Journal of Immunological Methods, 20 (1978) 365--383

365

© Elsevier/North-Holland Biomedical Press

AMPLIFICATION OF THE ENZYME-LINKED IMMUNOSORBENT ASSAY (ELISA) IN THE DETECTION OF CLASS-SPECIFIC ANTIBODIES

J.E. BUTLER, P.L. McGIVERN and P. SWANSON

Department of Microbiology, University of Iowa, Iowa City, IA 52242, U.S.A. (Received 26 July 1977, accepted 25 October 1977)

A modification of the standard enzyme-linked immunosorbent assay (ELISA) is described which circumvents the requirement for specifically purified antibodies from which antibody-enzyme complexes are made. The assay utilizes the principle of a soluble anti-alkaline phosphatase immune complex (AP-A-AP) and has been called the amplified ELISA. Methods for preparing and evidence for the specificity of rabbit anti-rat 7-FC, IgM (p) and IgA (~) are presented. These reagents are used to measure anti-DNP antibodies belonging to classes IgG, IgM and IgA in rat serum. Using antiglobulin and antienzyme reagents prepared in guinea pigs, anti-ovalbumin antibodies are measured in rabbit serum. Titration curves are similar when the amplified ELISA is compared to the standard ELISA. A change in slope suggesting an effect of saturation of antigen sites, occurs at the same input antibody concentration for both assays. Determination of the anti-DNP concentration of unknown sera by extrapolation from titration graphs of a known serum suggests that the value is overestimated, i.e., amplified when the amplified ELISA is used. In addition, the amplified ELISA has an improved ability to detect low levels of antibody. Evidence is presented which illustrates how the use of optimally conjugated DNP-proteins, age of conjugates, and optimal dilutions of secondary antiglobulins and the AP-A-AP reduce non-specific binding in the amplified ELISA. The amplified ELISA is capable of detecting 2.4 ng of antibody to ovalbumin in a one : one million dilution of rabbit serum with high reproducibility and low background.

INTRODUCTION An important aspect of our understanding of specific humoral immunity concerns the structure-function significance of antibody heterogeneity. The past decade encompasses a period during which the classes, subclasses, types and subgroups of immunoglobulins were identified and characterized. Examp l e s a r e a l r e a d y k n o w n in w h i c h t h e s e s t r u c t u r a l d i f f e r e n c e s a r e r e l a t e d t o p a r t i c u l a r b i o l o g i c a l a c t i v i t i e s ( I s h i z a k a a n d I s h i z a k a , 1 9 6 7 ; C a p r a e t al., 1972; Butler, 1974; Spiegelberg, 1974). One approach to understanding this structure-function relationship involves measuring the distribution of anti-

Research supported by grants from the National Institutes of Health numbers NO1 HR 5 3013 A and 5RO1 AI 11796.

366 b o d y activity among the various immunoglobulin classes and subclasses following antigenic stimulation. Three general methods are often employed to identify the class of a specific antibody. First, a number of inferential methods may be employed, i.e. the serum in question may be fractionated b y conventional methods of gel-filtration, ion-exchange chromatography, density gradient centrifugation or electrophoresis and the antibody activity shown to be associated with certain fractions which are rich in IgG, IgM, etc. Secondly, specific antibodies can be isolated by using antigen affinity columns or by dissociation of immune precipitates or microorganismantibody complexes. Such purified antibodies may then be subsequently identified by conventional techniques, i.e. single radial diffusion, radioimmunoassay, etc., which employ specific antiglobulin sera. Finally, there are a number of techniques which utilize specific antiglobulin reagents to identify the immunoglobulin class of the antibodies in situ. These include radioimmunoelectrophoresis (Yagi et al., 1963), the modified Coomb's assay (Coombs et al., 1965), double antibody 'Farr' assay (Minden and Farr, 1967) and more recently, the enzyme-linked immunosorbent assay (Engvall et al., 1971). The former two assays are at most qualitative while the latter two have been purported to be quantitative (Engvall and Perlmann, 1972). A major shortcoming of the ELISA is the requirement for specifically purified antibodies to be used in making enzyme-antibody complexes. If an investigator is content to work only with IgG class antibodies or to work in species where large quantities of IgA and IgM can be readily isolated in a pure state, this shortcoming is not serious. The use of the ammonium sulfate precipitated fraction of antisera, rather than specifically purified antibodies, for preparing antibody-enzyme conjugates has also been described. Unfortunately, experience in our laboratory (Sloan, 1975) has indicated that this is only useful when a significant portion of the globulin fraction of a specific antiserum (such as anti-IgG or anti-7) is in fact, specific antibody to the immunoglobulin in question. Such a finding would also be theoretically predictable. We describe here and evaluate a modification of the ELISA that: (i) eliminates the requirement for specifically purified antibodies and (ii) amplifies the sensitivity of the original ELISA. Our modification is modeled after the indirect peroxidase immunohistochemical assay described b y Sternberger et al. (1970) b u t utilizes antibodies to alkaline-phosphatase. The 'Amplified ELISA' was developed because of the difficulty we encountered in the specific purification of rabbit anti-rat ~-and ~-chaln antibodies. Data is presented on the use of the amplified ELISA to measure rabbit antibodies to ovalbumin. Data is presented on the'titration of the various antiglobulin, bridging and antibody-enzyme reagents employed. When used at o p t i m u m conditions, the amplified ELISA can detect less than 2.4 ng of antibody. We also report on comparisons made between the standard ELISA (that originally described by Engvall et al.) and the amplified ELISA for measuring IgG anti-dinitrophenyl (DNP) antibodies in the rat and also on the applica-

367 tion of the amplified ELISA for measuring IgM and IgA antibodies to DNP in rats. MATERIALS AND METHODS

Precipitin assays Immunoelectrophoresis (IEP) was performed in 2% Noble agar according to the method of Scheidegger (1955). Polyethylene-glycol immunodiffusion (PEG-ID) was performed according to Harrington et al. (1971). Single radial diffusion (SRD) and radioimmunodiffusion (RID) were carried out according to the methods of Mancini et al. (1963) and Yagi et al. (1962) respectively. 12SI-DNP-HSA was iodinated by methods previously used in our laboratory (Cambier and Butler, 1976). Turbidimetric precipitin assays were conducted as described by Leone (1971). Equivalence determinations were based on PEG-ID analyses of precipitin supernatants. Rat antibodies to DNP were quantitated using a double precipitin assay as previously described {Butler et al., in press). This involves incubation of antiDNP sera with dinitrophenylated human serum albumin (DNP-HSA) followed by the addition of bovine antiserum to HSA. The precipitate which forms after incubation is washed with cold TBS and resuspended in excess DNP-lysine. Rat IgG is then determined by SRD in agar which contains an excess of DNP-lysine to maintain antibody dissociation. Purified rat IgG used as a standard was measured spectrophotometrically, ~28Onm = 1.36. Previ"~ lcm,0.1% ous results have shown (Butler et al., in press) that the mean of four doubleprecipitin determinations was 0.717 mg/ml compared to 0.82 mg/ml determined by quantitative precipitation. Rabbit antibodies to ovalbumin were measured by quantitative precipitation in which the washed equivalence precipitate was measured using the microbiuret assay (Zamenof, 1957).

Chromatography Affinity chromatography was performed according to the methods of Cuatrecasas and Anfinsen (1971) using epichlorhydrinated Sepharose 4B (Porath et al., 1971) as the support medium. DE-52 cellulose (Whatman) and Sephadex CM-50 (Pharmacia) chromatography were used according to standard methods. Gel filtration on Sepharose 6B (reduced with alkaline NaBH4 and stabilized with epichlorihydrin (Porath et al., 1971)) was performed in a buffer containing 1.0 M NaC1, 0.01 N Tris--HC1, pH 8.6 with 0.003 M EDTA and 0.02% sodium azide added to prevent microbial growth.

Enzyme-linked immunosorbent assay (ELISA) The ELISA was conducted according to, and subsequently modified from, the method of Engvall and Perlmann (1972). The term 'Standard ELISA', as

368 used in this work, refers to the Engvall and Perlmann procedure of adsorbing primary antigen (DNP-histone) to polystyrene tubes at a fixed concentration followed by the addition of immune serum, the appropriate anti-immunoglobulin enzyme conjugate and finally the substrate. Antibodies to rat 7-Fc for the standard ELISA were purified from an antiserum to rat 7-Fc by applying it to an IgG1, IgG2ab affinity column and eluting the bound antibodies with 0.5 M acetic acid. Antibodies purified in this manner were conjugated to alkaline phosphatase by the method of Engvall and Perlmann (1972). Attempts to obtain specifically purified rabbit anti-rat IgM and IgA using a similar m e t h o d proved to be impractical. The 'Amplified ELISA' requires additional steps for the addition of the antibody-enzyme complex and the bridging antibody. The amplified ELISA is diagrammed in fig. 2 and is described in more detail later. Two separate systems are described. One system utilizes rabbit antiglobulin and anti-enzyme reagents and a goat anti-rabbit bridging antibody. The second system uses guinea pig reagents and a rabbit anti-guinea pig IgG as a bridging antibody.

Purification of rat immunoglobulin Secretory IgA was prepared from rat saliva by fractionation on DEAECellulose (Whatman, DE-52) followed by further purification on Sepharose 6B. Antiserum to this IgA was raised in a goat and rendered a-chain specific by absorption on an IgG affinity column. The specificity of this antiserum for IgA was verified using purified rat IgA kindly provided by Dr. Bazin, University of Louvain, Brussels (Cambier, 1975). Rat serum IgA was prepared by using an antibody-affinity column prepared from goat antibodies to rat IgA. The affinity fraction (bound protein) obtained after addition of whole rat serum to such a column was tested and shown to be largely IgA. IgA prepared in this manner was used to raise a rabbit antiserum to rat a-chains so that serum as well as secretory IgA anti-DNP antibodies could be detected without bias. Rat serum IgA was also purified by ion-exchange and gel filtration chromatography from a rat serum provided by Dr. Bazin which contained a large quantity of a myeloma-like IgA. Rat serum IgM was purified from the euglobulin fraction of pooled rat serum by gel-filtration on Sepharose 6B. The IgM-rich fraction (second major peak) was further separated from contaminating a2 macroglobulin (a2 M) by electrophoresis in Pevikon 870. The latter was present in the gel segment closest to the anode buffer while IgM remained near the origin. The a2 M obtained during IgM purification was used for antiserum absorption, specificity testing by PEG-ID and for the production of rabbit anti-a2M. Rat serum IgG, containing IgG1 and IgG2ab, was purified from the 33% a m m o n i u m sulfate insoluble fraction of rat serum prepared as described for rabbit IgG (see later) followed by fractionation on DEAE Sephadex A-50. The protein t h a t eluted at 0.02 M NaC1 in 0.01 M Tris--HC1 buffer, pH 8.6 was further purified on tandem 2.5 × 100 cm columns of Sepharose 6B. The

369 IgG prepared in this manner was digested for 16 h with cysteine activated papain (47 : 1 P : E ratio) and the digest chromatographed on Sephadex G-150. The 40,000--60,000 mol. wt. fraction was then fractionated on a Sephadex CM-50 column using a starting buffer of 0.01 PO4, pH 7.5 and a final buffer of 0.01 PO4, pH 7.5 containing 0.1 M NaC1. The protein eluted in t w o major peaks which were evaluated b y immunoelectrophoresis. The papain Fc and Fab fragments were identified by their analogous electrophoretic behavior to those of other species, e.g. bovine (Blakeslee et al., 1971) and the Fc was used to immunize rabbits. The resulting antiserum was rendered specific by absorption with a ~'-Fab affinity ~column.

Purification of rabbit and guinea pig IgG Rabbit IgG was prepared from the 33% ammonium sulfate insoluble fraction of rabbit serum. A 10 ml sample (162 mg/ml protein) was fractionated on a DE-52 cellulose column equilibrated with 0.01 M phosphate buffer, pH 8.0 and eluted using a linear gradient to 0.3 M phosphate, pH 8.0. Fractions containing only IgG as determined by PEG-ID and IEP were pooled and concentrated by ultrafiltration. Guinea pig IgG was prepared in a similar manner to rabbit IgG. Protein was eluted from a DE-52 column with steps of 0.01 M, 0.05 M, and 0.1 M phosphate buffer, pH 8.0. IgG containing fractions eluted with 0.01 M phosphate buffer, pH 8.0.

Preparation of primary antisera Rat anti-DNP antibodies were obtained from rats immunized intraperitoneally with alum precipitated DNP-BGG and B. pertussis and subsequently challenged i.v. with DNP-BGG in TBS. 'Early' antisera were obtained 10 and 20 days after primary immunization (no challenge) and referred to as pools A and B (table 1). Other pools (table 1) were obtained after challenge and differed in antibody content and affinity (Butler et al., in press). Rabbit antiserum to ovalbumin was prepared as described b y Richerson (1974).

Preparation of antiglobulins and other antisera Antiserum to human serum albumin (HSA) was raised in cattle as previously described (Butler et al., in press; Sloan, 1975). Rabbit antisera to guinea pig IgG, rat immunoglobulins, ~/-Fc fragments and a2M were raised in New Zealand rabbits by footpad immunization of 1--5 mg of the respective protein emulsified in complete Freund's adjuvant. Animals were challenged thirty days later by i.v. or intradermal immunization. Goat antiserum to rabbit IgG and rat SIgA were prepared in ewes using several subcutaneous and intramuscular injections consisting of a total of 5 mg of purified rabbit IgG or rat SIgA emulsified in complete Freund's

50 O

TABLE 1 R e p r o d u c i b i l i t y and sensitivity o f amplified E L I S A for t h e m e a s u r e m e n t o f r a b b i t a n t i - o v a l b u m i n . Replicate determination

Primary antibody

1:30,000 1:50,000 1:80,000 1:100,000 1:500,000 1:1,000,000

Xdiff -

Anti-OA NRS Anti-OA NRS Anti-OA NRS Anti-OA NRS Anti-OA NRS Anti-OA NRS

-

1

2

3

4

5

X

S.D.

C.V.

XNRS *

0.547 0.007 0.370 0.007 0.339 0.019 0.245 0.014 0.071 0.007 0.040 0.003

0.522 0.007 0.370 0.002 0.329 0.003 0.243 0.015 0.082 0.003 0.040 0.001

0.497 0.017 0.383 0.019 0.278 0.025 0.244 0.010 0.075 0 0.037 0

0.505 0.006 0.385 0.007 0.304 0.019 0.264 0.007 0.069 0.001 0.040 0.001

0.491 0.009 0.387 0.014 0.300 0.007 0.249 0.023 0.068 0 0.034 0.007

0.512 0.009 0.379 0.010 0.310 0.015 0.249 0.014 0.073 0.002 0.038 0.002

0.02 0.01 0.01 0.01 0.02 0.01 0.01 0.01 0.01 0.003 0.003 0.003

3.91

5588

2.64

3690

6.45

1966

4.02

1678

13.70

3550

7.89

1800

* N R S refers to n o r m a l r a b b i t serum.

100

371 adjuvant followed b y biweekly injections of IgG and incomplete adjuvant. Antiserum was prepared from blood obtained 10 days later. Rabbit and guinea pig antisera to alkaline phosphatase were prepared b y footpad (rabbit) or intradermal (guinea pig) immunization using 0.4--0.5 mg of alkaline phosphatase (Sigma, calf intestinal mucosa, t y p e VII) emulsified in complete Freund's adjuvant. Animals were boosted intradermally five weeks later with 0.5 mg alkaline phosphatase in saline and bled t w o weeks after boosting. Antiserum to rabbit IgG was also raised in 500 g outbred Hartley strain guinea pigs by multiple subcutaneous injections of 0.5 mg of rabbit IgG emulsified in complete Freund's adjuvant followed b y multiple intradermal injections of 0.5 mg of protein without adjuvant at two and three weeks. Animals were bled b y cardiac puncture two weeks later, and serum was collected.

Preparation of soluble antibody-enzyme complexes The turbidimetric precipitin test was used to estimate equivalence between guinea pig or rabbit anti-alkaline phosphatase and the enzyme. The equivalence precipitate was then washed three times in TBS and solubilized by adding a 10-fold mg excess of the enzyme. The soluble enzyme-antibody complex was stored at --20 ° C. The complex was designated as AP-A-AP after the PAP system of Sternberger et al. (1970).

Preparation of DNP-conjugates All DNP conjugates were prepared by standard methods (Little and Eisen, 1967) using 2,4-dinitrobenzenesulfonate recrystallized from hot ethanol. The degree of conjugation for the DNP-HSA and DNP-BGG was determined spectrophotometrically and ranged from 1.4 to 40 moles/mole. The insolubility and u n k n o w n extinction of calf thymus histone precluded its determination by the m e t h o d employed. RESULTS The specificity of the class-specific anti-rat antisera was evaluated b y PEGID and the results described below, and presented in fig. 1. Rabbit antiserum to rat IgM after absorption with an IgG1, IgG2ab affinity column, precipitated a line with purified IgM and this formed a line of identity with whole rat serum (fig. 1B). The antiserum did not precipitate with IgA or ~2M (fig. 1B, C) although a faint unidentified precipitin line formed with rat serum. When this antiserum was applied to an IgM-containing affinity column, the a m o u n t of specifically purified antibody recovered was extremely small. This was conjugated to alkaline phosphatase as described by Engvall et al. (1971).

372

B +

C +

X

X

~,:

X

iA

4

4 F

J

1

3

1 G

3

Fig. 1. I m m u n o d i f f u s i o n a n d i m m u n o e l e c t r o p h o r e t i c analyses of a n t i s e r a used for the E L I S A a n d a m p l i f i e d E L I S A . A = R a t s e r u m IgA p u r i f i e d f r o m m y e l o m a s e r u m . M = R a t s e r u m IgM p u r i f i e d as d e s c r i b e d in t e x t . S = PO 4 b u f f e r e d saline, pH 7.2. W = I n t e s t i n a l washings f r o m rats. X = W h o l e rat serum. Y = Fc-rich pool e l u t e d in 0.01 M PO4, pH 7.5 f r o m C M - S e p h a d e x . Z = F a b - r i c h pool e l u t e d at 0.1 M NaC1, 0.01 M PO4, pH 7.5 f r o m C M - S e p h a d e x . T h e following n u m e r i c a l d e s i g n a t i o n s are for a n t i s e r a p r e p a r a t i o n illust r a t e d in Fig. 1. 1 = R a b b i t a n t i - w h o l e rat s e r u m , 2 = R a b b i t a n t i - r a t IgM a b s o r b e d with IgG, 3 = G o a t a n t i - r a t IgG, 4 = R a b b i t a n t i - r a t aEM, 5a = R a b b i t anti-rat IgA a b s o r b e d w i t h IgG, 5b = S a m e as 5a b u t also a b s o r b e d w i t h a2M. 5 pi - R a b b i t a n t i - r a t IgA 5a e l u t e d f r o m a n IgA a f f i n i t y c o l u m n w i t h 0.5 M acetic acid, 5 pii - same as 5 pi b u t s e c o n d trial, 5u = R a b b i t a n t i - r a t IgA t h a t was n o t a b s o r b e d to t h e I g A - a f f i n i t y c o l u m n b u t e l u t e d in first p e a k , 6 = R a b b i t a n t i - r a t IgG - F c ( I ) , 7 = S a m e as 6 b u t a b s o r b e d o n an F a b a f f i n i t y c o l u m n , 8 = R a b b i t a n t i - r a t IgG-Fc (II).

After rabbit anti-rat IgA had been absorbed with the IgG1, IgG2ab affinity column, it still precipitated IgA and a2M (fig. 1C well 5a). Absorption on an a2M affinity column rendered this antiserum specific for IgA (Fig. 1B and 1C well 5b). Prior to absorption with a2M the anti-IgA was applied to an

373 affinity column containing purified monoclonal m y e l o m a serum IgA. Unfortunately only anti-a2M and soluble IgA were recovered in the unabsorbed peak (fig. 1A-5u) as well as in the acid-eluted fraction (fig. 1A-5pi). These results suggested that soluble antigen-antibody complexes were eluting from the column. Rabbit anti-~2M contained antibodies to t w o distinct a=M molecules of similar electrophoretic mobility as shown in PEG-ID (fig. 1C) and IEP (fig. 1F). This antiserum proved useful in testing the specificity of other preparations and antisera. Rat IgG Fc and Fab fragments separated on CM Sephadex were tested by IEP. Studies showed that Fc pools had trace contamination with Fab (fig. 1G), and sera from rabbits immunized with these Fc pools contained antibodies to Fc and Fab (fig. 1D, E). A n t i b o d y to Fab was removed b y affinity chromatography with an Fab immunosorbent column (fig. 1E, antiserum 7). This monospecific anti-7-Fc was then specifically purified b y affinity chromatography as described in Materials and Methods and conjugated to alkaline phosphatase. This anti-7-Fc-enzyme conjugate (antiserum 7, fig. 1E) was tested for specificity in the standard ELISA as follows. The IgG from the high titer rat anti-DNP serum was purified on an affinity column prepared from goat anti-rat IgG. Using this purified IgG antibody fraction at

THE

AMPLIFIED

ELISA

Anb - DNP SecPetlON

DN#-HIIST NE ~ ¢Y // ~1 T ]X#~) ~ |~1, U I~-, ~t~.

WASH

v

Measure O°Do400nm

A

r . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

#

I pr'epar'atlon of AP_A~AP 1~', \\'2"PL~ ~/~ . , ~ P ~ 1I Excess Alkaline

I

:

.

.

Ant,-Rat U [ ~ DN~ ~ ~ B I~Dk/p~ f . '-A E ~ ~1 m 'Y

.

.

.

.

.

.

.

.

.

.

.

.

[1

.

.

.

.

.

.

.

.

.

.

.

.

.

.

I % KJ-~AI~Ic-J7 ,'t~

]TUBE WALL

]

]r L-I p . Nitr,ophenyl P h o s p h ~ e rT II .........

-ID --""~AP A AP Goat A n t Added ~ ' ~ T Rabbit IgG h ~%r-~'

WASH

~L~-"~@

IU B

L

Fig. 2. Diagrammatic illustration to explain the steps in the amplified ELISA. Solid double-Y models represent rat secretory IgA, anti-DNP. Dotted Y models are rat IgG antiDNP. Pointed unshaded Y models represent rabbit anti-rat antiglobulin. Ruffled Y-models represent bridging antiglobulin against IgG. AP = alkaline phosphatase.

374

dilutions o f 1 / 5 0 0 , 1 / 7 5 0 , 1 / 1 0 0 0 and 1 / 1 5 0 0 , O.D. 4°° values o f 0.593, 0 . 3 2 8 , 0 . 2 3 0 and 0 . 1 5 8 respectively were o b t a i n e d a f t e r 120 rain using DNP-HSA c o a t e d t u b e s while the rabbit anti I g M - e n z y m e c o n j u g a t e did n o t give an O.D. 4°° value above b a c k g r o u n d (0.057). Before beginning r o u t i n e studies o f DNP-antibodies in rats using the E L I S A , pools I, II and III o f rat anti-DNP sera were o b t a i n e d , t h e i r c o n t e n t o f anti-DNP m e a s u r e d b y the d o u b l e precipitin assay and results expressed in table 2. Residual a n t i b o d y activity in the resulting s u p e r n a t a n t s after c o p r e c i p i t a t i o n in the q u a n t i t a t i v e precipitin assay was d e t e r m i n e d b y an E L I S A . When the residual a n t i b o d y activity in the s u p e r n a t a n t s was compared w i t h t h e starting material, it was f o u n d t h a t o n l y 2--3% r e m a i n e d . R a d i o i m m u n o d i f f u s i o n (RID) was t h e n e m p l o y e d t o qualitatively d e t e r m i n e the presence o f IgM, IgG and IgA a n t i b o d i e s in these pools. Pool I c o n t a i n e d only IgG b y R I D while P o o l II and III c o n t a i n e d IgG and IgM antibodies and possibly s o m e IgA antibodies. When a t t e m p t s were m a d e using the s t a n d a r d E L I S A t o o b t a i n useful data o n IgM anti-DNP a n t i b o d i e s in the above described sera t h e values were erratic and q u e s t i o n a b l y above b a c k g r o u n d . This u n d o u b t e d l y r e f l e c t e d the q u a l i t y o f the anti-IgM conjugate. T h e laborious m e t h o d o l o g y r e q u i r e d to: (i) p r o d u c e purified IgM f o r affinity c o l u m n s , (ii) to o b t a i n r e a s o n a b l e quantities o f purified anti-IgM antibodies f r o m such c o l u m n s and (iii) the difficulty in o b t a i n i n g purified r a b b i t anti-rat IgA t o g e t h e r with the weak b e h a v i o r o f the anti-IgM c o n j u g a t e p r o m p t e d the d e v e l o p m e n t o f the amplified ELISA. A l t h o u g h the amplified E L I S A was initially d e v e l o p e d to s t u d y rat antiDNP, we chose a p r o t e i n antigen (ovalbumin) t o establish m a n y o f the p a r a m e t e r s interrelating the c o n c e n t r a t i o n o f the d i f f e r e n t antiglobulin reagents and t h e soluble AP-A-AP c o m p l e x . R a b b i t serum c o n t a i n i n g precipTABLE 2 Antibody quantitation using ELISA and amplified ELISA. Quantity of IgG, anti-DNP antibody. Anti-DNP serum pools

Antibody affinity * (K50)

Double-precipitin assay (pg/ml) **

Standard ELISA (pg/ml)

Amplified ELISA (pg/ml)

Pool I *** Pool II *** Serum A SerumB Pool III

3.8 1.72 2.56 2.9 1.09

550 730 275 345 3250

Ref. Std. Ref. Std. trace 3.85 3575

Ref. Std. Ref. Std. 45 102 7051

× 107 × 107 x 106 X 106 × l0 s

* Association constants determined by modified Farr assay (From Butler et al., in press). ** Mean of two determinations made at six-month intervals. *** Reference pools against which other sera were compared by double-precipitin and ELISA assays.

375 itating antibodies to ovalbumin was measured by a quantitative precipitin assay as described in Materials and Methods and it was calculated that the rabbit antiserum contained 2.4 mg anti-ovalbumin per ml. Optimal conditions for the various antiglobulin reagents, etc. were established empirically (fig. 3) and eventually the following protocol was used: Ovalbumin (OA) incubation = 3 h at 37°C, (0.5 pg OA/tube); 1/20,000 to 1/1,000,000 dilutions of rabbit anti-OA = overnight at room temperature (RT); guinea pig anti-rabbit IgG = overnight at RT at 1/1000 dilution; rabbit anti-guinea pig IgG (bridging antiserum = 5--6 h at RT at a 1/500 dilution; guinea pig APA-AP = overnight at RT at a 1/1000 dilution; and finally the substrate reaction determined empirically on the basis of color intensity. Fig. 2, although illustrated for the rat anti-DNP system, shows the steps involved in the amplified ELISA. Fig. 3A illustrates the titration of the primary antiserum (rabbit anti-OA) using the o p t i m u m conditions described above as well as companion data obtained with the corresponding dilutions of serum from an unimmunized rabbit. The proportions described above and for fig. 3A were established by titrating the various components of the system. The results of titration studies on the secondary antiserum (antiglobulin reagent) are shown in fig. 3B, the rabbit anti-guinea pig bridge antiserum in fig. 3C and the soluble a n t i b o d y enzyme complex in fig. 3D. The data in fig. 3A--D also illustrate the difference in background O.D. 4°° obtained between one assay to the next. The logarithmic presentation of the data somewhat exaggerates the background effect as numerically the value is always less than 0.03 and often less than 0.010 when o p t i m u m conditions are employed. This is also seen from the data in table 1 which in addition shows the reproducibility of data points for the amplified ELISA. Because a rabbit anti-rat IgG-enzyme complex had been prepared for use in the standard ELISA and because the amplified ELISA had not been compared to other antibody assays, we chose to compare the standard and amplified ELISA in the measurement of rat IgG, anti-DNP antibodies. The amplified ELISA for the rat system was carried out using the antiglobulins described in fig. 1, the rat anti-DNP pools listed in table 2 and according to the protocol illustrated in fig. 2. Optimal conditions used for studying IgG anti-DNP antibodies are given below: Antigen adsorption = 2/Jg DNP-histone for 3 h at 37°C; primary antiserum = dilutions of rat serum pools I and II incubated 24 h at RT; secondary antiglobulin = rabbit anti-~,-Fc diluted 1 : 50 incubated overnight at RT; bridging antiserum = goat anti-rabbit IgG diluted 1 : 100 and incubated 24 h at RT; AP-A-AP = overnight at RT at 1 : 500 dilution and the substrate reaction time determined empirically on the basis of color intensity. The o p t i m u m antigen concentration was determined by experimentation using tubes to which 10, 5, 2, and 1 pg of DNP-histone had b'een added. Using a 1/1000 dilution of rat anti-DNP serum or normal rat serum, the following O.D. 4°° values for the respective dilutions were obtained after 16 min:

376

A

II

!0-

10

I 01-

/

01©

? O 001.

0 01 -

L.I

0!

1'0

10

01

R E C I P R O C A L of DILDT ION

I'O RECIPROCAL

X 10 5

X

I'O of D I L U T I O N 10 3

¢ 1.O-

01-

0 0 0 0

/

1.o-

0,1-

0.01 -

001

-

\ o.1

1.'o

/o

R E C I P R O C A L of D I L U T I O N X 10 -3

q

o.1

1:o REcIPRocAL

~b of D I L U T I O N

X 10 -3

Fig. 3. Titration o f amplified E L I S A used for rabbit anti-OA system. A = C o m p l e t e amplified E L I S A o b t a i n e d using o p t i m u m secondary antiserum, bridging antiserum and AP-AA P as determined in Fig. 3B, C and D. • = anti-OA prepared in the rabbit. • = Dilution of serum from control rabbit diluted as for experimental. B = Titration o f secondary antiserum, C = Titration o f bridging antiserum, D = Titration of AP-A-AP. Arrows indicate dilutions o f reagents that were selected for use in the c o m p l e t e amplified E L I S A in Fig. 3A. A 1 : 2 0 , 0 0 0 dilution of the primary rabbit anti-OA was used in the e x p e r i m e n t s s h o w n in Fig. 3B, C and D. E x c e p t for the reagent being titrated, the c o n c e n t r a t i o n s o f the other reagents were the same as those used in Fig. 3A. The color d e v e l o p m e n t times were A = 27 min, B = 16 rain, C = 18 min and D = 9 rain.

377

0.858/0.312, 0.493/0,080, 0.345/0.010 and 0.248/0.005. The values obtained using 2 g g DNP-histone per tube gave the greatest difference O.D. 4°° between anti-DNP serum {0.345) and normal rat serum (0.010). The effect of concentration of the secondary antiglobulin was compared using early, low affinity (Pool B) and late, higher affinity (Pool I) rat antiDNP sera (fig. 4). A comparison between the standard and amplified ELISAs in relationship to the input primary antibody dose was performed using rat serum pools II and III, rabbit anti-rat 7-Fc and the o p t i m u m conditions described above. The results of this comparison are shown in figs. 5A and 5B. The left-hand portions of fig. 5A and 5B obtained with Pool II also served as standard curves for determining the IgG anti-DNP activity of rat sera A, B and Pool III (table 2, right hand columns). Similar standard curves (not shown) were constructed with Pool I. Standards and unknowns were compared on the same day. Because of the difficulty in obtaining specifically purified anti-IgA and because of the poor quality of the IgM-enzyme complex, the dose dependency for IgA and IgM antibodies could n o t be compared using the two ELISAs. Instead, the amplified ELISA was used exclusively to study this function. The rat serum shown to contain IgG, IgM, and possibly IgA antibodies by RID was measured for these activities using the amplified ELISA. Fig. 6 shows a dose-response curve for this rat serum using all three antiglobulins. Dilutions of normal rat serum are used as controls. While fig. 6 illustrates a significant difference (2--10-fold) between the O.D. 4°° values of immunized versus control animals, the background O.D. 4°° of control animals varies as a function of dilution when IgA and IgG b u t not IgM antibodies are being measured. To determine the cause of non-specific binding by control sera, and to try and reduce this problem, two approaches were undertaken: (1) methycellulose (Fisher Chemical 400 centipoise) of 0.2 percent was added to all detergent-containing incubation steps, and (2)

l

-

0.9 0,6 0,7 O0 w 0.6 0.5 • 0

.

°. 111~0r, 50

4

1

~

,.

!

5~)0 1800 20C'C Reciprocal of Rabbit Antlglo::)alln

Fig. 4. E f f e c t o f anti-globulin (anti-TFc) c o n c e n t r a t i o n o n O.D. 4°° in amplified ELISA. • = 1 / 2 0 0 dilution o f high titer, Pool I. • = 1 / 5 0 0 dilutions o f early, l o w affinity antiDNP s e r u m , Pool B. Bridging a n t i s e r u m = 1 : 100, AP-A-AP. D e v e l o p m e n t t i m e s for the e n z y m e r e a c t i o n w e r e 45 a n d 75 rain, respectively.

378

A

1OO"

• Rat

anti

DNP

Pool

• ROt

ant t

DNP

Pool

~"

10-

E

c O O ~r r~

J

0 0.1

• J

O'I

.O1

I'O MIcro0rams of

I()O Antlbod y

IO()O

IB 10.O • Rat

anti

DNP

Pool

• Rat

anti

DNP

Pool

]I ]]]]

1.O¸

E

c O

O

j

°1

0

O. 1

i

O1

IO

I'-O Micrograms

16.O

160.O

10()O

of A n t l b o C l y

Fig. 5. Dose-response o f a m p l i f i e d E L I S A ( A ) and standard E L I S A (B). Legend is given o n graph and a n t i b o d y c o n c e n t r a t i o n s are presented in Table I. Each p o i n t is the m e a n of triplicate d e t e r m i n a t i o n s m a d e on the same day and are the O.D. 4°° o f the anti-DNP serum m i n u s the O.D. 4°° o b t a i n e d w i t h the same dilution o f normal rat serum. Dilutions o f various reagents were used as described in the text.

379

• R a t - A b - D N P Pool 2 I g G o N o r m a l Rat S e r u m Ig G A R a t - A b - D N P Pool 2 I g A Z~Normal Rat S e r u m Tg A • R a t - A b - D N P Pool 2 I g M o N o r m a l Rot S e r u m l g M

1.O 0 0 ,¢

(~ .!0

01 061

062 065 dl Reciprocal of Dilution

o2

Fig. 6. Dose response of amplified E L I S A for measuring IgM and IgA antibodies. Legend is given on figure. Serum source o f IgM and IgA antibodies was Pool II. C o n c e n t r a t i o n s of amplified E L I S A reagents were the same as used for anti-~/Fc and are described in the text. Values are mean of three determinations. TABLE 3 Effect of conjugate substitution and age o n background O.D. 400 in the amplified ELISA. Antiglobulin used

Anti-')' Anti-~ Anti-/~ Anti-~/ Anti-~ Anti-p Anti-7 Anti-~ Anti-p Anti-~ Anti-~ Anti-~ Anti-7 Anti-~ Anti-p Anti-T Anti-~ Anti-~ Anti-7 Anti-~ Anti-p Anti-')' Anti-~ Anti-p

Conjugate

Molar ratio DNP:protein

DNP-histone (one year old)

N.D.

DNP-histone (fresh)

N.D.

DNP-HSA

40.0

DNP-HSA

40.6

DNP-HSA

27.6

DNP-HSA

7.3

DNP-HSA

4.1

DNP-HSA

1.4

O.D. 4°0 1:100 anti-DNP Pool 2

1:100 NRS

0.211 0.171 0.204 0.557 0.326 0.382 0.543 0.376 0.393 0.554 0.429 0.423 0.455 0.290 0.269 0.271 0.164 0.200 0.201 0.156 0.156 0.216 0.140 0.124

0.286 0.149 0.184 0.053 0.080 0.129 0.138 0.162 0.168 0.156 0.168 0.212 0.140 0.145 0.135 0.135 0.119 0.137 0.162 0.101 0.117 0.121 0.126 0.111

Exp-control X 100% Control

Neg. 15.0 11.0 951 307 196 293 132 134 255 155 100 225 100 100 100 38 46 24 54 33 79 11 12

380 DNP-protein conjugates having different degrees of substitution or of different ages were tested. The data are shown in table 3. DISCUSSION Antisera of high specificity are essential for any antiglobulin assay. The immunoprecipitation results presented in fig. 1 were obtained in the presence of polyethylene glycol 6000. We have found this simple technique especially useful in our laboratory for detecting antigen-antibody reactions which are undetectable in the absence of PEG. Evaluation of antisera specificity b y PEG-ID have proven to be in agreement with specificity analyses at the immunohistochemical level (Butler et al., unpublished data). Therefore, the data we present in fig. 1 indicate the monospecificity of the anti ~/-Fc, anti-a, and anti-p reagents. The immunoelectrophoretic data on purified a:-macroglobulin (a:M) are consistent with those made by Heinberger (personal comm.) who demonstrated the polymorphism of this protein in the rat. Antiserum raised against this protein-complex has proven valuable in detecting a:M contamination in immunoglobulin preparations or the presence of anti-a2M antibodies in other antiglobulin reagents (fig. 1A, C). Incorporation of data on rat a:M is included in this report because: (1) rat a:M and serum IgA have almost identical electrophoretic mobility and are often confused, (2) some rat 'IgA preparations and anti-IgA sera' obtained from other investigators have been shown by us to be a2M and anti-a2M and (3) a2M prepared as we described, was also needed for specifically absorbing anti-IgA sera as shown in fig. 1C. Although our antiglobulin reagents were monospecific, preparing specifically purified anti-p of high titer and in sufficient quantity was disappointing. Preparing specifically purified anti-rat IgA proved to be impossible. Many of the problems encountered in specifically purifying anti-p and anti-a could be overcome in an animal species in which IgM and IgA could be purified in relatively large amounts. We have successfully purified specific antibodies to swine IgA (a) and IgM (g) and used them in the standard ELISA (Butler and Bibbe, unpublished). The failures reported here for antirat p and a, necessitated the development of the amplified ELISA. Initial studies with the amplified ELISA were undertaken to determine the optimal reagent concentrations. Detailed data on such titrations are presented for the rabbit anti-OA system. Because of an interesting behavior of the primary antiserum, data on the effect of the secondary antiserum (anti-~/-Fc) concentration are also shown for the rat system in fig. 4. A linear relationship was obtained for dilutions of 1 : 100 to 1 : 2000 for two rat anti-DNP sera of different antibody titer and affinity (table 2). Only using the combination of a high affinity antibody and low antiglobulin dilution ( 1 : 5 0 ) did the curve deviate from linearity. Although repeatable, this observation currently has no explanation. Because the abscissa of figure 4 is given in serum dilution and not anti-DNP content, the O.D. 4°° for the two

381 sera reflects the difference in their anti-DNP content. The validity of the amplified ELISA is indicated by comparison to the standard ELISA and the results of these studies are shown in fig. 5A and 5B. The 'plateau' region shown in fig. 5 has been a consistent p h e n o m e n o n in our laboratory and in those of others (J. Pitts, personal communications). When additional data points obtained with both the amplified and standard ELISA were plotted against antibody concentration on the abscissa, the point of change in the slope of the standard graph, i.e. where the plateau begins, occurred at the same antibody concentration. Hence, the plateau formation observed must be a simple case of affinity-independent antigen saturation and the two ELISA assays do not differ in their behavior in regard to this phenomenon. This is theoretically s o u n d because the primary antibody binding s'zep is identical for both assays. On the other hand, the fact that the change in slope for the amplified ELISA (fig. 5A) results in a flat plateau while for the standard ELISA (fig. 5B) the slope change does not result in a flat plateau, may represent a significant difference between the assays. Also, the O.D. 4°° values for Pool III are reproducibly greater than those for Pool II throughout the standard ELISA plot shown in fig. 5B. This difference may reflect yet another phen o m e n o n which may relate to antibody affinity. Data in support of this hypothesis is published elsewhere {Butler et al., in press). The calculation of the apparent antibody content of u n k n o w n anti-DNP sera, using standard ELISA curves like those shown in fig. 5A and 5B, resulted in the data of table 2. These data show that: (1) the extrapolated values obtained with the amplified ELISA are from two to more than 20 times greater than the values obtained with the standard ELISA, (2) except for Pool III, the ELISA data do not agree with total antibody measured by the double-precipitin assay and (3) the standard ELISA greatly underestimates the a m o u n t of antibody in Pools A and B. It is significant to note that both Pool A and B contain lower affinity antibody than Pools I, II and III. Because of the ability of our modification to readily detect antibodies n o t detectable in the standard ELISA we have introduced the name 'Amplified ELISA'. Finally, data in table 2 also indicate that values obtained with the ELISAs should not be expressed in gravimetric terms (pg/ml) but rather in arbitrary 'ELISA Units'. The detection of IgA and IgM antibody with the amplified ELISA also follows a predictable dose response (fig. 6) although non-specific background is more of a problem, especially at the highest and lowest serum dilutions. The use of methyl cellulose at a concentration of 0.2 percent did reduce the background O.D. 4°° obtained with control sera. The improvement was moderate and in some cases unreproducible. On the other hand, the preparation of DNP-conjugates with different degrees of substitution had a signifio cant effect as was shown in table 3. The data compared include that obtained with the fresh DNP-histone conjugate and a DNP-histone conjugate that had aged for one year. We had previously shown that histone alone did

382 not result in non-specific binding (high background with dilutions of normal serum). The data in table 3 show that: (1) the aged DNP-histone conjugate gave background values too high to make it useable in the assay, (2) fresh DNP-histone gave a larger difference between experimental and control values than highly conjugated DNP-HSA, (3) highly conjugated DNP-HSA (40 DNP groups/molecule) gave the lowest relative background of the DNP-HSA conjugates tested, and (4) poorly substituted DNP-HSA gave relative background values as high as the aged DNP-histone conjugate. Similar age-dependent deterioration occurs with TNP-conjugates (J.C. Cambier, personal communications). The data presented in fig. 3A and table 1 clearly show the usefulness of the amplified ELISA in the detection of rabbit antibodies to ovalbumin. Using the o p t i m u m conditions of secondary antiserum, bridging antiserum, and enzyme-antibody complex, the system is capable of detecting 2.4 ng of antibody with a high degree of reproducibility and moderate coefficients of variation. Even a 1 : 1,000,000 dilution of rabbit-OA differs from the background control b y a factor of 19-fold. The titration results we have presented for the rabbit system not only indicate the o p t i m u m reagent concentrations for the greatest sensitivity but also identify additional causes of non-specific background. Higher dilutions of both secondary antisera and antibody-enzyme complexes reduce background. One factor which may contribute to day-to-day or assay-to-assay background variation is the enzymatic reaction time. From the data presented, we believe the amplified ELISA to be a valid measure of IgA, IgM, and IgG antibodies to DNP and rabbit antibodies to ovalbumin. The use of the soluble AP-A-AP complex has eliminated the need for purified antibody-enzyme conjugates, which at least for rat IgM, or IgA was impractical or impossible. The amplified ELISA has been demonstrated with t w o sets of reagents prepared in primarily two different species; rabbits and guinea pigs. The use of rabbit anti-guinea pig IgG in the ovalbumin system eliminates the need for absorbing the bridging antiserum with IgG of the species of the primary antisera as was done for the goat anti-rabbit IgG used as a bridge in the rat anti-DNP system. The elimination of the specific antib o d y purification step can be considered as a simplification of the original standard ELISA assay although the greater number of incubation steps makes the amplified ELISA more complicated. Whether this should be considered simplification or complication depends on the species of the primary antibody. The method is a simplification if obtaining sufficient quantities of pure immunoglobulin {such as may be true for IgE in some species) is impractical. Although 'simplification' through elimination of the need to specifically purify antibodies was the stimulus for developing our modification, the ability of the amplified ELISA to detect low levels of antibody, especially when they are of low affinity (table 2 and Butler et al., in press) is perhaps of greater importance. This should be especially valuable when minor classes of

383 a n t i b o d y (such as IgA or IgM) are being studied. T h e i m m u n o c h e m i c a l basis f o r this a p p a r e n t a m p l i f i c a t i o n requires a d d i t i o n a l investigation. ADDENDUM A f t e r s u b m i s s i o n o f this m a n u s c r i p t we have f o u n d t h a t with the exception o f a d s o r p t i o n o f antigen and b i n d i n g o f p r i m a r y antisera, the t i m e o f all o t h e r i n c u b a t i o n steps can be r e d u c e d t o 2 h w i t h o u t loss o f sensitivity in the d e t e c t i o n o f r a b b i t a n t i - o v a l b u m i n antibodies. REFERENCES Blakeslee, D., J.E. Butler and W.H. Stone, 1971, J. Immunol. 107 (1), 227. Butler, J.E., 1974, In: Lactation: A Comprehensive Treatise, eds. B.L. Larson and V.R. Smith, Vol. III (Academic Press, N.Y.) pp. 217--255. Butler, J.E. and C.F. Maxwell, 1972, J. Dairy Sci. 55,151. Butler, J.E., T.L. Feldbush, P.L. McGivern and N. Stewart, Immunochemistry, in press. Cambier, J.C., 1975, Ph.D. Thesis, University of Iowa. Cambier, J.C. and J.E. Butler, 1976, J. Immunol. 116 {4), 994. Capra, J.D., M.N. Kehoe, R.C. Williams, T. Feizi and H.G. Kunkel, 1972, Proc. Natl. Acad. Sci. USA 69, 40. Coombs, R.R.A., W.E. Jonas, P.J. Lachmann and A. Feinstein, 1965, Int. Arch. Allergy 27,321. Cuatrecasas, P. and C.N. Anfinsen, 1971, In: Methods in Enzymology, eds. S.P. Colowick and N.O. Kaplan (Academic Press, N.Y.) Vol. XXII, p. 345. Engvall, E. and P. Perlmann, 1972, J. Immunol. 109,129. Engvall, E., K. Jonsson and P. Perlmann, 1971, Biochim. Biophys. Acta 251 (3), 427. Harrington, J.C., J.W. Fenton II and J.H. Pert, 1971, Immunochemistry 8,413. Ishizaka, K. and T. Ishizaka, 1967, J. Immunol. 99 (6), 1187. Lamed, R., J. Levin and M. Wilchek, 1973, Biochim. Biophys. Acta 304, 231. Leone, C.A., 1968, In: Methods in Immunology and Immunochemistry, eds. C.A. Williams and M.W. Chase (Academic Press, N.Y.) p. 174. Little, J.R. and H.N. Eisen, 1967, In: Methods in Immunology and Immunochemistry, eds. C.A. Williams and M.W. Chase (Academic Press, N.Y.) Voi. I, pp. 128--133. Mancini, G., J.P. Vaerman, A.O. Carbonara and J.F. Heremans, 1963, Protides Biol. Fluids 11, 30. Marchalonis, J.J., 1969, Biochem. J. 113, 299. Minden, P. and R.S. Farr, 1967, In: Handbook of Experimental Immunology, ed. D.M. Weir (F.A. Davis Co., Philadelphia) p. 463. M~iller-Eberhard, H., 1960, Scand. J. Clin. Lab. Invest. 12, 33. Porath, J., J. Janson and T. L~i~is, 1971, J. Chromatog. 60, 167. Richerson, H.B., 1974, Ann. N.Y. Acad. Sci. 221,340. Scheidegger, J.J., 1955, Int. Arch. Allergy 7,103. Sloan, G.J., 1975, M.S. Thesis, University of Iowa. Spiegelberg, H.L., 1974, Adv. Immunol. 19,259. Sternberger, L.A., P.H. Hardy, Jr., J.J. Cuculis and H.G. Meyer, 1970, J. Histochem. Cytochem. 18,315. Yagi, Y., P. Maier and D. Pressman, 1962, J. Immunol. 89, 736. Zamenhof, S., 1957, In: Methods in Enzymology, Vol. III, eds. S.P. Colowick and N.O. Kaplan, p. 702.