The sensitivity (lower detectzion limit) depends only on the affinity of the solid phase antibody according to the law of mass action. Therefore you should search for a hiogh affinity antibody. 

With most other tricks you arbe able to decrese the loewer detection limit only by factor 2 - 5.

Some tricks: 

1. Increse the incubation time (ever over night

2. add Polyethylene glycole (up to 1 ... 3%) to the incubation buffer

3. Amplification by using biotinylated secodary antibodies followed by streptavidin-HRP. SA has four binding sites for biotin (but you are doing this already)

I agree with Serafeim, increase the length of your incubations. Also, I think it is odd that changing the dilutions of coating and detection antibody did not have any result. Increasing coating concentrations usually increases sensitivity, as does increasing detection antibody (although that may lead to increased background). Also, try different blockers. Pierce sells a streptavidin;HRP polymer that is supposed to increase the sensitivity (i.e. 25 HRP molecules attached to each streptavidin, versus 4 )

Have you defrosted the serum samples before thawing them for the ELISA? If these samples have been defrosted and then refrozen they will not work to give you a higher concentration, it doesn't matter how hard you try! Why not purchase a normal ELISA kit for human serum samples and do not use primary and secondary antibodies. I have always used the HumanTNF-alpha High Sensitivity ELISA from eBioscience (BMS223HS) and have always had great results without having to worry about primary and secondary antibodies.

The sensitivity (lower detectzion limit) depends only on the affinity of the solid phase antibody according to the law of mass action. Therefore you should search for a hiogh affinity antibody. 

With most other tricks you arbe able to decrese the loewer detection limit only by factor 2 - 5.

Some tricks: 

1. Increse the incubation time (ever over night

2. add Polyethylene glycole (up to 1 ... 3%) to the incubation buffer

3. Amplification by using biotinylated secodary antibodies followed by streptavidin-HRP. SA has four binding sites for biotin (but you are doing this already)

though I understand your need to optimize tests.... I have to say that this did not work in moost cases. There are hundreds of factors to consider with ELISAs. Working for 15 years in this field and many years in the technical support for these kind of questions I have seen always the same picture. Researchers try to optimize kits which tends to be cheaper by direct coating and mixing reagents. But all failed with this attempt. Starting with the plate, the buffers up to antigens/antibodies and enzymes it is a high complex procedure with a lot of raw material and in process and final release testings inside an immunoassay company. You may have one good shot, but the next time you will see that it is not reproducible... and in the end you have wasted a lot of time and money and just have unsecure results.

I hardly recommend to look for a good ELISA company, test several kits and ask for a good price for selected ones. Then you will enjoy working with these kits and create reliable results      

Serafeim, thanks, I will definitely give a shot to o/n incubations.

Evan, right, I should have said "no positive results" because I got a high background with no increased sensitivity thus I rejected it. Pierce poly-SA sounds interesting, have you tried it yourself or compared it with other SA? I use SA purchased form Jackson Immunoresearch and I am satisfied with its results so far.

Maritza, yes, I am fully aware of the freeze-thaw cycles of the samples but I am 100% sure that the samples work fine because I tested them at other ELISAs (and I also tried with fresh samples). A kit would be a nice but I need to test 100 (4  dilutions each) samples, so imagine the number of the kits I actually need!!

Stephan, I will try PEG, but one question: which PEG (molecular weight) do you use? Does it matter? I have one of ~3000 M.W. and another of ~6000 M.W. Thank you.

I would recommend incubating your antigen overnight at 4 degree. I have also used the human TNF-alpha kit from Peprotech. They have mini kit for this target for $100. Might be worth a try. They also carry a lot of the individual antibodies to engineer a kit if it works.

High background could also be because you are working with serum. Sometimes blocking proteins in serum can inhibit your target from binding. and why elisas may work for cell and tissue lysates and not serum.

I suggest reviewing the technology used for TNF-a EASIA (Enzyme Amplified Sensitivity Immunoassay (EASIA) kit. The assay is based on oligoclonal system in which a blend of monoclonal antibodies directed against distinct epitopes of TNF-a are used. The kit sold by the life technologies and there is link  for the kit manual :

http://tools.lifetechnologies.com/content/sfs/manuals/KAC1751_Direct_ELISA.pdf

Incubating your sample overnight at 4°C is a good idea for serum/plasma. Secondary antibody incubation can be a bit longer as well, up to 2.5 hrs. I would also recommend all incubations up to & including *avidin be done at 4°C, on an orbital shaker.

Have you tried an indirect ELISA "checkerboard" titration for antibody activity? Coat rows of a plate with 100, 50, 25, 12.5, 6.25, ... pg/mL peptide, then block, and detect with columns of 1:2000, 1:1000, 1:500, ... dilutions of antibody. Do this for primary & secondary, to get an idea of their individual sensitivities. If you can't distinguish 5 pg/mL from 10 pg/mL in an indirect ELISA, you may not be able to see this on a sandwich ELISA either.

If, after all possible optimizations, you still have low signal, you may still be able to use this for an assay! Check to make sure background is low, also that your sample signals are interpolatable into the signal range of your standard curve. If you don't have linearity at the bottom end of your standard curve, check out 5-PL curve fitting. There's a great free tool at www.readerfit.com which I used to use quite regularly.

Really the points on which you can optimize an ELISA are virtually endless... Happy experimenting!!

Hello Spylla Tsiomita,

I have tried polyHRP-SA from Pierce, and it does improve the sensitivity. It works by amplifying the signal due to the presence of over 400HRPs/polymer-SA. I would also suggest you to use 1-step Ultra TMB substrate solution, instead of OPD or other substrate as substrate. I found an improvement in sensitivity for IFN-g detection upto 5 pg/mL, by applying this two components. Here are links where you can get the product. 

https://www.lifetechnologies.com/order/catalog/product/21140

https://www.lifetechnologies.com/order/catalog/product/34028

Hope this helps.

Vigneshwaran Mani thank you for your reply. I have already tried polyHRP-SA from Pierce but unfortunately I didn't see any significant changes in sensitivity plus it gave me high background. In which dilution did you use it? Working buffer as usual (PBS with 1% BSA and 0.1% Tween)? As for TMB I, too, use one ready-made from KPL, similar to the one you describe, with generally great resutls so TMB isn't the problem! Thanks again!

Hi Spylla, I had tried 1:5000/10000 PolyHRP-SA in 1-2%BSA with 0.05% tween-20. I recommend you to wash atleast 5x PBS-T, and 1x with PBS after incubation with Strep-PolyHRP. Washing is key. I had done this on plate surface (Maxisorp) as well as magnetic bead surface. Btw what is your blocking buffer, Did you try higher BSA conc. like 5% in PBS? 

Dear Spylla,

PEG 6000 is fine, 3000 will work as well. But again, only the affinity of the solid phase antiobody defines the detection limit. Just think about the law of mass action. It's the base also for immune reacions.  All the other items mentioned are from non-optimized ELISA. 

The best TMB is available from Seramun: http://www.seramun.com/products/substrates/horseradish-peroxidase.html

http://www.seramun.com/products/substrates/horseradish-peroxidase.html

Try incubating at 37 degrees with shaking on a plate shaker (600 rpm).

The higher temperature and shaking increase the chance your antibody and cytokine find each other resulting in higher sensitivity.