A New End-Point for ELISA Titrations
JOURNAL OF IMMUNOASSAY&IMMUNOCHEMISTRY
Vol.25,No.4,pp.371–383,2004
A New End-Point for ELISA Titrations
Jose′Vidal*
Departamento de Personalidad,Facultad de Psicolog?′a,Universidad de
Barcelona,Barcelona,Spain
ABSTRACT
This report describes a new ELISA procedure based on end-point
titrations.This end-point ELISA takes advantage of the change of color
intensity that occurs when peroxidase-containing wells of an ELISA
plate are revealed with diaminobenzidine–nickel and further intensi?ca-
tion with silver:as antibody concentration and,therefore,peroxidase
concentration,decreased,the color became stronger in some wells
and,afterwards(i.e.,at lower antibody and peroxidase concentrations),
the color faded toward clear background.It is proposed that the reciprocal
of the sample dilution at which the color intensi?es can be used as a
measure of the sample antibody content.This report veri?es the validity
and precision of that procedure.
Key Words:ELISA;End-point;Silver enhancement;Diaminobenzi-
dine;Poly(lys,phe);Immunoglobulin measurement.
*Correspondence:Jose′Vidal,Departamento de Personalidad,Facultad de Psicolog?′a, Universidad de Barcelona,Passeig de la Vall d’Hebron171,08035Barcelona,Spain; E-mail:jvidal@https://www.wendangku.net/doc/6217227318.html,.
371
DOI:10.1081/IAS-2000338391532-1819(Print);1532-4230(Online) Copyright#2004by Marcel Dekker,https://www.wendangku.net/doc/6217227318.html,
INTRODUCTION
Diffusion-in-gel enzyme-linked immunosorbent assay (DIG-ELISA)[1]is a simple,cheap,and sensitive procedure to measure the antibody con-centration in a sample.In this technique,a layer of gel (agar or agarose)is deposited onto a polystyrene surface that has antigen adsorbed on it,then holes are bored in the gel and they are ?lled with antibody-containing solution;after several hours,the gel is removed and antibody-containing areas are visualized enzymatically.The original technique [1]was modi?ed in order to increase its sensitivity and reliability.[2]Nevertheless,the technique still has a drawback,namely,performance takes a long time (24hr to measure IgG concentration and 48hr to measure IgM concentration).Those long times are due to the slow diffusion of IgG and IgM antibodies in the gel.
When antibody-containing areas,in DIG-ELISA,were revealed with an excess of peroxidase conjugate and the substrate diaminobenzidine–nickel–silver,[2]those areas had the aspect shown in Fig.1,i.e.,a circular spot
of Figure 1.DIG-ELISA revealed with diaminobenzidine–nickel–silver.Murine serum,containing IgG antibodies to rat erythrocytes diffused,in an agar layer,over polystyrene with erythrocyte stromata bound to it.Antibody-containing areas were disclosed with (an excess of)peroxidase-conjugated anti-IgG antibodies and diamino-benzidine plus nickel,with color intensi?cation with silver.A ring of stronger color can be seen in the periphery of the spots.
Vidal
372
New End-Point for ELISA Titrations373 pale color was surrounded by a ring of stronger color.Because the antibody concentration decreased from the center of the spot toward the periphery, at a certain(low)antibody concentration,the color became more intense, and thereafter,i.e.,at lower antibody concentrations,the color faded away.In principle,this increase in color intensity would take place also in the wells of an ELISA plate that contains a serial dilution of antibodies,and the intensi?cation of color could be used as an end-point.Consequently, one could perform an enzyme-linked immunosorbent assay(ELISA)and quantitate the results by the increase in color intensity,since the dilution corresponding to the color increase would re?ect the antibody content. In this manner,there would be no need to use a spectrophotometer,and ELISA would become simpler and cheaper;besides,this end-point ELISA would be faster than DIG-ELISA.
This report shows that the end-point ELISA provides an approximate estimate of the antibody content of a sample,and that the procedure is relatively precise.
EXPERIMENTAL
Reagents
3,30-Diaminobenzidine tetrahydrochloride dihydrate(DAB)was from Fluka(Alcobendas,Madrid,Spain),and the following reagents were from Sigma(Alcobendas,Madrid,Spain):glutaraldehyde(25%aqueous solution), phosphate–buffered saline(PBS;10mM phosphate, 2.7mM potassium chloride,0.137M sodium chloride,pH7.4),poly(lysine,phenylalanine) hydrobromide(lysine:phenylalanine?1:1,molecular weight20,000–50,000),ethanolamine hydrochloride,human IgG(reagent grade),human IgM (reagent grade),af?nity-puri?ed goat antibodies to human IgG(g chain-speci?c),af?nity-puri?ed goat antibodies to human IgM(m chain-speci?c), peroxidase-conjugated af?nity-puri?ed goat antibodies to human k light chain,peroxidase-conjugated af?nity-puri?ed goat antibodies to human l light chain,peroxidase-conjugated af?nity-puri?ed goat antibodies to mouse IgG(g chain-speci?c)and peroxidase-conjugated af?nity-puri?ed goat antibodies to mouse IgM(m chain-speci?c).
Animals and Immunizations
Ten female mice of the RjOrl:Swiss strain,approximately3months of age,were purchased from Janvier Espan?a,Madrid,Spain;the mice were
housed 4per cage and received food and water ad libitum;the temperature of the room was 218C +18C.The mice were injected intraperitoneally with varying numbers of erythrocytes from a rat of the Long Evans strain:two mice were injected with 1?107erythrocytes,two mice with 5?107erythrocytes,three mice with 1?108erythrocytes,and three mice with
1.5?108erythrocytes;the red cells were suspended in PBS and each mouse received 0.1mL of the suspension.Each mouse was boosted,16days later,with the same dose of erythrocytes,and was bled 5days after the booster.The mice were bled from the tail vein.
DIG-ELISA
This technique was performed as previously described [2]with two modi-?cations:(i)erythrocyte stromata were bound to poly(lysine,phenylalanine)by means of glutaraldehyde [instead of suberic acid bis (N -hydroxysuccin-imide ester)],and (ii)the concentration of the silver reagent used to intensify the spots was one half of the concentration used before.In brief,the procedure was as follows:polystyrene petri dishes of 13.6cm in internal diameter were coated with poly(lysine,phenylalanine),40m g /mL,and rat erythrocytes (20mL /dish of a 1%suspension in saline)were allowed to settle;after three washes with saline,bound red cells were lysed,and the stromata were linked to poly (lysine,phenylalanine)with 0.1%glutaraldehyde in PBS (20mL /dish)for 30min;peroxidase activity of the stroma was quenched with 0.3%hydrogen peroxide,unreacted glutaraldehyde residues were quenched with a mixture of ethanolamine and gelatin,and loosely bound material was detached by a wash with PBS–Tween (0.05%Tween 20in PBS).After addition of agar and serum samples,diffusion took place for 1day (for IgG),or 2days (for IgM);afterwards,peroxidase label (peroxidase–anti-IgG antibodies or peroxidase–anti-IgM antibodies)was added and areas of antibody content were revealed with diaminobenzidine–nickel reagent with further intensi?cation with silver reagent.
End-Point ELISA
End-point ELISA was used to measure concentrations of human IgM,or human IgG,in samples thereof,and to estimate the content of antibodies to rat erythrocytes in mouse sera.To measure concentrations of human immuno-globulins,antibodies to human IgM,or IgG,were bound to 384-well Maxi-Sorp plates (Nunc,Roskilde,Denmark)by Hobbs’procedure;[3]the plates were incubated overnight with poly (lysine,phenylalanine),40m g /mL in Vidal 374
New End-Point for ELISA Titrations375 PBS,25m L/well,washed three times with water,and incubated with0.5% glutaraldehyde in PBS(25m L/well)for30min,washed twice with water and once with saline,incubated with human IgM(15m g/mL in PBS),or IgG(15m g/mL in PBS),25m L/well,for60min,washed twice with saline, incubated with ethanolamine hydrochloride(1mg/mL in PBS)plus gelatin (5mg/mL),50m L/well,for30min,washed twice with saline and once with PBS–Tween.
To measure concentrations of antibodies to rat erythrocytes,red-blood-cell-stromata were linked to384-well plates as follows:plates were coated with poly(lysine,phenylalanine)as above,a1%suspension of erythrocytes in saline was added,25m L/well,and allowed to settle for30–40min;there-after,the unbound erythrocytes were washed away with saline and bound erythrocytes were lysed with PBS:water(1:10,v/v).After several washes with PBS:water1:10(until the washing?uid was no longer pink),0.1% glutaraldehyde in PBS was added,25m L/well,and kept there for30min; after three washes with saline,the wells were incubated for15min with 0.3%hydrogen peroxide in PBS,25m L/well.After three washes with saline,the wells were incubated for30min with ethanolamine hydrochloride (1mg/mL in PBS)plus gelatin(5mg/mL),50m L/well;the plates were washed twice with saline and once with PBS–Tween.
Samples(containing human IgM,human IgG,or antibodies to rat eryth-rocytes)were serially diluted in the wells of the ELISA plate;dilutions,by a factor of1.3,were done in PBS–Tween with10mg/mL gelatin.Each well received25m L of?uid.After90min,the plates were washed three times with PBS–Tween,and the appropriate peroxidase conjugate was added(for human immunoglobulins,anti-human kappa light chaintanti-human lambda light chain antibodies;for murine antibodies,anti-mouse IgM or anti-mouse IgG antibodies);after approximately90min,wells containing antibodies were revealed with diaminobenzidine–nickel and further treatment with a silver developer(this is the procedure described by Luda′ny et al.[4]with a minor modi?cation:the concentration of the silver developer was one half of that used by Luda′ny et al.).The incubation time for diaminobenzidine–nickel was20–30min,whereas the time for silver enhancement was10–15min.
A reference sample(a solution of human immunoglobulins of known concentration or a murine serum rich in antibodies to rat erythrocytes)was run together with the test samples,and results are expressed as the quotient: antibody concentration in the test sample/antibody concentration in the reference sample(this quotient equals the quotient:reciprocal of test-sample dilution when color increases/reciprocal of reference-sample dilution when color increases).
The working dilution of each peroxidase conjugate has to be found by cross titration,since the working concentration of the enzyme conjugate is
usually higher than the one suggested by the manufacturer for ELISA work (a concentration that is too low does not produce the color increase mentioned in the Introduction).
Statistics
Straight lines were ?tted to scatter-plots by a least-squares method.Correlation coef?cients are Pearson product-moment coef?cients.The statis-tical software used was Statistica 6.0(Tulsa,OK).
RESULTS
The Basic Phenomenon
A solution of human IgG (10mg /mL),and a solution of human IgM (1mg /mL),were titrated by ELISA as described above,dilutions being from 1:10,000to 1:20,480,000for IgG (wells B1–B12in Fig.2),and from 1:25to 1:51,200for IgM (wells E1–E12).Figure 2shows the plate after development:an increase in color intensity is observed at dilution 1:80,000for IgG (well B4)and at dilution 1:1600for IgM (well E7).This assay was performed on a 96-well plates because the wells of a 384-well plates were too small and too close together to yield a good-quality picture.
Rows C and F represent assays run on wells coated with goat albumin:no color is observed in any well,which shows the absence of nonspeci?c adsorp-tion by IgM,IgG,or the peroxidase conjugate (at the dilutions used).Wells D1,D2,and D3were coated with antibodies to human IgG,but did not receive IgG:no color is perceived in those wells.Similarly,wells G1,G2,G3were coated with antibodies to human IgM did not receive IgM,and did not develop color.Wells D4,D5,D6were coated with IgG,and wells G4,G5,G6were coated with IgM,but did not receive the peroxidase conjugate:no color is observed in those wells.
Validity of End-Point ELISA as a Procedure to Measure
Antibody Concentration
The abrupt change in color that takes place with dilution could be used as an end-point for the measurement of the antibody content in a sample.Vidal 376
This idea was tested thus:several samples were prepared by diluting a human IgM solution (1mg /mL),and the IgM concentration of the ensuing samples was determined by end-point ELISA (the undiluted solution was run in the plate as the reference sample).The upper part of Fig.3shows a plot of the estimated concentration vs.the actual concentration:the equation of the line was (measured IgM)?20.004t1.190?(actual IgM);the standard errors of the intercept and the slope were 0.005and 0.050,the correlation coef?cient was 0.98,and the number of points was 17.The difference between the intercept (20.004)and 0did not reach statistical
signi?cance Figure 2.The basic phenomenon.Row B:serial dilution of a human IgG solution (10mg /mL),from 1:10,000(well B1)to 1:20,480,000(well B12);the wells had anti-IgG antibodies bound to them.Row C:serial dilution as for row B,but the wells had goat albumin bound to them.Wells D1–D3:wells with anti-IgG antibodies bound to them,but did not receive IgG.Row E:serial dilution of a human IgM sample (1mg /mL),from 1:25(well E1)to 1:51,200(well E12);the wells had anti-IgM antibodies bound to them.Row F:serial dilution of the IgM sample on wells coated with goat albumin.G1–G3:wells coated with anti-IgM antibodies and not incubated with IgM.Wells D4,D5,D6were coated with IgG,and wells G4,G5,G6were coated with IgM,but did not receive the peroxidase conjugate.The plate was developed with peroxidase and diaminobenzidine–nickel,with further intensi?ca-tion of color with silver.Color intensi?cation is observed at dilution 1:80,000for IgG (well B4),and at dilution 1:1600for IgM (well E7).
New End-Point for ELISA Titrations 377
(95%con?dence limits:20.014to0.006),and the slope(1.190)was slightly higher than1(95%con?dence limits:1.083–1.298).
The same validity test was performed with the IgG sample:the equation
of the line was(measured IgG)?25.79t1.35?(actual IgG).The lower part
New End-Point for ELISA Titrations379 of Fig.3shows the plot of the estimated concentrations(as log10)vs.the actual concentrations(as log10):the equation of the line was(log measured IgG)?0.13t0.97?(log actual IgG);the standard errors of the intercept and the slope were0.04and0.02,the correlation coef?cient was0.99,and the number of points was16.The intercept(0.13)was slightly above0 (95%con?dence limits:0.05–0.22)and the slope(0.97)was not different from1(95%con?dence limits:0.92–1.03).
The above results showed a good linear correlation between the actual immunoglobulin concentration and the concentration measured with the ELISA described here.Yet,all the samples had antibodies of the same af?nity, because the samples were obtained by diluting the same original solution. Therefore,it was necessary to verify that the same linear association between measured concentrations and actual concentrations of immuno-globulins held for different sera.To do this,the anti-rat-erythrocytes antibody content of10mouse sera was measured with an established method(DIG-ELISA)and with the end-point ELISA described here.Figure4shows linear plots for both IgM and IgG.For IgM,the equation of the line was (IgM by end-point ELISA)?0.71t0.54?(IgM by DIG-ELISA)and the correlation coef?cient was0.87.For IgG,the equation was(IgG by end-point ELISA)?0.03t0.63?(IgG by DIG-ELISA);the correlation coef?-cient was0.91.
Reliability of End-Point ELISA
To assess the intra-assay reliability,the IgM,or IgG,concentration of four replicates of a human serum was measured,in the same plate,by end-point ELISA as described above;the measurement was again repeated with two more plates.The root mean square coef?cient of variation,a pooled coef?cient of variation from several assays,[5]was12%for IgM and ,1%for IgG.
To assess the inter-assay reliability,the IgM,or the IgG,concentration of a human serum was measured,by end-point ELISA,in four different plates. The root mean square coef?cient of variation was31%for IgM and12% for IgG.
The sensitivity of the procedure(the antibody concentration at which color intensi?cation took place)could be estimated from the human serum used for reliability assessment:for IgM,the mean(+standard error of the mean)of four determinations was1.5(+0.6)m g/mL,whereas,for IgG, the mean(+standard error of the mean)of four determinations was0.10 (+0.02)m g/mL.
New End-Point for ELISA Titrations381
DISCUSSION
In this report,it is proposed that the intensi?cation of color that took place at certain antibody dilution,in plates developed with diaminobenzidine–nickel–silver,could be used as an end-point for antibody titration.This implies that the method should be accurate(i.e.,that the estimated antibody concentration is close to the actual concentration)and precise(i.e.,that its coef?cient of variation is not too high).This report intends to verify both conditions.
Accuracy was tested by dilution experiments:several samples were created by diluting a solution of human immunoglobulin(IgM or IgG)of known concentration,their immunoglobulin concentration was estimated by end-point ELISA as described in this report,and the estimated values were plotted against the actual concentrations(Fig.3).Ideally,the intercept and the slope of the ensuing straight line should be,respectively,0and1. For the IgM line,the intercept(20.004)was not different from0(Results) and the slope(1.19)was close to1(95%con?dence interval:1.08–1.29). For the IgG line,estimated and actual concentrations were transformed into logarithms(to the base10)in order to dampen the in?uence of the higher concentrations in the regression(the concentration interval was from0to 3000m g/mL).After the transformation,the intercept(0.13)was close to0 (95%con?dence interval:0.05–0.22)and the slope(0.97)was not different from1(95%con?dence interval:0.92–1.03).
Besides testing the accuracy of a new method,a comparison with other current methods is necessary.Consequently,the end-point method reported here was compared with DIG-ELISA(an established method);[1,2]for the measurement of the level of anti-rat erythrocytes antibodies(Fig.4).The correlation coef?cients for IgM and IgG were,respectively,0.87and0.91. These coef?cients are comparable with the ones between DIG-ELISA and ELISA[r?0.84,[6]],DIG-ELISA and hemagglutination[r%0.90,[7]; r?0.89,[8]],DIG-ELISA and immuno?uorescence[r%0.80,[7]; r?0.90,[8]],DIG-ELISA and complement?xation[r?0.83,[8]],ELISA and Farr assay and immuno?uorescence[0.53,r,0.85].[9]
The use of384-well plates has some advantages over the use of96-well plates:(i)more samples can be processed in a plate,and(ii)the dilution factor may be lowered from2to1.3,thus achieving a more continuous distribution of antibodies in a row(if dilutions were done by a1.3factor in a96-well plates,each sample would take up a lot of wells and only a few samples could be processed in a plate).
The end-point procedure reported here has two advantages over ELISA: (i)it is simpler and cheaper(because it does not require a plate reader),and(ii) it may work when the test sample and the reference sample yield non-parallel
lines in a plot of absorbance vs.dilution (because the end-point does not depend on the color content of the wells).Therefore,the end-point method described here is appropriate for ?eld studies,laboratories with a low budget,and situations where antibody content is to be correlated with a less precise variable (e.g.,some behaviors).
ABBREVIATIONS
ELISA
Enzyme-linked immunosorbent assay DAB
3,30-Diaminobenzidine tetrahydrochloride dihydrate PBS
Phosphate-buffered saline DIG-ELISA Diffusion-in-gel enzyme-linked immunosorbent assay
ACKNOWLEDGMENT
This work was supported by a fellowship (Project No.BSO 2000-0661)from the Spanish Ministry of Education.
REFERENCES
1.Elwing,H.;Lange,S.;Nygren,H.Diffusion-in-gel enzyme-linked immunosorbent assay (DIG-ELISA):optimal conditions for quantitation of antibodies.J.Immunol.Meth.1980,39,247–256.
2.Vidal,J.Improvements to the enzyme-developed radial immunodiffusion technique.J.Immunol.Meth.2002,270,163–170.
3.Hobbs,R.N.Solid-phase immunoassay of serum antibodies to peptides.Covalent antigen binding to adsorbed phenylalanine–lysine copolymers.J.Immunol.Meth.1989,117,257–266.
4.Luda
′ny, A.;Gallyas, F.;Gaszner, B.;Andra ′sfalvy, B.;Szu ¨cs,G.;Kellermayer,M.Skimmed-milk blocking improves silver post-intensi?ca-tion of peroxidase–diaminobenzidine staining on nitrocellulose membrane in immunoblotting.Electrophoresis 1993,14,78–80.
5.Davies,C.Concepts.In The Immunoassay Handbook ,2nd Ed.;Wild,D.,Ed.;Nature Publishing Group:London,2001;78–107.
6.Berntsson,https://www.wendangku.net/doc/6217227318.html,parison between enzyme-linked immunosorbent assay (ELISA)and diffusion-in-gel enzyme-linked immunosorbent assay (DIG-ELISA)for detection of antibodies to pneumococcal polysacchar-ides.J.Immunol.Meth.1984,73,226–22
7.
Vidal 382
New End-Point for ELISA Titrations383 7.Cursons,R.T.M.DIG-ELISA for the serologic diagnosis of toxoplasmosis.
Am.J.Clin.Pathol.1982,77,459–461.
8.Requejo,H.I.Z.;Nakamura,P.M.;Vaz, A.J.;Pialarissi, C.S.M.;
Hoshino-Shimizu,S.;Matsumoto,T.K.;Nakamura,H.Diffusion-in-gel enzyme-linked immunosorbent assay(DIG-ELISA)for Chagas disease serodiagnosis.Braz.J.Med.Biol.Res.1991,24,471–483.
9.Smeenk,R.A comparison of four different anti-DNA assays.In Immunoassay
Technology;Pal,S.B.,Ed.;Walter de Gruyter:Berlin,1986;Vol.2,145–166. Received March23,2004
Accepted May14,2004
Manuscript3134