Dear Chelsey,

I just realized your request and the answers above are really facing major aspects correctly. I would just  like to emphazise another aspect of your ELISA. For every immunoassay the intended use is important. This aspect is highly relevant in terms of the second aspect mentioned by Hans Lutz. If you like to recognize in a serological format those sera, which have high affinity antibodies against the virus antigen (So the "patients" had a "clear and targeted" immune response.), your assay should perform different compared to the situation where you want to see weak infection status with only low affinity antibodies. (But some of those "individuals" don't have or had a relevant infection and don't need any treatment of whatsoever. Most of them had only contact with a cross-reacting antigen and not with the virus, you like to investigate.). If you are interested in a serological assay, which clearly and without any doubt detects only sera after infection and after a real response from the immune system, you should develop your assay in a way that low to medium affinity binding does not occur at all. This makes a serological assay highly reliable, but on the other hand not sensitive during the first few days of infection.

Low to medium affinity binding is the major cause for cross-reactivities. And so if you optimize your assay to show only high affinity binding effects and not low to medium affinity binding - which is done by choosing the assay diluent - you will most probably get rid of any cross-reactivities.

But if you like to have an assay, which also gives a signal from the weakest immune response, that would mean, that your assay will have to show also low affinity binding effects. In that case you will not get rid of cross-reactivities and other potential interference effects, but you will have positive results even in earlier stages and with weaker immune responses (maybe false-positives due to cross-reactitivies of unknown molecular reasons). In other words: a serological assay either gives nearly no false-positives and is very reliable to detect a strong infection reaction, but you maybe loose the first days of infection as false-negatives. Or your assay will detect all potential infections, but will show many false-positives due to cross-reactivities and interference. Both assay types are good and well-done, but it is important to differentiate and to understand which assay gives the one or the other kind or answer. For many diseases and situations the weak immune response is not relevant. Unfortunately most kit manufacturers and also scientist do not realize and mention this difference in the intended use and the answer of an assay.

Again: It is a question of your research task, whether you should avoid potential interference due to cross-reactivity and similar effects or not. You can avoid it by using interference-reducing assay diluents or antibody diluents. But this is not always good to do so.

Furthermore I would like to add a comment to the point 1 of Hans Lutz, which is in general correct, but it maybe doesn't fix the complexity of coating to ELISA plates. If you coat molecules on an ELISA plate you will normally not get a monolayer of molecules, which are "directed" like polar or non-polar. In fact you normally get a completely uncontrollable and complex layer of molecules, which are in part monolayer in part double- or multilayered and which are oriented in any thinkable and non-thinkable direction. As a result only very few of the coated molecules (regardless whether you coat proteins or peptides or antibodies in other kinds of ELISA) will be presented in a way that binding to other molecules like antibodies will be possible. For example in an extremely good optimized sandwich ELISA only 4-8 % of all coated capture antibodies can bind to the target molecule. All the other coating antibodies are either denatured or misoriented or both. So it is not the case that your capture molecules on an ELISA plate will be completely oriented according to the polarity of plate and molecule. You will also see all other orientations and in most cases even multilayer and complexes of molecules. There are only few options to change this situation, which is difficult to control. One option is choosing different pH values during the coating process. You will get very good results in many cases with a Coating Buffer at pH 9.6, but sometimes pH 7.2 (or maybe others) will give you different results in your ELISA. Normally you can't predict the outcome. Just test different pH-values during coating and you will sometimes get different results in your assay.

So first decide, which kind of question your assay should answer and then optimize your assay to answer exactly this question (avoiding of low to medium affinity binding or not, ...). You should keep in mind that you can get two completely different assays with the same antigens and antibodies, if you change aspects like coating procedure, assay diluent, incubation times.

Best wishes

Tobias

Dear Chelsey,

There may be 2 reasons for the discrepancies you have observed:

1. The 1. reason lies in the fact that in all  ELISA procedures  antigens are not displayed in the same way as in IFA. When you coat your antigens to the polystyrene surface in carbonate buffer, the antigens will mostly be separated from their original 3-dimensional context and will bind to the plastic by non-convalent attachment by their non-polar part. The polar part is directed toward the lumen of the well where antibodies can readily bind to their epitopes. In IFA, only relatively few  epitopes are accessible to antibodies as they may still be in the context of the original position within the cells. Fixation of the cells may not completely make all epitopes accessible to antibodies.

2. The 2nd reason can be found in the affinity and the "polyclonality" of the antibodies. Early on in an infection, you may expect an antibody response characterized by presence of antibodies to a wide variety of epitopes with relatively low affinity. During the course of infection the polyclonality decreases while at the same time the affinity of the antibodies increases. Low affinity antibodies are not readily picked up by ELISA but by by IFA. High affinity antibodies -which are not removed by vigorous washing- are  easily detected by ELISA but may be missed by IFA due to lack of accessibility of the epitopes.

Thus, both types of discrepancies (ELISA pos, IFA neg and vice versa) can be observed . ELISA techniques based on well defined rec. antigens are therefore ideally suited to produce reproducible results.         

Best wishes, Hans

I adhere to the comments made by Hans Lutz; this phenomena is observed in many in-vitro assays. We published a way to analyse such discrepencies in data (see: Saez-Alquézar A et al. (2000) J Clin Microbiol. Feb;38(2):851-4).

Best wishes,

Maan

I am in general agreement with Mr. Hans Lutz. However, if I may add something, according to our experience with ELISA, low affinity immune complexes might be easily detected if your ELISA is directly detecting complexed antibodies since they involve numerous antibodies. On the other hand, if your ELISA is detecting serum blocking activity, low affinity immune complexes might pass detection since they are not efficiently blocked as they also involve numerous epitopes. Your ELISA is showing cross reactivity, I grasp, increase antigen dose is one way of getting rid of cross reactivity (without loss of sensitivity). Increase antigen dose might also increase test sensitivity since low content of high avid antibodies, in sera of mixed population of antibodies, are apt to react more preferentially. Of course increase antigen dose above certain levels decrease sensitivity and that would be observable.

Thank you for all of your feedback, this really helps!

Dear Chelsey,

I just realized your request and the answers above are really facing major aspects correctly. I would just  like to emphazise another aspect of your ELISA. For every immunoassay the intended use is important. This aspect is highly relevant in terms of the second aspect mentioned by Hans Lutz. If you like to recognize in a serological format those sera, which have high affinity antibodies against the virus antigen (So the "patients" had a "clear and targeted" immune response.), your assay should perform different compared to the situation where you want to see weak infection status with only low affinity antibodies. (But some of those "individuals" don't have or had a relevant infection and don't need any treatment of whatsoever. Most of them had only contact with a cross-reacting antigen and not with the virus, you like to investigate.). If you are interested in a serological assay, which clearly and without any doubt detects only sera after infection and after a real response from the immune system, you should develop your assay in a way that low to medium affinity binding does not occur at all. This makes a serological assay highly reliable, but on the other hand not sensitive during the first few days of infection.

Low to medium affinity binding is the major cause for cross-reactivities. And so if you optimize your assay to show only high affinity binding effects and not low to medium affinity binding - which is done by choosing the assay diluent - you will most probably get rid of any cross-reactivities.

But if you like to have an assay, which also gives a signal from the weakest immune response, that would mean, that your assay will have to show also low affinity binding effects. In that case you will not get rid of cross-reactivities and other potential interference effects, but you will have positive results even in earlier stages and with weaker immune responses (maybe false-positives due to cross-reactitivies of unknown molecular reasons). In other words: a serological assay either gives nearly no false-positives and is very reliable to detect a strong infection reaction, but you maybe loose the first days of infection as false-negatives. Or your assay will detect all potential infections, but will show many false-positives due to cross-reactivities and interference. Both assay types are good and well-done, but it is important to differentiate and to understand which assay gives the one or the other kind or answer. For many diseases and situations the weak immune response is not relevant. Unfortunately most kit manufacturers and also scientist do not realize and mention this difference in the intended use and the answer of an assay.

Again: It is a question of your research task, whether you should avoid potential interference due to cross-reactivity and similar effects or not. You can avoid it by using interference-reducing assay diluents or antibody diluents. But this is not always good to do so.

Furthermore I would like to add a comment to the point 1 of Hans Lutz, which is in general correct, but it maybe doesn't fix the complexity of coating to ELISA plates. If you coat molecules on an ELISA plate you will normally not get a monolayer of molecules, which are "directed" like polar or non-polar. In fact you normally get a completely uncontrollable and complex layer of molecules, which are in part monolayer in part double- or multilayered and which are oriented in any thinkable and non-thinkable direction. As a result only very few of the coated molecules (regardless whether you coat proteins or peptides or antibodies in other kinds of ELISA) will be presented in a way that binding to other molecules like antibodies will be possible. For example in an extremely good optimized sandwich ELISA only 4-8 % of all coated capture antibodies can bind to the target molecule. All the other coating antibodies are either denatured or misoriented or both. So it is not the case that your capture molecules on an ELISA plate will be completely oriented according to the polarity of plate and molecule. You will also see all other orientations and in most cases even multilayer and complexes of molecules. There are only few options to change this situation, which is difficult to control. One option is choosing different pH values during the coating process. You will get very good results in many cases with a Coating Buffer at pH 9.6, but sometimes pH 7.2 (or maybe others) will give you different results in your ELISA. Normally you can't predict the outcome. Just test different pH-values during coating and you will sometimes get different results in your assay.

So first decide, which kind of question your assay should answer and then optimize your assay to answer exactly this question (avoiding of low to medium affinity binding or not, ...). You should keep in mind that you can get two completely different assays with the same antigens and antibodies, if you change aspects like coating procedure, assay diluent, incubation times.

Best wishes

Tobias