In House ELISA - how to increase sensitivity? (Feb/09/2011 )

Hey guys,

I'm in the process of developing a cytokine sandwich ELISA with monoclonals we have generated. I did a BIAcore experiment and found a few paired antibodies and am now struggling to increase sensitivity into the pg/ml range. The assay maxes out (>2.0AU) at concentrations ~50ng/ml and then sharply drops to 0.5AU at ~6.5ng/ml. In fact, the absorbance dropped exactly by half each dilution (as I'm diluting my antigen 1:2). I'm not quite sure what I can do to increase sensitivity 1000fold. Most of the commercial kits I've seen have very linear AU readings from 1000pg/ml to 6.25pg/ml, (from like 1.5AU to 0.2AU). Could it be that my antibodies just don't have the affinity to detect cytokine at such low levels? I'm currently running a checkerboard but I doubt it would increase sensitivity 5 fold, let alone 1000fold...

Any help or comments would be appreciated!

EDIT: Also my background appears to be quite high, which I suppose can be some reactivity of the detection antibody with the capture!

The abs you have may not be sensitive enough. Things that will increase binding sensitivity (push cuve to the left): increase incubation time, increase incubation temperature, addition of polymers to assay buffer (sample/buffer/ab coated plate rx), and mixing plate during incubation steps.

If you are trying to duplicate an existing commercial kit and can purchase one...you can substitute your ab in their system to see how far off you are.

High background and low sensivity can be caused by too much Abs, both coating and detection. You could test a serial dilution of both Abs in a grid experiment and see if going lower with the concentrations doesn't help.

SebS on Wed Feb 9 10:36:25 2011 said:

Hey,

thanks for the prompt responses. Unfortunately there isn't a commercial kit available so I have no real reference point. What do you mean when you say I can try incubating with polymers? What polymers for example? I do have the worry that my antibodies aren't good enough to get into the pg/ml range... Also, should I always dilute my antigen of interest in blocking buffer? Is 3% BSA/PBS too much?

What concentration of cytokine should I test in a checkerboard? My boss has the opinion I should try 10ng/ml (which I have been able to detect in the past). I think that's way too much considering the assay maxes out at concentrations not much higher than that. I'd be tempted to try 500pg/ml and see if I can detect that...

OK. You have no reference kit. Biacore exp shows some promising pairings.

Since you need to bind in the pg area I would use serial 10 fold dilutions of antigen for screening and start in the pg/ml area ie 10pg/ml. If a short incubation time is not critical then increase it to several hours (seal the plates!). If you have a plate shaker then use that as well. The polymers (ie PEG) will increase binding...however, you may see some increase also in NSB. This chemical is usually used in the buffer for diluting ag (ie ag-ab binding step). I imagine you are screening one mab with one mab...ie a matched pair. You may also try using 2 mabs for capture or detection. (as an aside take a look at test kits for human TnI and you will see they use multiple mabs since they also require sensitivity in pg/ml range)

Your bsa/pbs is fine 1-2%BSA/PBS is fine. Your ab coating can be carbonate buffer at hight pH (ie pH 8-9)

One other trick is to "shock" your ab before coating...ie open up the structure. This requires glycine buffer low pH dilute ab in this 1st allow to sit for 10-15 min then add it to your coating buffer at higher M and pH.

Hey, thanks for the reply.

I performed a checkerboard and found that when I use very high concentrations of coating and detection (10ug/ml and 5ug/ml respectively) using a combination of 2 mAbs for coating I max out the readings within minutes for 8ng/ml. I realise this is very high but my boss wouldn't let me try any lower. I'll try diluting out antigen next and also have a 0pg/ml control to look at background. In the last experiment I only had 1 well with no antigen and background was almost 0 so I'll see how far I can push this!

Thanks for the advice and I'll report my findings!

Maxing out the reading at 8ng/ml and your blank is 0...run that curve in the pg/ml area 0.1, 1, 10, 100, 1000, 10,000 pg/ml and,you just may see your high sensitivity dose response curve.