Appearance in mice during infection with lactate ... - Wiley Online Library

2 downloads 0 Views 694KB Size Report
... Mary Jo Jaqua, and. Peter G.W. Plagemann ... carried out in the presence of detergent, for example, 0.05% Tween 20 [Bullock and. Walls, 19771. We have ...
Journal of Medical Virology 19: 175-186 (1986)

Immunoglobulins That Bind to Uncoated ELISA Plate Surfaces: Appearance in Mice During Infection With Lactate-DehydrogenaseElevating Virus and in Human Anti-Nuclear Antibody-Positive Sera William A. Cafruny, Daniel P. Heruth, Mary Jo Jaqua, and Peter G.W. Plagemann Departments of Microbiology (W.C., D.H.) and Laboratory Medicine (M.J.), School of Medicine, University of South Dakota, Vermillion, South Dakota and Department of Microbiology (P.P.), University of Minnesota Medical School, Minneapolis, Minnesota Immunoglobulins present in the blood plasma of mice infected with lactatedehydrogenase-elevating virus (LDV) were found to bind strongly in the presence of 0.05% Tween 20 to the uncoated surfaces of wells of certain ELISA plates with previously recognized high protein-binding capacity. The binding was readily distinguishable from non-specific background binding of immunoglobulins present in normal mouse plasma. The binding components absorbed to protein A and had molecular weights in the 150-300 kDa range. Binding of the purified IgG fraction was progressively inhibited by increasing the concentration of Tween 20 in the diluent and by preincubation of the fraction at pH 3-4 for 10 min. The appearance of plate-binding IgM and IgG during LDV infection corresponded approximately with previously reported time courses of appearance of IgM- and IgG-containing circulating immune complexes and of specific IgM and IgG anti-LDV antibodies in LDV-infected mice. We conclude that complexes of IgG and IgM with LDV antigens have a much higher affinity for ELISA plates with high protein-binding capacity than uncomplexed immunoglobulins. Immune complexes did not significantly bind to ELISA plates with low protein-binding capacity, which, therefore, are suitable for measuring specific antiviral antibodies. Preliminary experiments with human anti-nuclear antibody-positive serum samples demonstrated markedly elevated non-specific binding of immunoglobulins to high-binding-capacity ELISA plates. Key words: ELISA, immunoglobulin, lactate-dehydrogenase-elevatingvirus

INTRODUCTION The enzyme-linked immunosorbent assay (ELISA) [Engvall and Perlman, 19713 is useful for the detection of antigens and antibodies in a wide range of experimentally

Accepted for publication February 11, 1986. Address reprint requests to Dr. William Cafruny, Department of Microbiology, School of Medicine, University of South Dakota, Vermillion, South Dakota 57069.

0 1986 Alan R. Liss, Inc.

176

Cafruny et a1

and clinically useful tests [Voller et al, 1978; Yolken, 19821. For use in the detection of specific antibodies, the plastic wells of microplates are coated with antigen, followed by sequential addition of test samples containing antibodies, enzyme-conjugated second antibodies to the antibody to be assayed, and the assay of bound enzyme activity [Voller et al, 19781. Antigens that are protein in nature attach to the plastic surfaces of microplates by hydrophobic interactions [Kochwa et al, 1967; Van Oss and Singer, 19661. The degree of binding of proteins to plastic surfaces is dependent on the type of plastic used [Lee et al, 19741 as well as on the source of the microplates themselves [Kenny and Dunsmoor, 19831. To reduce non-specific interaction of antibodies with antigen-coated microplate wells, reactions with antibodies are usually carried out in the presence of detergent, for example, 0.05% Tween 20 [Bullock and Walls, 19771. We have previously developed an ELISA for measuring IgG and IgM to the lactate-dehydrogenase-elevatingvirus (LDV) of mice using MicroELISA microtiter plates from Dynatech Laboratories [Cafruny and Plagemann, 1982al. After these plates were no longer available, we obtained Nunc Immunoplates I (Vanguard International) but found that anti-LDV IgG and IgM from persistently infected mice strongly bound to these plates, even though they had not been coated with antigen. This binding was readily distinguishable from the normal non-specific binding of mouse immunoglobulin from uninfected mice to the plates; it apparently reflects binding of viral antigen-antibody complexes. Thus these plates might be useful for the quantitation of immunocomplexes or immunoglobulins altered by virus infection. In the present report, we describe these results and those from preliminary studies dealing with the nature of the anti-LDV immunoglobulins binding to the Nunc plates. Similar immunoglobulins were also observed in certain human sera. MATERIALS AND METHODS Mice and Infection With LDV

BALB/c mice were obtained from Sasco (Omaha, NE) or bred in the Department of Microbiology, University of Minnesota. Outbred Swiss mice were obtained from Sasco (CF-1 strain) or Bio-Labs (St. Paul, MN). All mice used were female. Groups of five or more mice were infected with LDV as described previously [Plagemann et aI, 19631. The mice were bled from the orbit at the times indicated, and plasma samples from individual mice were pooled and frozen at -70°C until used. 3H-labeled LDV was prepared by propagating the virus in primary macrophage cultures in the presence of [3H]-uridine [Brinton-Darnell and Plagemann, 19751. ELSA Technique

Nunc immunoplate I, lot Nos. 4683 and 3583, were obtained from Vanguard International (Neptune, NJ) . Other plates were obtained from Dynatech Laboratories (Alexandria, VA), Flow Laboratories (McLean, VA), and Costar (Cambridge, MA). All experiments were carried out using microplates that had no previous antigen coating or other treatment. Blood plasma from normal control or LDV-infected mice was serially diluted in phosphate-buffered saline, pH 7.4, containing 0.05 % Tween 20 (PBS-Tween) and 200 pl added to each well. In some experiments the concentration of Tween 20 was varied as stated. The plates were normally incubated at 4°C

Immunoglobulin Binding to ELISA Plates

177

overnight (16-18 hr), then washed three times with PBS-Tween, and alkalinephosphatase-conjugated antibody was added. Alkaline-phosphatase-conjugatedantibodies were obtained and used as follows: goat anti-mouse IgM from Kirkegaard and Perry Labs (Gaithersburg, MD) diluted 1:400; goat anti-mouse IgG from Cappel (Cochranville, PA) or Sigma (St. Louis, MO) diluted 1:150 or 1:300 all added at 200 pl/well. After 2.5-3 hr at room temperature, the plates were washed three times with PBS-Tween, and substrate was added (150 or 200 pl/well). After 30-60 min of further incubation, the reaction was stopped by addition of 25 p12 N or 50 pl 1 N NaOH per well, and the absorbancy was measured at 405 nm in a Beckman spectrophotometer or a Dynatech Minireader II. Since the reaction rate varied with the amount, type, and source of second antibody added, NaOH was added when visual examination indicated an absorbancy between 1 and 2.5 units in any of the wells (generally between 30 and 60 min) or at 60 min of incubation. The ELISA titer was designated the reciprocal of the highest dilution of immune plasma that yielded an absorbancy of 50-100% above that obtained with mouse plasma from uninfected mice under identical experimental conditions (see Fig. 1B).

Purification of IgG IgG was adsorbed onto a column of protein A-Sepharose essentially as described previously [Cafruny and Plagemann, 1982al. Plasma pooled from BALB/c mice infected with LDV 8 months previously, or from control uninfected BALB/c mice, was diluted 1:1 in PBS, and 0.275 ml was applied to a column of protein A-Sepharose

I

I

40

80

I

120

1

I

160

200

240

TIME pi. (DAYS)

Fig. 1. Binding of IgG from LDV-infected mice to Nunc Immunoplate I microtiter plates not coated with antigen. Seven BALB/c mice (about 2 months of age) were infected with LDV and bled at intervals. Their plasma was pooled and dilutions thereof in PBS-Tween (1:50 to l:I,600)were assayed with alkaline-phosphatase-conjugated goat anti-mouse IgG, as described in Materials and Methods. A) The absorbancies observed with the 1: 100 dilution of plasma are plotted as a function of time p.i. B) Only representative dose-response curves are shown in B. T indicates the time p i . the mice were bled.

178

Cafruny et a1

(Sigma) that had a bed volume of 0.5 ml. The column was washed with 5 ml PBS, collecting 0.275-ml column fractions. Then, 0.250 ml citrate-phosphate buffer (pH 3.3) was applied to the column, followed by PBS, and the acid fraction containing the eluted IgG was immediately neutralized with 2 N NaOH. The eluted fraction from LDV-infected plasma contained a protein concentration of 2.1 mg/ml as estimated by absorbancy at 280 nm. Other Reagents

Mouse IgGI, IgG2, IgG2,,, and IgG3 subclasses were obtained from Miles Laboratories (Elkhart, IN). Human Serum Samples

Sixteen human serum samples were selected on the basis of clinical laboratory findings of positive anti-nuclear antibody (ANA) titers between 1:40 and 1: 1280 by indirect immuno-fluorescence [Fritzler, 19801. For comparison, 16 control sera were obtained from apparently healthy college students (ten females and six males). The clinical samples were analyzed for total IgG and IgM by nephelometry using the Beckman Immunochemistry System.

RESULTS Appearance of ELISA Plate-Binding IgG and IgM in Mouse Plasma After LDV Infection

The results in Figure 1 show that IgG binding to Nunc ELISA plates appeared in the plasma of LDV-infected BALB/c mice within 10 days postinfection (p.i.). Maximum titers of 1,600 to 3,200 (Fig. 1B) were attained about 40 days p.i. and then remained approximately constant until at least 8 months p.i. Since the differences in the ELISA reaction between immune and normal plasma were greatest at plasma dilutions of 1:100 or 1:200, and the absorbancy values fell on the approximately linear portion of the dose-response curves and thus were proportional to the titer (see Fig. lB), we have used these values as a measure of the level of immunoglobulin that binds to these plates (Fig. 1A). To allow a valid comparison of individual plasma samples by this method, all samples to be directly compared were analyzed under identical conditions with respect to reagents, processing of the plate, and incubation times, generally on a single plate. In the experiments illustrated in Figure 2, groups of LDV-infected mice were bled at 1- and 2-day intervals, to assess more clearly the early time course of appearance in the plasma of plate-binding IgG. We similarly assayed the plasma for plate-binding IgM. The results show that plate-binding IgG, as well as IgM, appeared very rapidly in infected mice, both reaching an initial peak about 6 days p.i. However, the levels of plate-binding IgM were lower (the maximum titers were 1:400 to 1:800; data not shown) than those for IgG, and they subsequently decreased to the level of normal mouse plasma by 30-40 days p.i. In contrast, the time course of appearance of plate-binding IgG was biphasic; the level decreased between 6 and 10 days p.i. It then increased again to a second maximum level by 40 days p.i., which, as illustrated in Figure 1, persisted for many months. Our workmg hypothesis therefore is that the IgM and IgG from LDVinfected mice that bind to Nunc plates represent immune complexes containing virus

Immunoglobulin Binding to ELISA Plates

TIME

179

PI.(DAYS)

Fig. 2. Early time course of development of Nunc plate-binding immunoglobulins during LDV infection. Groups of five to ten LDV-infected A) BALB/c or B) Swiss mice were bled at the indicated times p.i. The plasma samples from each group were pooled, and samples of a 1:200 dilution thereof in PBSTween were added to uncoated wells of Nunc Immunoplate I microtiter plates. Alkaline-phosphataseconjugated anti-mouse IgM ( 0 )or IgG ( A ) was used to detect bound immunoglobulins. A11 points represent averages of duplicate wells.

or viral antigens. LDV-antibody complexes are known to be present in persistently infected mice and to entrap practically all LDV in the plasma in an infectious form [Cafruny and Plagemann, 1982a; Notkins et al, 1966; McDonaId et al, 19831. Most of the following results support this hypothesis.

Physical and Chemical Analysis of ELISA Plate-Binding IgG The material from the plasma of infected mice that bound to Nunc plates and reacted with rabbit anti-mouse IgG was completely retained on a protein A-Sepharose column (Table IA), confirming that it was IgG. The material that was eluted from the column at pH 3.3 still bound to Nunc plates at a detectable concentration of about 5 pglml, but its titer was reduced by about 75 76. A more detailed analysis of the pH sensitivity of the binding activity of the IgG fraction eluted from the protein ASepharose column is presented in Table IB. The binding of IgG was inhibited progressively with a decrease in treatment pH. Binding was practically abolished by the 10 min incubation at pH 3.3 or 4.0. There was no detectable reactivation of binding activity when the fractions were maintained at pH 7.4 for up to 6 hr on ice (data not shown). Nunc-plate-binding activity in immune plasma from LDV-infected mice was not sedimented by high-speed centrifugation (100,OOO x g for 2 % hr, data not shown), indicating that the binding activity was not due to the presence of intact virions or to intact virus complexed with antibody. This conclusion is supported by an isopycnic sedimentation analysis of immune plasma from %month infected mice (Fig. 3A). None of the Nunc-plate-binding activity was associated with LDV, which equilibrates at a density of about 1.14 g/cm3 in this type of gradient [Brinton-Darnel1 and Plagemann, 19751. Furthermore, extraction of plasma from LDV-infected mice with ethyl ether, which is known to inactivate LDV infectivity and cause the disinte-

180

Cafruny et a1

TABLE I. Effects of Protein A Adsorption and Pretreatment at Low pH on Binding of IgG From LDV-Infected Mice to Nunc Plates Exp.

Test material

PH

Absorbance at 405 nm

A'

IMP, unadsorbed IMP, adsorbed Protein A, eluted fraction

7.4 7.4 7.4

1.47 0.03 0.69

Bb

Protein A. eluted fraction

7.0 6.0 5.0 4.0 3.3

0.80 0.68 0.50 0.22 0.18

'Plasma from BALB/c mice 8 months post-infection with LDV (see Fig. 1) was passed through a protein A-Sepharose column and the flow-through (adsorbed) fraction, the IgG fraction eluted from the column at pH 3.3, and unadsorbed plasma were titrated (150 to 1:3,200) in an uncoated Nunc Immunoplate I. The values shown are for the 1: 100 dilution in PBS-Tween. bSamples from the eluted fraction described above were diluted 1: 10 in buffer of the indicated pH. After 10 min at 4"C, the mixtures were diluted in PBS-Tween, pH 7.4, and titrated as described (]:lo0 to I :600). The values shown are for the I :100 dilution.

gration of virions [Notkins et al, 19661 had little effect on the binding of IgG to Nunc plates (Table IIA). In fact, the results in Figure 3A suggest that the Nunc-plate-binding material was of relatively low molecular weight, since it had hardly entered the gradient. However, it seemed to sediment more rapidly than the bulk of the proteins of the plasma. The molecular weight of the Nunc-plate-binding material was analyzed in greater detail by zone sedimentation in a sucrose density gradient (Fig. 3B). The bulk of the protein of the plasma cosedimented with bovine serum albumin, whereas the bulk of the Nunc-plate-binding activity sedimented more rapidly than bovine yglobulin. The binding material was heterogenous in size, ranging from about 150300 kDa. Furthermore, it was highly active; 5 pg protein of gradient fractions 30-36 yielded a maximum ELISA response. In fact, as little as 0.01 pg protein/well yielded a positive response. These results are consistent with the view that the binding activity reflects complexes between individual LDV proteins and anti-LDV IgG. In this context, we have also ruled out that binding activity resulted from a subclass of IgG that might be enhanced in persistently infected mice during the course of the polyclonal activation induced by LDV infection [Cafruny and Plagemann, 1982a1, since there was little binding of purified mouse IgG,, IgG2,, IgG2b, or IgG3 to Nunc plates (Table IIB). Some of the IgG from LDV-infected mice bound very rapidly to the plates (Fig. 4B), but the total amount bound increased up to 73 hr of incubation (Fig. 4A). Furthermore, the binding of both IgG and IgM from infected mice was markedly reduced with an increase in Tween 20 concentration in the binding solution (Table 111); there was little binding in PBS containing 4.5 % Tween 20. Binding of IgG From LDV Infected Mice to Different Microtiter Plates

The binding of IgG from LDV-infected mice was not limited to Nunc plates. Extensive binding was observed with Dynatech Immulon I1 and Costar EIA plates, as well as with tissue culture microtiter plates not designed for ELISA (Table IV). On

Immunoglobulin Binding to ELISA Plates

181

15I

n 1 2

a I -10U

0

t

z 0

a

U

.

IL LL

5

n

-

_I

0-

I

Top

FRACTION NUMBER

Fig. 3. Analyses of Nunc-plate-binding IgG of plasma from LDV-infected mice by A) isopycnic and B) zone sedimentation in sucrose density gradients. Samples of 0.5 ml of immune mouse plasma from a group of 50 2-3 month LDV-infected Swiss mice were centrifuged A) on a linear 0.5-1.9 M sucrose gradient in a SW 27 rotor at 22,000 rpm (Beckman ultracentrifuge) at 6°C for 12 hr, or B) on a linear 0.1-0.9 M sucrose gradient in a SW 41 rotor at 38,000 rpm at 6°C for 23 hr. Fractions of 1 ml or 0.25 ml were collected from the gradients, respectively, and samples thereof were analyzed for protein (0) by the method of Lowry et al, 1951, and for IgG binding to Nunc plates in the absence of antigen coating ( 0 ) Gradient . fractions were serially diluted from A) 1: 10 to 1:80 or from B) 1:20 to 1:640 and were analyzed by ELISA, as described in Materials and Methods. The absorbancies for the 1:40 dilutions are plotted. In A), LDV ( p = 1.14 g/cm3), which had been labeled with t3H] uridine, was sedimented and gradient fractions were analyzed for acid-insoluble radioactivity. In B) samples of 1 mg/ml solutions of bovine serum albumin (BSA, 66 kDa) and bovine y-globulin (150 kDa) were cosedimented in companion gradients, and fractions from the gradient were analyzed for protein. The arrows indicate the positions of peaks of the proteins in these gradients.

the other hand, there was little IgG binding to Linbro EIA and Dynatech Immulon I plates. Binding of Human Immunoglobulins to E L S A Plate Wells With High Binding Capacity

Human sera diluted 1:400 in PBS-Tween 20 (0.05%) were added to wells of high binding capacity (Nunc Immunoplate I) or low binding capacity (Linbro EIA) ELISA plates. After 18 hr at 4"C, the plates were washed three times and developed with alkaline-phosphatase-conjugatedpolyvalent antisera to human immunoglobulins G and M (Sigma). Figure 5 shows results using high-binding-capacity plates, indicating that the samples from ANA-positive individuals exhibited significantly higher binding to the plates (mean A405 = 0.343 k 0.160) relative to the control samples = 0.153 f 0.029). Results from low-binding-capacity plates were mark(mean hO5 edly different (data not shown), with absorbancy readings for all wells below 0.030, except for one ANA-positive sample that gave a value of 0.141. DISCUSSION

The time course of appearance of IgM and IgG binding to Nunc plates agrees, in general, with the formation in LDV-infected mice of immune complexes containing

182

Cafruny et a1

TABLE 11. Ether Extraction of Plasma from LDV-Infected Mice Does Not Inhibit IgG Binding to Nunc Plates, and Purified Mouse IgG Subclasses Exhibit Minimal Binding to Nunc Plates Exp.

Absorbance at 405 nm

Test Material

A=

IMP IMP, ether extracted NMP

Bb

1.10

1.20 0.19

IgGl IgG2,

0.25 0.40 0.30 0.48 2.15 0.49

kGZb

IgG3 IMP NMP

aPlasma from BALB/c mice 3 months post-infection with LDV (IMP) was titrated (150 to 1:6,400) in an uncoated Nunc Immunoplate I before or after three extractions with ethyl ether (performed as described by Notkins et al, 1966) along with plasma from uninfected BALB/c mice (NMP). Values are shown for the 1: 100 dilution. bPurified mouse IgG subclasses were dissolved in PBS (pH 7.4) at 1 mg/ml and then titrated (1: 100 to 1:6,400) in an uncoated Nunc Immunoplate I along with IMP and NMP. Values are shown for the 1:400 dilution.

o.o@ 0

--

-

- ,

n

10

20

30

9

v,

40

50

60

70

Fig. 4. Time course of binding of purified plasma IgG from LDV-infected mice to Nunc plates. The IgG fraction from plasma of 8-month LDV-infected BALB/c mice was diluted 1:200 in PBS-Tween. Aliquots thereof were added at timed intervals to the wells of a Nunc Immunoplate I microtiter plate that was kept at 4°C.Bound IgG was then quantitated with alkaline-phosphatase-conjugatedgoat anti-mouse IgG. All points represent averages of duplicate wells. Open circles represent normal mouse IgG control.

Immunoglobulin Binding to ELISA Plates

183

TABLE 111. Effect of Tween 20 Concentration on Nunc Plate-Binding Activity of IgG and IgM From LDV-Infected Mice Tween 20 (%)

IMP" (days pi)

0.05

0.15

0.5

1.5

4.5

6 240

0.92 1.33

N D ~ 1.14

0.35 0.91

ND 0.60

0.08 0.26

"Plasma pooled from 6-day and 8-month LDV-infected BALB/c mice (IMP) was titrated (1:200 to 1 :6400)in uncoated wells of Nunc Immunoplate I microtiter plates with alkaline-phosphatase-conjugated goat anti-mouse IgM and anti-mouse IgG, respectively. Plasma dilutions were made in PBS (pH 7.4) containing the indicated concentrations of Tween 20. Values are in absorbancy at 405 nm and are for the 1:200 dilution. bND, not determined.

TABLE IV. Comparative Binding of IgG From Plasma of LDV-Infected Mice to Various Microtiter Plates Absorbance at 405 .ma Plate Nunc immunoplate I Costar EIA Dynatech immulon I1 Dynatech immulon I Linbro EIA Linbro tissue cultureb Falcon microtest IIIb

LDV infected

Uninfected

No mouse plasma

1.060 0.845 0.801 0.040 0.043 0.495 0.523

0.066 0.040 0.040 0.031 0.033 0.251 0.254

0.039 0.037 0.034 0.030 0.031 0.060 0.033

aPlates received aliquots of 1:200 dilutions in PBS-Tween of plasma from uninfected mice or from mice 8 months p.i. with LDV, or they received PBS-Tween only (no mouse plasma). The plates were developed with goat anti-mouse IgG. All values are averages from duplicate wells. bNot marketed for ELISA procedure.

IgM and IgG that bind to Clq, as reported previously [McDonald et al, 19831. These investigators showed that immune complexes containing IgM were present between 3 and 9 days p i , with a peak at 6 days. Immune complexes containing IgG, were observed between 3 and 12 days p i , which corresponds to the early peak of Nuncplate-binding IgG (Fig. 2), whereas immune complexes containing IgG2 began to appear about 15 days p.i. The specificities of the IgM and IgG in immune complexes that bind to C l q were not determined. They probably reflect antibodies to LDV proteins, since their appearance coincides with the formation of specific anti-LDV antibodies. The latter were measured using ELISA with plates-such as Dynatech Immulon I or Linbro EIA-IgM and IgG from LDV-infected mice do not bind to these plates to a significantly greater extent than those from uninfected mice (see Table IV), except when they are coated with LDV at lo8 to lo9 IDSO/well[Cafruny and Plagemann, 1982b, and unpublished data]. Although immune complexes are formed during persistent LDV infection of mice that entrap practically all LDV in the blood in an infectious form [Cafruny and Plagemann, 1982a; Notkins et al, 19661, the ELISA plate-binding IgG cannot be due to the presence of immune complexes containing intact virions, since the binding IgG was not sedimented at 100,OOO X g for 2 % hr, did not cosediment with LDV in sucrose density gradients, and was not affected by extraction of the mouse plasma with ethyl ether.

184

Cafruny et a1

On the contrary, our results place the molecular weight of the Nunc-platebinding fraction in the 100-300 kDa range, which, combined with the heterogeneous size of the material, makes it likely that it is composed of anti-LDV IgG complexed to individual LDV proteins. An alternative possibility is that the binding material reflects mouse IgG that has been structurally altered as a result of LDV infection or that is produced as part of the polyclonal activation induced by LDV [Cafruny and Plagemann, 1982al. Whichever is the case, our data suggest that the reactive immunoglobulins develop a hydrophobic interaction with the microplate surface, which appears to be responsible for the binding affinity, as indicated by the rapid rate of reaction and its inhibition by increasing detergent concentrations. Dissociation of antigen-antibody complexes in low pH [Campbell and Weliky, 19671 presumably leads to loss of the hydrophobic interaction with the plate surface. It has previously been reported [Kenny and Dunsmoor, 19831 that Dynatech Immulon I1 plates and Nunc Immunoplates I and I1 exhibited a particularly high binding capacity for certain proteins, such as bovine serum albumin, whereas the binding of human IgG to a variety of commercial plates differed little. The Nunc plates also exhibited a relatively high non-specific binding of conjugated antibodies, which was progressively reduced by an increase in the concentration of Tween 20 in the diluent. Thus, the binding of immunoglobulins or immune complexes from LDVinfected mice to certain ELISA plates is a function of the high binding capacity of the plates. Preliminary results have demonstrated that human immunoglobulins also display non-specific reactivity with high-binding-capacity ELISA plates in the presence of 0.05 % Tween 20 and enhanced reactivity correlated with the presence of anti-nuclear

-

0.7

:

EC 0.6

0" 0.5

t -

0.4 0 2

a 0.3 m

R

$ 0.2 m

< 0.1 0.0

A N A -?ositive

Control

SERUM Fig. 5. Binding of human immunoglobulins to ELISA plate wells with high binding capacity. Serum samples containing positive anti-nuclear antibody titers (ANA positive) or obtained from apparent healthy college students (control) were diluted 1:400 in PBS-Tween 20 (0.05%) before addition to the wells of Nunc Irnmunoplate I ELISA plates. Bound immunoglobulins were detected with alkalinephosphatase-conjugated anti-human IgG and IgM. The average of duplicate determinations is shown for each sample. The bars represent mean rt SD.

Immunoglobulin Binding to ELISA Plates

185

antibodies (Fig. 5). Plate-binding titers in the ANA-positive samples ranged from 1:400 to 1:3,200 and were not directIy related to the total IgG or IgM content of the samples, since these were within the normal range for about one-half of the samples. The nature of the human immunoglobulins that bound to high-binding-capacity plates is not known, but they may represent immune complexes known to occur in patients with positive ANA titers [Fritzler, 19801. These preliminary data suggest that the phenomenon of non-specific binding we have described may be important during analysis of human antibodies in ELISA plates with high binding capacity. Our results suggest that caution is needed in the application of ELISA to measuring specific viral antibodies in sera of infected animals and humans, particularly with respect to selection of appropriate plates. Furthermore, certain ELISA plates with high binding capacity might be usefid for quantitating antigen-antibody complexes. ACKNOWLEDGMENTS The authors thank George Kenny for helpful discussion, Shida Yousefi for competent technical assistance, and Pat Hovinen and Yvonne Guptill for secretarial assistance. This work was supported by grants from The Parsons Fund and the General Research Committee (WAC) and by Public Health Service research grant A1 15267. REFERENCES Brinton-Darnel1 M, Plagemann PGW (1975): Structure and chemical-physical characteristics of lactate dehydrogenase-elevating virus and its RNA. Journal of Virology 16:420-433. Bullock SL, Walls KW (1977): Evaluation of some of the parameters of the enzyme-linked immunospecific assay. Journal of Infectious Diseases 136:S279-285. Cafruny WA, Plagemann PGW (1982a): Immune response to lactate dehydrogenase-elevating virus: Isolation of infectious virus-immunoglobulin G complexes and quantitation of specific antiviral immunoglobulin G response in wild-type and nude mice. Infection and Immunity 37: 1001-1006. Cafruny WA, Plagemann PGW (1982b): Immune response to lactate dehydrogenase-elevating virus: Serologically specific rabbit neutralizing antibody to the virus. Infection and Immunity 37: 10071012. Campbell DH, Weliky N (1967): Immunoadsorbents: Preparation and use of cellulose derivatives. In Williams CA, Chase MW (eds): “Methods in Immunology and Immunochemistry.” NY: Academic Press, p 371. Engvall E, Perlman P (1971): Enzyme-linked immunosorbent assay (ELISA): Quantitative assay of immunoglobulin G. Immunochemistry 8:871-874. Fritzler, MJ (1980): In: Rose NR, Friedman H (eds): “Manual of Clinical Immunology.” Washington, DC: American Society for Microbiology, pp 852-857. Kenny GE, Dunsmoor CL (1983): Principles, problems, and strategies in the use of antigenic mixtures for the enzyme-linked immunosorbent assay. Journal of Clinical Microbiology 17:655-665. Kochwa SM, Brownell M, Rosenfield RE, Wassermann LR (1967): Absorption of proteins by polystyrene particles. Journal of Immunology 99:981-986. Lee RG, Adamson C, Kim SW (1974): Competitive absorption of plasma proteins onto polymer surfaces. Thrombosis Research 4:485-490. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ (1951): Protein measurement with the Folin phenoI reagent. Journal of Biological Chemistry 193:265-275. McDonald TL, Donnelly T, Weber A, Quenette L (1983): Antibody classes and subclasses in circulating immune complexes isolated from mice infected with lactic dehydrogenase virus. Immunology 48: 5 11-5 17.

186

Cafruny et a1

Notkins AL, Mohar S, Schele C, Goffman J (1966): Infectious virus-antibody complex in the blood of chronically infected mice. Journal of Experimental Medicine 124:81-97. Plagemann PGW, Gregory KG, Swim HE, Chan KKW (1963): Plasma lactic dehydrogenase elevating agent of mice: Distribution in tissues and effect on lactic dehydrogenase isoenzyme patterns. Canadian Journal of Microbiology 9:75-86. Van Oss CJ, Singer JM (1966): The binding of immune globulins and other proteins by polystyrene latex particles. Journal of Reticuloendothelial Society 3:29-40. Voller A, Bartlett A, Bidwell DE (1978): Enzyme immunoassays with special reference to ELISA techniques. Journal of Clinical Pathology 3 I :507-520. Yolken RH (1982): Enzyme imrnunoassays in the detection of infectious antigens in body fluids: Current limitations and future prospects. Review of Infectious Diseases 4:35-68.