ELISA sample titration effects

4 posts / 0 new
Last post
Becky
Becky's picture
ELISA sample titration effects

We have developed many in-house sandwich-type ELISAs for human cytokine, chemokines, ECM, etc. All work beautifully for cell culture supernatant fluids. We need to measure analytes in bronchoalveolar lavage fluid and concentrated the sample 10X (using centrifugal filter unit with 3kDA cutoff). Expecting that some samples might be over the standard curve, all were run neat and also diluted 1:5 (2X conc) in PBS/ 1% albumin/ 0.05% tween.  Indeed some of the 10X samples were over the curve, but analyte was undetectible at 1:5 dilution.  I  expected that the diltued samples would give higher value (dilution of an inhibitor, better fit to the standard curve, etc), but I can't explain the total disappearance of the analytes with dilution. This occured for some ELISAs and not others and wasn't specific to the species or isotype of coating or detecting ab. I'm in a crunch to publish these data, but don't want to publish an artifact. Any ideas why we would see a dilution effect and how to determine if signal is specific to the analyte?
Thanks for any suggestions.

Sami Tuomivaara
Sami Tuomivaara's picture
Becky,

Becky,

One thing that I can suggest is to try the dilution without albumin. One major function of albumin is to bind various molecules in serum to stabilize them... 1% albumin can saturate your molecules that are present no doubt in very small amounts in your samples.

Cheers,

Becky
Becky's picture
Thanks for the suggestion

Thanks for the suggestion Suola. Will give it a try. 

Edit Sat 10/3: my apologies to Suola and anyone who is following this thread.  I originally asked Suola why albumin would increase the signal, hence her reply below. 

Latter, realizing my question didn't make sense....the signal was decreased not increased, I edited the response. I had the site open at the time and did not realize that Suola had responded to the reply.  My apologies for the confusion, I am new at "blogging".  My thanks to Suola for the suggestions.  I will pass the information along to my technician and give  the  BAL ELISA a small test run (these human sample are quite precious) without albumin and/or without detergent. 

Just for clarification, the odd ditution effect was seen in  3-4 of  72 samples that were run and only in 5 of 6 different ELISA assays.  The  remaining samples had to be run at 10X conc for analytes to be detected. We have 10+ years experience with these assays and have not seen the dilution effect before now.  The samples with the "dilution effect" were from subjects with very inflammed airways and the samples had very high levels of protein.  Hopefully the test without albumin and/or detergent will solve the problem. Thanks again to Suola for the responses.

Sami Tuomivaara
Sami Tuomivaara's picture
Becky,

Becky,

Adding albumin would not increase the signal, it would decrease it by saturating the binding capability of your target molecules, so they cannot bind antibodies anymore -> no ELISA signal. That is why I suggested to drop albumin from the dilution buffer altogether.

I didn't realize that bronchoalveolar fluid has lot of albumin in it... maybe dropping albumin from the dilution media wouldn't have big effect then.

I've done lot of ELISAs and never used any detergent, maybe you can try dilution without tween as well.

Why do you have to do the concentration/dilution cycle in the first place... if your signal is a bit weak with the native sample, use more sample per well, increase the color development time, lots of different things you can try to get higher signal. It could be beneficial to avoid concentration steps since you can lose some protein in the membrane, etc... All sample manipulations that can be avoided, should be avoided.

Cheers,