fractions (3 ml) containing aggregates of lgG were pooled and concentrated using Centricon 30 microconcentrators (Amicon, Danvers, MA, USA). Subsequently, a buffer change of the IgG solutions was performed on a PD l0 column (Pharmacia, Uppsala, Sweden) with 5 mM sodium phosphate buffer pH 7.0 containing 0.1 g / I Thiomersal. The total lgG recovery was > 90% as calculated from the absorbance at 28(1 nm.

A dilution series of a MAb-latex conjugate ({).85% w / w ) was prepared in (}.05 M glycine buffer pH 8.5. Samples (100 p,I) of the dilution series were pipetted (four replicates) into the wells of type I and II coated microtitration plates. The plates were sealed and incubated for 2 h at 37°C. Subsequently, the wells were aspirated followed by four washes wi,il 300 p,I PBS-Tween (20 mM phosphate buffer pH 7.2 containing 0.13 M NaCI and (I.05% (v/v) Tween 20). Finally, 110 p,I of 0.1 M NaOH were addcd to the wells in order to dissociate the immune complexes and redisperse the bound anti-hCG latex conjugate. After 15 rain the absorbancc of ~ach well was measured at 540 nm. Wells incubated with a BSA-latex conjugate were taken as blanks'.

ELISA for mouse IgG Dilution series of both native and pH 2 treated MAb solutions were prepared in PBS buffer. Samples (100 /~l) were pipetted (four replicates) into the wells of type l and type 11 coated microtitration plates and incubated at RT (18-24°C)for 30 rain. The microtitration plates were then washed three times with PBS-Tween (300 p.I/well) and incubated at RT for 30 rain with 100 ~.l/well sheep anti-mouse IgG-HRP conju-


gate, 1/31100 diluted in PBS-Tween. The excess of enzyme-labelled antibody was removed by another three washes with PBS-Tween and subsequently IIX)/.tl of the substrate solution (0.45 mM 3,3',5,5'-tetramethylbcnzidine, !.5 mM H.,O_, and II.l M sodium acetate pH 5.5) were added. After color development (30 min), the reaction was terminated by adding 100 p.I of 2 M H_,SO4. The absorbance of each well was measured at 450 nm.

Determhlation of antibody affinity The affinity constants (K,,) of the MAbs in solution were determined as previously described (Van Erp, 1991; Van Erp et al., 1991c). After incubation of a constant amount of MAb (4.5 nM, both native and pH 2 treated lgG) with various concentrations of hCG, the free and MAb bound hCG were separated with the aid of an excess of rabbit anti-mouse lgG coupled to cellulose particles (DASP). The free hCG was subsequently determined using a SPIA agglutination procedure for hCG. The determination of the (apparent) K~, values of the immobilized MAbs was essentially similar. After two washes of the latex conjugates with assay buffer, a constant surface area of 2.25 cm z was incubated with various concentrations of hCG for 4 h at RT. Bound and free hCG were subsequently separated by centrifugation for 10 min at 10 "~ N/kg. The free hCG concentration in the supernatant was determined using the SPIA agglutination procedure for hCG. Experiments with both normal mouse IgG and BSA coated latices showed that non-specific adsorption of hCG to the latex conjugates was negligible under the assay conditions used. All experiments were performed at least in duplicate.

Calculation of binding parameters The binding of antigen (hCG) to immobilized MAb does not necessarily reflect the corresponding interaction between Ag and MAb in solution (Stenberg and Nygren, 1988). It has been assumed that the M A b / A g interaction at a solidliquid interface is governed by the law of mass action (no mass transport limitations) and reaches equilibrium. In this study, the binding of antigen to immobilized MAbs is quantitatively expressed by an apparent affinity constant.

The (apparent) K,, values for anti-hCG MAbs and hCG were obtained using the Scatchard and Sips equations (Steward, 1986; Van Erp et al., 1991c). For homogeneous aqueous solutions, r is defined as mol A g / m o l MAb. In the case of heterogeneous systems (latex bound MAb) r is calculated as mol Ag bound per mol of effective MAb. The effective number of MAb is taken to be equivalent to the maximum amount of bound Ag, since hCG is a monovalent Ag for each particular MAb used.


Monitoring of the pH 2 pretreatment of different MAbs HPSEC analyses The technique of HPSEC showed that pH 2 treatment of anti-hCG MAbs yielded a considerable amount of aggregates immediately after adjustment of the pH to 7.0. The major portion of these aggregates consisted of di- and trimeric lgG. The level of aggregation, however, was different for each particular MAb and varied between 20-42% of total peak area. In addition, storage of the treated MAbs at 4-8°C for 24 h showed no decrease in the extent of aggregation indicating that the aggregation was not reversible upon prolonged incubation in solution (pH 7.0) at 4-8°C (data not shown).

Affinity determhzations The effect of pH 2 treatment on MAb affinity in solution was determined by comparing the affinity constants (K,,) of native and pH 2 treated MAbs. The results are presented in Table 1 and show that the acid pretreated MAbs retained their affinity in comparison with the native MAbs. Furthermore, comparison of the average number of binding sites per MAb molecule in solution ( h C G / M A b ratio) revealed that there was no (or only a slight) decrease in the number of active binding sites after pH 2 treatment.

Equilibrium binding studies of immobilized MAbs Affinity The main advantage of a latex particle bound antibody is that the mobile particle facilitates rapid reaction with the antigen and it was found

that an incubation time of 3 h was sufficient for the system to reach (at least an apparent) equilibrium. Furthermore, the latex conjugates remained eolloidally stable in the assay buffer during the 3 h incubation period as measured by quasi elastic light scattering. The measured apparent K~, values of four different, immobilized anti-hCG MAbs are summarized in Table It(A). As can be seen, the (apparent) K,, values of native and pH 2 pretreated MAbs were found to be similar after physical adsorption onto polystyrene latex particles. This indicates that no changes in apparent K~ values occurred after pH 2 prctreatment in comparison with the native MAbs. It must be noted, however, that the (apparent) K,, values for antigen binding to physically adsorbed MAbs (heterogeneous reaction, Table II(A)) were higher than those of the homogeneous reaction in solution (Table l) by a factor which varied slightly between the MAbs.

Antigen binding capacity No preferential adsorption of either native or pH 2 pretreated MAb solutions was observed (Table ll(A)). Nevertheless, the antigen binding capacities were considerably different (Fig. l). As summarized in Table ll(A), the highest binding capacities as well as maximum h C G / M A b ratios TABLE ! THE EFFECT OF pH 2 TREATMENT ON MAb AFFINITY IN SOLUTION MAb


pH 2 treat ment ~' + +


+ + +

Scatchardh K,, (I/nmol) 1.8 1.8 1.9 2.0 1.6 1.4 10.7 111.7 5.2 5.4

Sips b.c K;, (l/nmol) 1.8 1.9 2.(I 2.0 1.5 1.6 11).7 1{).5 5.3 5.5

Ratio d hCG/ MAb 2.1) 1.9 2. I 1.9 2.1 1.9 2.2 2.(1 1.9 1.9

a Native ( - ) and pH 2 treated ( + ) MAb. h The values are the means of duplicates. c The heterogeneity indices obtained from the Sips equation were not significantlydifferent from 1. d Mol Ag hound per mol of MAb.

were obtained for the pH 2 pretreated MAb fractions.

Contribution of the IgG aggregates to the enhanced antigen binding capacity As described above, pH 2 treatment leads to aggregation of the MAbs. These aggregates of lgG, however, may result in a better accessibility of the MAb binding sites by a decrease in sterie hindrance (spacer effect). In order to determine w h e t h e r the lgG aggregates a n d / o r the monomeric lgG, both obtained after pH 2 treatment, were responsible for the enhanced antigen binding capacity, the following adsorption experiments were performed with purified (Sephacryl S-3(10 H R gel filtration) aggregates and monomeric pH 2 treated IgG. Equal amounts (0.33 rag, based on A,~),m) of both purified monomeric lgG and aggregateswere added to the polystyrene latex surface (0.15 mE). The amount of a d ~ r b e d ' l g G ' was calculated from the SPIA agglutination procedure for mlgG. The results are summarized in Table II(B). For all MAbs tested, the amount of adsorbed monomeric lgG was slightly lower in c o m p a r i ~ n with the lgG aggregates. In spite of this lower IgG concentration per latex surface area, the monomeric pH 2 treated IgG latex conjugates had a much higher antigen binding capacity than the aggregates. These results suggest that the aggregates of lgG obtained after the acidic pretreatment do not contribute (significantly) to the enhanced antigen binding capacity of the latex conjugates by formation of a so-called spacer.

Immobilization of F(ab') 2 fragments In order to examine the effect of MAb structural integrity on the enhanced antigen binding capacity, native and pH 2 treated F(ab') 2 fragments of three anti-hCG MAbs were adsorbed to the polystyrene latex particles. The amount of adsorbed fragments was determined by the SPIA agglutination procedure for mouse IgG using native and pH 2 treated F(ab') z fragments for the respective standard curves. Table II(C) shows that equal amounts of both native and pH 2 treated fragments were adsorbed to the latex surface although less than found for the intact MAbs (Table Ii(A)).

Ith~ Moreover, it is remarkable that no significant differences in a p p a r e n t K , values and Ag binding capacities were observed between native and pH 2 treated F{ab') 2 fragments. This su[;gests that the Fc region is essential if an enh~.nced antigen binding capacity is to be obtained after pH 2 treatment, when the antibody is adsorbed to latex.

MAb orientation on the latex surface T h e increased antigen binding capacity after p H 2 p r e t r e a t m e n t might be due to better orientation of the lgG molecules on the latex surface (i.e., the Fab s e g m e n t s exhibit a p r e f e r r e d orientation towards the solvent). Therefore, the orientation of the M A b s after adsorption to latex was investigated by incubation of the latex conjugates

in microtitration plates previously coated with goat antibodies specific for either Fc or F ( a b ' ) 2 of m o u s e IgG. The results are shown in Fig. 2. It was found that the anti-Fc coated (type I) microtitration plates b o u n d m o r e latex particles w h e n coated with native M A b rather than p H 2 treated MAb, w h e r e a s the opposite was obtained for the a n t i - F ( a b ' ) 2 coated (type 11) plates. Moreover, no preferential binding of either native or acidic p r e t r e a t e d M A b to the coated microtitration plates (type I and II) was observed in experiments using M A b s in solution (both native and p H 2 treated) and subsequently SaM l g G H R P conjugate (Fig. 3). Similar results were obtained with Sam lgG coated mierotitration plates indicating that there was also no preferential binding of Sam IgG to native or p H 2 treated M A b s (data



pH 2 treatment "

Surface cone. (p.g/cm2)

K,, values I, (I/nmol)

Binding capacity {pmol/cm 2)

Ratio ¢ hCG/MAb



+ + + +

211 20 21) 21 21 21 2{) 21

4.3 4.6 3.4 3. I 26 30 12.4 13.4

().4 1.6 1).9 1.3 0.3 1.6 11.7 1.5

0.3 1.2 0.7 1.0 0.2 1.2 1).5 1.1


21 19 21 20 20 19 2{1 19

4.5 4.4 3.6 3.3 28 25 13.I 14.3

0.5 I.l 11.9 1.2 0.4 1.1 {).3 1.3

0.4 0.9 0.6 1.0 0.3 0.9 I).2 1.0

+ + +

8 8 12 13 II 11

4.3 4.9 3.3 3.6 26 33

0. I 0.1 0.4 0.4 0.5 0.5

0. I 1}.I 0.3 0.3 0.4 0.4



C F(ab' ),


Native MAb ( - ): pH 2 treated MAb ( + ) total fraction; aggregated (A) and monomeric (M) MAb fraction after pH 2 treatment. , The (apparent) K~ values are the means of duplicates (Scatchard equation). " The highest amount of I'xmndAg (tool) per total amount of immobilized MAb (tool).


A450 nm (pH 2 treated MAb)

Bound hCG (pmol/cm2) 2.0-






1.0 -


0.5" 0.4-






MAb Fig. I. The hCG binding capacities of native (open bars) and pH 2 treated (shaded bars) MAb latex conjugates. The data

representmean+ standarddeviation(error bars)of three repeatedexperiments.


0.0 r 0.0

i 0.2

l l I ' 0.4 0.6 0.8 1.0 A450 nm (native MAb)

I " 1.2

Fig. 3. Binding of native and pH 2 treated MAb (OT-4E) to

type I ( + ) and type II ([]) coatedmicrotitrationplates.The correlation coefficient for both type I and type II coated microtitratkm plates is 0.99. The line represents y = x.

540 nm

A540 nm







0.6= 0.4 0.40.2 0.2-








OT- 1C OT-4D MAb


Fig. 2. Binding of both native (open bars) and oH 2 treated (shaded bars) MAb latex conjugates ( I / 5 dilution) to microtitration plates previously coated with GaM Fc-specific (A) and F(ab') 2 specific (B) IgG fractions. The values are corrected for the blank (BSA coated latex conjugate) and represent mean + standard deviatkm (bars) of quadruplicate determinations.


not shown). These results suggest that there are more Fab segments oriented towards the solvent when pH 2 treated MAbs are adsorbed to latex.

Discussion The pH 2 treatment of MAbs prior to adsorption onto the polystyrene latex surface resulted in an increased antigen binding capacity of the final conjugate as compared to the native MAbs based conjugate. Four possible explanations for the increased binding capacity after pH 2 treatment arc: (!) an increase in antibody affinity due to conformational changes in the Ag binding site, (2) an increased solid phase adsorption of pH 2 treated MAbs, (3) the aggregates of MAbs obtained increase the number of accessible binding sites by a spacer effect, (4) changes in structural integrity of the MAb facilitate an improved orientation of the MAb molecule. The validity of each of these explanations is discussed successively. Binding studies of several anti-hCG MAbs, both in solution and physically adsorbed onto latex particles, demonstrated that no changes in K~ values occurred after a pH 2 pretreatment in comparison with the native MAbs. This indicates that the first explanation is unlikely. The (apparent) K~ values of the immobilized MAbs, however, were higher than of those in solution. The reason for this phenomenon is not yet clear but it is unlikely that the enhancement in K~b values is due to experimental differences since the same procedure for determining K~, was used for both the MAbs in solution and Mabs adsorbed onto polystyrene latex. On the other hand, the solid phase adsorption might give a favorable conformational change of the MAbs yielding increased K~ values. Another factor could be the increase in the local MAb concentration due to the immobilization; this effect is absent in the case of MAbs in solution (homogeneous system). Both options are currently under investigation. In contrast with the considerable loss of active binding sites of polyclonal antibodies in solution after low pH treatment (Lin et al., 1989), little or no decrease in total active binding sites was observed for the anti-hCG MAbs in solution. The acidic treatment of a MAb prior to its physical

adsorption gives a positive contribution to the total number of active binding sites per surface area (Ag binding capacity). Conradie et al. (1983) observed an increased solid phase adsorption of polyclonal antibodies after low pH treatment. In the present experiments, however, equal amounts of both native and pH 2 treated MAbs were adsorbed to the polystyrene latex particles indicating that the second explanation is less likely for the enhanced antigen binding capacities observed in this study. The occurrence of a higher desorption (leakage) of the native immobilized MAb as compared to the pH 2 treated MAb is also unlikely as determined by the SPIA agglutination procedure for mouse lgG. Monitoring of the pH 2 treatment by HPSEC showed that exposure to a low pH environment followed by adjustment to pH 7.0 caused MAb aggregation. Application of aggregated MAbs could increase the number of accessible binding sites by a spacer effect (decrease of steric hindrance by the solid phase). However, experiments with either aggregated or (pH modified) monomeric IgG fractions, both adsorbed onto the latex particles, demonstrated that the highest number of active binding sites were found for the monomeric pH 2 treated lgG latex conjugates. The lgG aggregates obtained after pH 2 treatment obviously did not. significantly enhance the antigen binding capacity of the physically adsorbed MAbs. Explanation 3 can therefore be ruled out. Infrared (Abaturov et al., 1969), circular dichroism (Conradie et al., 1983) and fluorescence spectroscopy studies (Lin et ai., 1989) have shown that lgG molecules in solution undergo structural changes upon lowering the pH. These changes mainly occur in the Fc region as indicated by infrared and circular dichroism measurements using Fab and Fc fragments. The lower conformational stability of the Fc fragment compared to that of the Fab fragments is confirmed by both the data on proteo[ytic hydrolysis (Tijssen, 1985) and the similar K~ values for native and pH 2 treated MAbs as described in this paper. Also the data obtained from physically adsorbed F(ab') 2 fragments clearly demonstrate the involvement of the Fc region in the increased

I~,r¢ Ag binding capacity u p o n p H 2 t r e a t m e n t of the M A b s prior to immobilization (Table iI(C)). Finally, the indirect information with respect to the orientation of the physically adsorbed M A b s suggested that the Fc region is more oriented towards the solvent for the native M A b coated latex particles, w h e r e a s the Fab s e g m e n t s are m o r e oriented towards the solvent for the p H 2 p r e t r e a t e d MAbs. In conclusion, all o f the observations in these studies s u p p o r t explanation 4, namely that the p H 2 t r e a t m e n t of a M A b partially modifies the Fc region of the M A b molecule, thereby increasing the local hydrophobicity of the lgG molecule ( V a n d e n b r a n d e n et al., 19811 and consequently facilitating the oriented adsorption of the modified m o n o m e r i c l g G fraction such that the Fab s e g m e n t s are m o r e oriented towards the solvent, favoring antigen binding capacity.

Acknowledgements D. White and L. v.d. Weyer are kindly acknowledged for their helpful discussions.

References Abaturov, L.V.. Nezlin, R.S., Vengerova. T.I. and Varshavsky. J.M. 11969) Conformational studies of immunoglohulin G and its subunits by the methods of hydrogen-deuterium exchange and infrared spectroscopy. Bioehim. Biophys. Acta 194, 386-396. Bernard, A.M., Foidart, J.-M., Mahieu, P., Viau, C. and Lauwerys, R.R. (1986) Detection of anti-laminin antibodies in sera by latex agglutination. Clin. Chem. 32, 146S1472. Conrad[e, J.D., Govender, M. and Visser, L. (1983) ELISA solid phase: partial denaturation of coating antibody yields a more efficient solid phase. J. Immunol. Methods 59, 289-299. Fritz, R.B. and Rivers, S.L. (1972} Hepatitis-associated antigen: detection by antibody sensitized latex particles. J. |mmunol. 108, 108-11 I. Ishikawa, E., Hamaguchi, Y. and Imagawa, M. (19801 An improved preparation of antibody coated polystyrenebeads for sandwich enzyme immunoassay. J. Immunoassay I, 385-398. Kakabakos, S.E., Livaniou, E., Evangelatos, G.P. and lthakissios, D.S. f1991D Immobilization of immunoglobulins onto surface-treated and untreated polystyrene beads for radioimmunoassays.Clin. Chem. 36, 492-496.

Leuvering, J.II.W., Govcrd¢, B.C., Thal, P.J.H.M, avd Sehuurs. A.II.W.M. (19831 A luum~gcneous .,~d purtielc immum~assay for human chorhmie gonadotrophin using mom)chmal ant[billies. J. Immunol. Methods 611,9-23. Lin, J.N., Andradc, J.D. and Chang, I,-N. (19891 Tim influence of ad.~)rpfion of native and nmdified antilmdics on their activity. J. Immunol. Metht~.Is 125, 67-'77. Moylc, W.R,, Andcr.'~m. D.M. and Ehrlieh, P. (1983) A circular antibody-antigen complex is responsible for increased affinity s h ~ n by mixtures of tmlm)chmal antibodies to human chorkmie gonadotrophin. J. Immunol. 131, Iq}01905. Parson. G.|I. (1981 } Antibody-coated plastic tubes in radkdmmunoassay. Methods Enzytrafl. 73, 224-239. Sch~inherr. O.T., Van Gelder, P.T.J.A., Van flees, P.J,. Van Os, A.M.J.M. and Rcmlofs,H.W.M, (19871 A hollmv fibre dialysis system for the in vitro production of n~moclonal antibodies replacing in vivo pnnluction in mice. Dev. Bkfl. Stand. 66. 211-220. Stenberg, M. and Nygren, If. (1988)Kinetics of antigen-antitardy reactions at ~did-liquid interfaces. J. Immunol.

Methods 113, 3-15. Steward, M.W. (1986) lntrtgluction to meth~e,ls used to study the affinity and kinetics of antilxgly-antigcn rcactkms. In: D.M. Weir, L.A. Herzenberg and C. Blaekweli (Eds.), Handbtmk of Experimental Immunology,4th ¢dn. Blackwell Scientific Publications, Oxfi)rd, pp. 25.1-25..'~J. Tijssen. P. (1985) The immobilizatkm of immunoreaetants on solid phases. In: Practice and theory of enzyme immunoassays: laboratory techniques in biochemistry and molecular

biology. Elsevier Science Publishers. Amsterdam, pp. 297328. Vandenbranden. M.. De Cimn..I.L., Jeener. R., Kanarek, L. and Ruysehaert, J.M. (19SI) Interactions of immunoglobulins with lipid mono- or bilayers and lip~.~mes. Existence of two conformations of immunoglobulins of different hydrophobicitics, Mol. Immunol. 18, 621-631. Van Erp. R. (1991) Mom)chmal antilx~dies in diagnostics. Monitoring of mom)chmal antibody charactcristics during (large scale} productkm, purification and applicatkm in diagnostic systems. Ph.D. Thesis, University of Nijmegen, Nijmegen. Netherlands. Van Erp, R. Gribnau. T.C.J.. Van Sommeren, A.P.G. and Bloemers, H.P.J. (1991a) Affinity of mom)clonal antilm-,dies: Interpretafion of the positive cooperative nature of anti-hCG/hCG interactions. J. lmmunol. Methods 140, 235-241. Van Erp, R. Adorf, M., Van Sommeren, A.P.G. Sch6nherr, O.T. and Gribnau, T.C.J. (1991b) Monitoring of the product[on of mom)ckmal antib~dies by bybridomas. Part I: long-term cultivatkm in hollow fibre bioreactors using serum-free medium. J. Bk)tcchnol. 211,235-248. Van Erp. R. Grihnau, T.C.J., Van Sommeren, A.P.G. and Bloemers. H.P.J. (1991c) Applicatkm of a sol particle immunoassay to the determination of affinity constants of mommhmal antibodies. J. Immure)assay12, 425-443.