I haven't done the experiment you mention, but I will point out that the sensitivity of the blot will be limited by the fact that plasma contains a huge amount of albumin protein, and a well in an ELISA plate can only hold so much protein. If you can remove the albumin first (there are special columns available to do this, although I don't know how well they work with pig albumin), you can load much more of the remaining protein and improve the sensitivity of the ELISA for the protein you are interested in.

I have a good experience with ELISAs and my paper where I developed ELISAs for brain transporter proteins is currently under review. From the knowledge I have about ELISAs, I can tell you that you should first try a couple of different dilutions on your ELISA plate. For e.g.: a 1:10, 1:50 or a 1:100 dilution of some test samples in the diluting buffer. An ideal dilution would be one where the corrected values fall within 25% to 75% range of your standard curve scale. You dont want your corrected optical densities or luminescence units after background subtraction to be at the extreme ends of your standard curve (close to 0 or the highest point).

Please feel free to ask any further related to ELISA development.

Good luck,

Chandana

Thanks Adam for pointing out the interference from albumin. Our kits do not indicate any need for purification, so I'm not sure this is going to be the first thing we try, but I'll keep this in mind as an alternative.

Chandana, thanks for the tip about the dilutions. We first diluted our samples to the same total protein concentration, and then diluted 1:2 and 1:10, since the protein we're looking at is in low abundance normally. After all our analysis and following the manufacturer's instructions, we ended up getting NEGATIVE concentrations! Do you have any tips for trouble shooting this? More dilutions? This seemed to give higher optical densities in our output. Also, our standard dilution only went to 0.156 ng/mL, would further serial dilution (to 0.078ng/ml and lower) help in determining these concentrations?

Thanks for your help!

Hi Ivana

Yes, the way you are doing is also valid, where in you first normalize all samples to the same protein concentration and then dilute either 1:10 or 1:2.

The way I did was, I started with the same weight of tissue across different tissue samples. Then, I did different dilutions of the samples to check and see what dilution falls within 25%-75% of the standard curve. I also did a BCA assay to determine the protein concentration of the samples. Once I got the protein concentration of the specific antigen from the ELISA data, I divided that with the total protein concentration to appreciate the actual differences of the specific antigen between samples.

Yes, you are right. If your samples after background subtraction are negative, then you should extend the lower end of your standard curve to determine the lower limit of quantification (LLOQ) of that assay. To do this, I ran 3 sets of standard curves in triplicate on two different plates, calculated the % CV (a sister of SD) and % recovery for each set of the standard curve. If % CV (SD/Mean*100) is 10% or 15% and % recovery (observed conc/true conc *100) is between 80%-120%, then that lowest point on the standard curve (across all standard curve sets) that obeyed the rules you set for %CV and % recovery cut-off was considered the LLOQ. Again, you can determine if you want the % recovery to be 80-120% or 90-110% and for the % CV to be below 10% or below 15% or below 20% (which is the max).

It is also possible that your antigen is not completely soluble in the sample diluent. Adding 0.1% SDS or 0.1% Triton-X-100 to the sample and standard diluent might help solubilize your antigen better, thereby help interact with the capture/detection Abs. When doing this, I would first test just the standard curve, with or without SDS/Triton-X-100, and see how the signal improves and then extend it to the samples, at which point you should test the standard too.

Another possible explanation is that your antigen may not be completely extracted while processing the plasma. This happened with my antigen. I used a number of different methods (RIPA/SDS/Triton) to extract the antigen. SDS worked very well for my antigen. Mine was a transporter protein, which is a transmembrane protein. Because it is not a soluble protein, but a membrane protein, it needed 10% SDS during the sample processing step to completely solubilize the protein.

Also, unless the kit that you purchased from the company has passed good quality control, and is a reliable company (like Innotest or Covance), ELISA may not work, just because the company sells. I had a not-so-good experience with one company out in the market, and ended up developing my own. Sometimes, companies sell products with a part of the recombinant protein (fragment), rather than using the full length protein, whose folding will be different from that of the full-length antigen in the sample.

I hope these tips help improve the ELISAs. Always, please  feel free to contact if you have any further questions.

Best wishes,

Chandana