I think that the most significant part in ELISA technique is to find a protein (antibody) which reacts with the protein you want to detect. The procedure is very simple: you add a cerebrospinal fluid sample (the axonal protein is considered as an antigen) in the wells of ELISA plates, then you add the first antibody which reacts with the antigen, then you add the second antibody which is attached to an enzyme (such as alkaline phosphatase) and reacts with the first antibody. Finally, you add a chromogenic substrate which reacts with the enzyme and gives a colored product... If the antibody is present in the wells, you will see them colored... Don't forget to add an inert protein (such as BSA) in order to "cover" the wells where the antigen has not been adsorpted... Also you have to use a buffer solution so that you can wash out the antigen and the second antibody which have not been adsorpted or reacted.
I have the manufacturer's protocol and I need to optimise it to my project. So should I determine what are the best capture antibody, secondary antibody + enzyme and chromogenic substrate to use and,if needed, alter the quantities in the original protocol?
I am not sure if I got your question write... Why do you need "elaborate" the ELISA, there are no kit available? The quantification provided by the ELISA test is based in a standard and purified protein. Once you have a stable protein where you can be sure of the concentration, you can use the method you are trying to develop to quantify your samples based on that standard protein. Another point is how the color (colorimetric test) intensity is related to the concentration, and this relation is usually not linear. That is why is very useful use a kit that was tested for those conditions. And you shouldn't extrapolate points out of the curve, meaning not having concentrations that are not covered by the standard curve. In order to to that, you should run a first ELISA test with a few dilutions (1:10, 1:100, 1:500, for example). Until you find the dilution that give a result that is covered by the standard curve. After that, you will have an idea of the dilution factor necessary. I also worked with a high concentrated sample and after a few preliminar trials, I standardized and always run my samples in 2 concentrations (1:250 and 1:500). In this way, the results are always covered by the standard curve... Good luck!