Improvements to the enzyme-developed radial

immunodiffusion technique

Jose Vidal*

School of Psychology, University of Barcelona, Passeig de la Vall d’Hebron, 171, 08035 Barcelona, Spain

Received 21 December 2001; received in revised form 13 May 2002; accepted 15 July 2002

Abstract

An enzyme-developed radial immunodiffusion technique, previously known as the diffusion-in-gel enzyme-linked

immunosorbent assay (DIG-ELISA), has been improved in two ways: (a) antibody-containing spots have been made larger and

more distinct by revealing them with a mixture of hydrogen peroxide, 3,3V-diaminobenzidine and nickel, and further

intensification of the ensuing spots with silver; (b) the reliability of the method has been enhanced by chemically coupling the

antigen to a layer of a polyamino acid (poly(lysine, phenylalanine)) adsorbed to the bottom of the polystyrene petri dish. The

usefulness of the improved technique is illustrated by reference to the measurement of serum concentrations of IgM and IgG,

and in the assessment of antibody levels against a particulate antigen (erythrocytes).

D 2002 Elsevier Science B.V. All rights reserved.

Keywords: DIG-ELISA; Silver enhancement; Diaminobenzidine; Poly(lys, phe); Immunoglobulin measurement

1. Introduction

The so-called diffusion-in-gel enzyme-linked

immunosorbent assay (DIG-ELISA; Elwing and Nyg-

ren, 1979) was developed as a sensitive, simple, and

inexpensive technique to measure antibody concen-

trations (Elwing et al., 1980; Cursons, 1982; Demat-

teis et al., 1989; Gunnarsson and Svedhem, 1998). In

this technique, the antibody diffuses in a gel over an

antigen-coated surface and areas of bound antibody

are revealed enzymatically: The diameters of the spots

are linearly related to the logarithm of the serum

dilutions. This technique is a variation of the radial

immunodiffusion technique, since areas of antibody

content are revealed not as precipitation rings but as

spots. Therefore, DIG-ELISA should more appropri-

ately be called ‘‘enzyme-developed radial immuno-

diffusion’’. In general, the system used to reveal the

spots has been peroxidase and a mixture of hydrogen

peroxide and p-phenylenediamine in a gel: The

insoluble product that results from polymerization of

p-phenylenediamine is trapped in the gel at the site of

antibody location and becomes visible as a brown

spot. Despite its simplicity, this procedure to reveal

spots has the following drawbacks when the antibody

concentration is low: (i) after moderate incubation

0022-1759/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved.

PII: S0022 -1759 (02 )00300 -9

Abbreviations: DIG-ELISA, diffusion-in-gel enzyme-linked

immunosorbent assay; DAB, 3,3V-diaminobenzidine tetrahydro-

chloride dihydrate; PBS, phosphate-buffered saline.

* Tel.: +34-93-3125112; fax: +34-93-4021362.

E-mail address: jvidal@psi.ub.es (J. Vidal).

www.elsevier.com/locate/jim

Journal of Immunological Methods 270 (2002) 163–170

(10–30 min), to avoid background emergence, the

spots have fuzzy contours, which makes diameter

measurement imprecise; (ii) if incubation proceeds

overnight (in an attempt to increase spot definition),

the spot contours become distinct, despite a high

background. However, the procedure is already

time-consuming (Gunnarsson and Svedhem, 1998),

and increased incubation times are undesirable.

Another peroxidase substrate, 3,3V-diaminobenzi-

dine (DAB), has been used, together with silver, in

immunohistochemistry to stain tissue sections (Mer-

chenthaler et al., 1989; Frigo et al., 1991) and to

reveal protein bands in immunoblots (Frigo et al.,

1991; Ludany et al., 1993). Therefore, the same

procedure could reveal antibody spots in the DIG-

ELISA and could overcome the abovementioned

drawbacks of disclosure with p-phenylenediamine.

As reported here, development with DAB–nickel

and silver yielded sharper, and larger, spots than

development with p-phenylenediamine.

While the author of this report was measuring

immunoglobulin concentrations in human serum, it

became evident that the DIG-ELISA was unreliable;

i.e., two measurements of IgG concentration, in the

same 22 samples, were poorly correlated (Pearson

r = 0.36), as were two measurements of IgM concen-

tration in 21 samples (Pearson r =� 0.15). Accord-

ingly, a method to increase the reliability of the

technique is described here. In this method, a copoly-

mer of lysine and phenylalanine was adsorbed to a

plastic petri dish, and thereafter the antigen was

covalently bound to the copolymer (Wood and

Gadow, 1983; Hobbs, 1989). This procedure

increased the reliability of DIG-ELISA so that for

IgG the Pearson r = 0.80 and for IgM the correlation

was r= 0.90.

2. Materials and methods

2.1. Reagents

1,4-Phenylenediamine was obtained from Aldrich

(Alcobendas, Madrid, Spain), 3,3V-diaminobenzidine

tetrahydrochloride dihydrate (DAB) was from Fluka

(Alcobendas, Madrid, Spain), and the following

reagents were from Sigma (Alcobendas, Madrid,

Spain): phosphate-buffered saline (PBS; 10 mM phos-

phate, 2.7 mM potassium chloride, 0.137 M sodium

chloride, pH 7.4), suberic acid bis(N-hydroxysuccini-

mide ester), poly(lysine, phenylalanine) hydrobromide

(lysine/phenylalanine = 1:1, molecular weight

20,000–50,000), ethanolamine hydrochloride, affin-

ity-purified goat antibodies to human IgG (g chain-

specific), affinity-purified goat antibodies to human

IgM (A chain-specific), peroxidase-conjugated affin-

ity-purified goat antibodies to human n light chain,

peroxidase-conjugated affinity-purified goat antibod-

ies to human E light chain, peroxidase-conjugated

affinity-purified goat antibodies to mouse IgG (g

chain-specific).

2.2. Animals and immunizations

Female mice of the CD1 strain, of approximately 3

months of age, were housed four mice per cage and

received food and water ad libitum; the temperature of

the room was 21F1 jC. The mice were injected

intraperitoneally with erythrocytes from a rat of the

Long Evans strain (4.8� 107 erythrocytes/mouse in

PBS). A booster immunization of 4.2� 107 erythro-

cytes was performed 10 days later. Plasma levels of

erythrocyte-specific IgG antibodies were measured, 8

days post-booster injection, by DIG-ELISA as modi-

fied in this report.

2.3. Human blood samples

Students, most of them women, from the Univer-

sity of Barcelona volunteered to donate a few drops of

blood, which were drawn from the fingertip by a nurse

in the campus health center. The blood, anticoagulated

with heparin and preserved with 0.01% thimerosal,

was kept in a refrigerator (4 jC) until assayed.

2.4. Enzyme-developed radial immunodiffusion tech-

nique (DIG-ELISA)

This technique was a modification of that

described by Elwing et al. (1980). For the measure-

ment of immunoglobulin content, polystyrene petri

dishes of 13.6-cm internal diameter (Nunc, Ros-

kilde, Denmark), were coated with poly(lysine,

phenylalanine) by overnight incubation with a 40

Ag/ml solution thereof in PBS, 20 ml/dish. After

three washes with water, lysine residues were

J. Vidal / Journal of Immunological Methods 270 (2002) 163–170164

activated for 15–30 min by adding 1 mM suberic

acid bis(N-hydroxysuccinimide ester), 20 ml/dish, in

a mixture of PBS and dimethyl sulfoxide (1.5:1, v/

v). The dish was then washed twice with a mixture

of PBS and dimethyl sulfoxide (2:1) and once with

PBS. Immediately thereafter, 20 ml of a solution of

antibodies to human IgM, or IgG, in PBS, 15 Ag/ml,

was added to each dish and kept there for 30

min at room temperature. After two washes with

PBS, unreacted residues were inactivated by the

addition of ethanolamine (1 mg/ml in PBS), 20 ml/

dish, for 30 min. The dishes were then washed

twice with PBS, once with PBS–Tween (0.05%

Tween 20 in PBS) and once with saline before

processing immediately or storage in the refrigerator

for up to 2 weeks. Twenty ml of 1% agar in PBS

containing gelatin (1 mg/ml, final concentration)

was added to each dish. While the gel was still

liquid, a template with 26 two-mm-diameter pins

was inserted and left there until the gel solidified.

When the template was removed, it left 26 regularly

distributed holes in the agar, the holes being 2 cm

apart. Each hole was filled with 5 Al of serum

diluted in PBS (addition of some trypan blue to

PBS yielded a deep blue solution, which facilitated

the filling of the holes). After diffusion time (1 day

for IgG, 2 days for IgM), the gel was removed, the

dish washed 3 times with PBS–Tween, and perox-

idase-labeled antibodies to human n and E light

chains were added (20 ml per dish of a 1:4000

dilution in PBS–Tween with gelatin, 4 mg/ml final

concentration). After a 2-h incubation the dish was

washed three times with PBS–Tween and the spots

were developed by silver intensification of DAB

spots (see below).

For the measurement of the concentration of

antibodies to rat erythrocytes, polystyrene petri

dishes (13.6-cm internal diameter; Nunc) were

coated with poly(lysine, phenylalanine) and acti-

vated with suberic acid bis(N-hydroxysuccinimide

ester) as described above. Each dish received 20 ml

of a 1% erythrocyte suspension (previously washed

three times with PBS) and was left undisturbed for

45–50 min at room temperature. Then, the eryth-

rocytes were carefully resuspended and discarded.

The erythrocyte monolayers were thoroughly

washed with saline and lysed by the addition of

PBS diluted 1:10 in water; the stroma layers were

repeatedly washed with saline until the washing

fluid no longer was pink. Inactivation of the

unreacted lysine residues was achieved with etha-

nolamine as described above. After inactivation,

peroxidase activity of the stromata was quenched

by incubation with 0.3% hydrogen peroxide in PBS

for 15 min (15 ml per dish) and three washes with

saline. Addition of agar gel and sample dilutions

was as above. After incubation, the gel was

removed and the plate incubated for 2 h with

peroxidase-labeled antibodies to mouse IgG. The

spots were developed by silver intensification of

DAB spots (see below).

2.5. Methods to reveal antibody-containing spots

Polystyrene petri dishes of 9-cm diameter, divided

in two equal halves by a plastic strip (Bibby Sterilin;

Stone, Staffs, United Kingdom), were used. Both

halves of each dish were coated with antibodies to

human IgM or IgG (by overnight incubation with a

solution of each in PBS (10 Ag/ml), 4 ml in each half),

washed with PBS–Tween, and 3 ml of 1% agar

containing gelatin (1 mg/ml) was laid on each half

of the dish. Thereafter, two holes were punched in

each half of the dish and filled with 4 Al of IgM

solution in PBS (5 Ag/ml or 100 Ag/ml) or with 4 Al ofIgG solution (0.5 Ag/ml or 20 Ag/ml). After a 1-day

incubation (for IgG), or a 2-day incubation (for IgM),

the gel was removed, the dishes washed with PBS–

Tween, and peroxidase-labeled antibodies to human nand E chains were added. After 1 h and 40 min, spots

in each half of the dish were revealed by one of the

two methods given below.

(a) With p-phenylenediamine: This is the method

described by Elwing et al. (1980). A mixture of warm

agar solution (1% in PBS), 1,4-phenylenediamine (1

mg/ml final concentration) and hydrogen peroxide

(0.01% final concentration) was added to one side

of the dish (5 ml/side), and the enzymatic reaction was

allowed to proceed for 35 min.

(b) By silver intensification of DAB spots: This is

the method described by Ludany et al. (1993) to

reveal spots on nitrocellulose membranes. A solution

of 3,3V-diaminobenzidine tetrahydrochloride dihydrate

(0.2 mg/ml), nickel chloride hexahydrate (0.5 mg/ml)

and hydrogen peroxide (0.002%) in 50 mM Tris–HCl

buffer, pH 7.6, was added to one side of the dish

J. Vidal / Journal of Immunological Methods 270 (2002) 163–170 165

(5 ml/side). After 20 min, the dish was washed three

times with distilled water. Thereafter, the physical

developer described by Ludany et al. (1993) was

added (5 ml/side); the reaction was allowed to con-

tinue for 10 min and was terminated by three washes

with distilled water. The physical developer was

prepared using 1 ml of ascorbic acid (2 mg/ml), 1

ml of sodium tungstate (50 mg/ml) and 8 ml of the

following mixture: sodium acetate (160 mg/ml), acetic

acid (7 Al/ml), silver nitrate (0.125 ml of 1% silver

nitrate/ml), cetylpyridinium chloride (0.0125 ml of

1% cetylpyridinium chloride/ml), Triton X-100

(0.075 ml of 1% Triton/ml).

3. Results

3.1. Spot development

In the studies described, development with DAB–

nickel and silver yielded sharper, and larger, spots

than development with p-phenylenediamine. The left

side of Fig. 1 shows the IgG spots obtained with

DAB–nickel–silver, while the right side shows the

spots with p-phenylenediamine after 35 min of color

development: The DAB–nickel–silver spots were

larger and better defined, especially the spot corre-

sponding to the lower IgG concentration (0.5 Ag/ml;

top spot).

The left side of Fig. 2 shows the IgM spots

obtained with DAB–nickel–silver, while the right

side shows the spots with p-phenylenediamine after

35 min of color development: Again, the DAB–

nickel–silver spots were larger and better defined,

especially the spot corresponding to the lower IgM

concentration (5 Ag/ml; top spot).

3.2. Parallelism of the regression lines of several sera

To be able to measure the antibody content of

several sera from a reference serum, a plot of

diameter vs. logarithm of dilution should yield par-

allel lines for all the sera. This was verified with the

improved DIG-ELISA described in this report: Five

dilutions of each of four human plasma samples

were assayed and the corresponding (IgG) diameters

were plotted against the logarithm (to the base 10) of

the dilutions. This was done in duplicate, in two

different petri dishes. Table 1 gives the intercept and

the slope of the regression lines. For each plasma

sample, the diameters were linearly related to the

logarithm of the dilutions (the correlation coefficient

of any regression line was z 0.990). Furthermore,

the regression lines were approximately parallel (test

for homogeneity of the slopes in dish 1: F(3,12) = 2.89,

p = 0.079; test in dish 2: F(3,12) = 1.11, p = 0.38), and

the slopes in dish 1 and dish 2 were of similar

magnitude (t test for repeated measures: t(3) = 0.097;

p = 0.92).

The same measurements were performed with

IgM, and the results were similar (results not

shown).

3.3. Reliability of the enzyme-developed radial

immunodiffusion technique (DIG-ELISA)

The IgM concentration of 21 human plasma sam-

ples was measured twice, each time in a different petri

dish, and the paired concentrations exhibited a poor

correlation (Pearson r =� 0.15; p>0.10). Similarly,

the IgG concentrations of 22 plasma samples were

also poorly correlated (r = 0.36; p>0.10). These meas-

urements were performed with the technique

described by Elwing et al. (1980), although the dis-

closure of spots was performed using the DAB-nickel-

silver method.

It was thought that this unreliability might be

caused by differing efficiencies of adsorption of

antigen on different dishes. To correct this, chemical

binding of antigen on to adsorbed poly(lysine,

phenylalanine) was attempted (Wood and Gadow,

1983; Hobbs, 1989). To check the reliability of the

IgG measurement, the IgG content of 19 human

plasmas was measured twice, each time on a diffe-

rent petri dish. The IgG concentrations (mg/ml)

(meanF standard error) were: 7.04F 0.58 in dish A,

and 7.37F 0.58 in dish B with a correlation

coefficient of 0.80 ( p < 0.0009). The intraassay and

interassay coefficients of variation could be estimated

from the reference sample and both were comparable

(7.5%).

To assess the reliability of the IgM measurements,

the same test was performed with 20 human plasma

samples. The IgM concentration (mg/ml; mean

F standard error) was 1.78F 0.21 in dish A and

2.08F 0.24 in dish B with a correlation coefficient

J. Vidal / Journal of Immunological Methods 270 (2002) 163–170166

of 0.94 ( p < 0.0009). The intraassay and interassay

coefficients of variation, estimated from the reference

sample, were similar (10%).

The reference sample could also be used to

estimate the sensitivity of the technique (i.e., the

minimal concentration of immunoglobulin that

Fig. 1. Development of IgG spots with diaminobenzidine–nickel–silver vs. development with p-phenylenediamine. Left side: Spots obtained

with DAB–nickel–silver; right side: spots obtained with p-phenylenediamine after 35 min of color development (development of DAB–

nickel– silver spots required 30 min). IgG concentrations, from top to bottom: 0.5 and 20 Ag/ml.

J. Vidal / Journal of Immunological Methods 270 (2002) 163–170 167

could be reliably measured) and this was 0.1 Ag/ml

for IgG and 1.1 Ag/ml for IgM.

3.4. Performance of the enzyme-developed radial

immunodiffusion technique (DIG-ELISA) with a

particulate antigen (erythrocytes)

As shown above and previously (Wood and

Gadow, 1983; Hobbs, 1989), soluble proteins can be

covalently bound to poly(lysine, phenylalanine)

although it has not been shown that particulate anti-

gens (e.g., cells) can also be bound. An experiment

was therefore conducted to ascertain whether the

technique could be used to measure antibodies to

erythrocytes. The plasma of five mice, immunized

twice with rat erythrocytes, was serially diluted and

the level of specific IgG antibodies was determined by

the technique described in this report. Table 2 shows

that (a) the diameters of the five dilutions of each

plasma sample were linearly related to the (base 10)

logarithm of the dilutions (lowest r>0.99; (b) the lines

corresponding to the five plasmas were approximately

Fig. 2. Development of IgM spots with diaminobenzidine–nickel–silver vs. development with p-phenylenediamine. Left side: Spots obtained

with DAB–nickel– silver; right side: spots obtained with p-phenylenediamine after 35 min of color development (development of DAB–

nickel– silver spots required 30 min). IgM concentrations, from top to bottom: 5 and 100 Ag/ml.

Table 1

Parallelism of IgG curves

Dish 1 Dish 2

Intercept Slope r Intercept Slope r

Plasma 1 25.53 4.06 0.996 25.84 4.15 0.990

Plasma 2 24.64 3.73 0.997 24.89 3.71 0.996

Plasma 3 22.94 3.57 0.999 23.47 3.69 0.999

Plasma 4 25.23 4.15 0.997 24.54 3.93 0.999

Five dilutions of four human plasma samples were assayed. The

IgG spots were revealed by the modified DIG-ELISA described in

this report and the spot diameters were plotted against the log (to the

base 10) of the dilutions. The assay was performed in duplicate, in

two different petri dishes.

Equations are of the form diameter (mm) = intercept+(slope*log10dilution).

r is the correlation coefficient of the regression line.

J. Vidal / Journal of Immunological Methods 270 (2002) 163–170168

parallel (test for homogeneity of the slopes in dish 1:

F(4, 15) = 2.35, p = 0.10; test in dish 2: F(4,

15) = 0.37, p = 0.83); (c) the slopes in dish 1 were

comparable to the slopes in dish 2 (t test for repeated

measures: t(4) = 1.40; p = 0.23).

4. Discussion

In the enzyme-developed radial immunodiffusion

technique (or DIG-ELISA), the development of anti-

body spots with DAB–nickel and intensification with

silver was found to offer some advantages over the

former development with p-phenylenediamine: (i) it

was more sensitive, i.e., it yielded larger spots, espe-

cially at low antigen concentrations (Figs. 1 and 2);

and (ii) it produced more distinct spots (Figs. 1 and 2),

which made measurement of diameters more precise.

One disadvantage of DAB over p-phenylenediamine

is its carcinogenic potential, but this fact should not

detract from its advantages in spot development.

In principle, DIG-ELISA in its former version

(Elwing et al., 1980) should be a reliable technique

and, therefore, it was surprising to find a poor corre-

lation between two measurements of immunoglobu-

lins on the same samples (correlation coefficients:

0.36 for IgG and � 0.15 for IgM). It was hypothe-

sized that a possible cause of the low reliability was

differing efficiencies of adsorption of the antigen on

each dish (Engvall, 1980; Parsens, 1981), and, if so, a

procedure yielding homogeneous binding should

solve the problem. One such procedure has been

reported (Wood and Gadow, 1983; Hobbs, 1989):

The antigen is chemically bound to poly(lysine, phe-

nylalanine) previously adsorbed on the petri dish. This

simple modification was indeed found to yield a

reliable assay. Hobbs (1989) activated the polyamino

acid with bis(sulfosuccinimidyl) suberate, whereas

Wood and Gadow (1983) used glutaraldehyde. In

the author’s experience, bis(sulfosuccinimidyl) suber-

ate worked better than glutaraldehyde, which agrees

with the results by Hobbs (1989). However, suberic

acid bis(N-hydroxysuccinimide ester) worked as well

as bis(sulfosuccinimidyl) suberate and was cheaper

(its insolubility in water was remedied by dissolving

the chemical in a mixture of PBS and dimethyl

sulfoxide, 1.5:1, v/v).

The modified DIG-ELISA described here permit-

ted the reliable measurement of IgG and IgM concen-

trations in serum, but, in addition, the same technique

also allowed the measurement of antibodies to cells.

The diameters of the spots produced by several

dilutions of serum (containing IgG antibodies to rat

erythrocytes) were linearly related to the logarithms of

the dilutions, and the regression lines from five sera

were parallel (Table 2 and Results).

Since DIG-ELISA is a time consuming technique

(z 2 days), should it not be more convenient to use

ELISA instead? To answer this question, one has to

consider the advantages and drawbacks of each tech-

nique. ELISA is faster and is more amenable to

automation, but DIG-ELISA has these features: (a) it

requires very simple equipment (for instance, there is

no need for a plate reader), which lends itself to use in

the field (e.g., in veterinary applications (Bautista-

Garfias et al., 1989; Ibarra et al., 1998) and in tropical

medicine (Kumar et al., 1987; Castilla et al., 1988));

(b) it allows low affinity antibodies to encounter an

excess of antigen (Gunnarsson and Svedhem, 1998);

(c) the reference and sample lines are parallel, which

makes it possible to read sample concentrations from

the reference line. In the author’s experience, and also

that of other’s (Gripenberg and Gripenberg, 1983;

Franc�ois-Gerard et al., 1988), lines are not always

parallel in the ELISA procedure. In general, ELISA is

the technique of choice if one has the appropriate

equipment and if the reference and sample lines are

parallel.

Table 2

DIG-ELISA can measure the level of IgG antibodies to rat

erythrocytes

Dish 1 Dish 2

Intercept Slope r Intercept Slope r

Plasma 1 18.20 3.84 0.999 18.52 3.62 0.994

Plasma 2 16.91 3.47 0.999 17.55 3.94 0.999

Plasma 3 15.18 3.38 0.995 15.83 3.48 0.999

Plasma 4 16.50 3.63 0.998 16.73 3.42 0.999

Plasma 5 15.84 3.69 0.999 17.00 3.55 0.998

Five mice were immunized with rat erythrocytes and the diameters

of the (anti-erythrocytes) IgG spots, revealed by the DIG-ELISA

described in this report, were plotted against the logarithm (to the

base 10) of the dilutions. The assay was done in duplicate in two

different petri dishes.

Equations are of the form diameter (mm) = intercept+(slope*log10dilution).

r is the correlation coefficient of the regression line.

J. Vidal / Journal of Immunological Methods 270 (2002) 163–170 169

In conclusion, DIG-ELISA has been improved in

two ways: size and sharpness of the spots have been

increased by development with diaminobenzidine–

nickel–silver, and the technique has been made reli-

able by chemically attaching the antigen to a poly-

amino acid (poly(lysine, pheylalanine)) adsorbed on

the petri dish.

Acknowledgements

This work was supported by a fellowship (Project

no. BSO 2000-0661) from the Spanish Ministry of

Education.

References

Bautista-Garfias, C.R., Lopez-Arellano, M.E., Sanchez-Albarran,

A., 1989. A new method for serodiagnosis of sheep fascioliasis

using helminth excretory–secretory products. Parasitol. Res. 76,

135.

Castilla, M.M., Santos-Gomez, M., Guzman-Bracho, C., Bautista-

Garfias, C.R., 1988. A new method for diagnosis of Chagas’

disease: diffusion-in-gel enzyme-linked immunosorbent assay. J.

Parasitol. 74, 805.

Cursons, R.T., 1982. DIG-ELISA for the serologic diagnosis of

toxoplasmosis. Am. J. Clin. Pathol. 77, 459.

Dematteis, S., Arburuas, G., Marco, M., Migues, M., Nieto, A.,

1989. Quantitative determination of anti-hydatid antibodies by

Elisa without a colorimetric reading. Int. J. Parasitol. 19, 229.

Elwing, H., Nygren, H., 1979. Diffusion-in-gel enzyme-linked im-

munosorbent assay (DIG-ELISA): a simple method for quanti-

tation of class-specific antibodies. J. Immunol. Methods 31, 101.

Elwing, H., Lange, S., Nygren, H., 1980. Diffusion-in-gel enzyme-

linked immunosorbent assay (DIG-ELISA): optimal conditions

for quantitation of antibodies. J. Immunol. Methods 39, 247.

Engvall, E., 1980. Enzyme immunoassay ELISA and EMIT. Meth-

ods Enzymol. 70, 419.

Franc�ois-Gerard, C., Gerard, P., Rentier, B., 1988. Elucidation of

non-parallel EIA curves. J. Immunol. Methods 111, 59.

Frigo, B., Scopsi, L., Patriarca, C., Rilke, F., 1991. Silver enhance-

ment of nickel-diaminobenzidine as applied to single and double

immunoperoxidase staining. Biotech. Histochem. 66, 159.

Gripenberg, M., Gripenberg, G., 1983. Expression of antibody ac-

tivity measured by ELISA. Anti-ssDNA antibody activity char-

acterized by the shape of the dose– response curve. J. Immunol.

Methods 62, 315.

Gunnarsson, H., Svedhem, A., 1998. The usefulness of diffusion-

in-gel-ELISA in clinical practice as illustrated by a Campylo-

bacter jejuni outbreak. J. Immunol. Methods 215, 135.

Hobbs, R.N., 1989. Solid-phase immunoassay of serum antibodies

to peptides. Covalent antigen binding to adsorbed phenylala-

nine– lysine copolymers. J. Immunol. Methods 117, 257.

Ibarra, F., Montenegro, N., Vera, Y., Boulard, C., Quiroz, H., Flores,

J., Ochoa, P., 1998. Comparison of three ELISA tests for seroe-

pidemiology of bovine fascioliosis. Vet. Parasitol. 77, 229.

Kumar, G.S., Mak, J.W., Lam, P.L., Tan, M.A., Lim, P.K., 1987.

Diffusion-in-gel enzyme-linked immunosorbent assay, ELISA

and IFA test in the detection of malarial antibodies. Southeast

Asian J. Trop. Med. Public Health 18, 502.

Ludany, A., Gallyas, F., Gaszner, B., Andrasfalvy, B., Szucs, G.,

Kellermayer, M., 1993. Skimmed-milk blocking improves silver

post-intensification of peroxidase–diaminobenzidine staining

on nitrocellulose membrane in immunoblotting. Electrophoresis

14, 78.

Merchenthaler, I., Stankovics, J., Gallyas, F., 1989. A highly sensi-

tive one-step method for silver intensification of the nickel–

diaminobenzidine endproduct of peroxidase reaction. J. Histo-

chem. Cytochem. 37, 1563.

Parsens, G.H., 1981. Antibody-coated plastic tubes in radioimmu-

noassay. Methods Enzymol. 73, 224.

Wood, W.G., Gadow, A., 1983. Immobilisation of antibodies and

antigens on macro solid phases—a comparison between adsorp-

tive and covalent binding. A critical study of macro solid phases

for use in immunoassay systems. Part I. J. Clin. Chem. Clin.

Biochem. 21, 789.

J. Vidal / Journal of Immunological Methods 270 (2002) 163–170170