Jump to content

  • Log in with Facebook Log in with Twitter Log in with Windows Live Log In with Google      Sign In   
  • Create Account

Submit your paper to J Biol Methods today!
Photo
- - - - -

high slope (b-value) of standard curve

standard curve

  • Please log in to reply
11 replies to this topic

#1 elisa2013

elisa2013

    member

  • Active Members
  • Pip
  • 7 posts
0
Neutral

Posted 05 July 2013 - 07:20 AM

I am running an elisa using Bethyl IgG-Fc capture antibody @ 10ug/ml coating (CBC coating buffer per Bethyl), using a medium binding plate. The standard is a Bethyl IgG human reference serum. I am starting the dilution of the std at 500ng per the protocol & diluting 2-fold. The secondary IgG aby dilution is 1:15,000 (all 3 components come as a kit). I am coating the rest of the plate with a specific protein for which I am trying to quantify antibody amounts and running random serums in them.
The OD of the IgG STD start at about 3.0, but once they come to the well that the conc. is at 125ng (3rd well), the ODs drop way past 2-fold & continue to drop fast resulting in a high slope of the standard curve & very little "plottable" points. The ODs for the wells coated with the specific protein (random serums run) dilute out just fine, it's only the STD column that is giving the issue, so I can't quantitate until this is resolved.
I have tried:
-starting the IgG STD at 250ng & 1000ng, the same thing happens once the dilution of it hits 125ng, no matter the starting conc.
-today I am trying a high binding plate with various 2nd ABY conc.
-today I am trying the same medium bind plates and increasing the 2nd ABY diln from 15,000 to 5,000

Why is this only happening in the capture ABY/IgG STD column & not the specific protein coated wells? I am using the same conc. of 2nd aby across the plate.
I am not getting background in my blank wells.

#2 elisa2013

elisa2013

    member

  • Active Members
  • Pip
  • 7 posts
0
Neutral

Posted 05 July 2013 - 10:41 AM

Just read the results using the high-binding plates. The ODs are much better, not getting the huge decrease is ODs, but slopes are still higher than desired (1.5ish). Is this just a balance of an antigen/aby amount available to bind ratio? or maybe too much or not enough incubation times, but at which step?
thanks for any insight

#3 Ben Lomond

Ben Lomond

    Enthusiast

  • Active Members
  • PipPipPipPipPip
  • 75 posts
6
Neutral

Posted 07 July 2013 - 04:41 AM

I am having difficulty in understanding what you are trying to measure and what the format of the assay is.

In order to compare quantities of unknown against your human IgG standard curve, the entire plate should be coated with the same thing (i.e bethyl anti-IgG Fc or your specific protein, but not both). It seems that you are running two different assays on one plate:

assay 1 (curve): capture with anti-human IgG Fc, human IgG standard, detect with another anti-human IgG-HRP (or other conjugate)

assay 2 (samples): capture with specific protein (presumably some viral or vaccine antigen), samples containing unknown titres/quantities of anti-antigen antibodies, detect the antibodies with the anti-human IgG-HRP

As the two formats rely on different coating efficiencies and binding kinetics, it is not surprising that the two curves exhibit lack of parallelism (can not be super-imposed).

Unless I am missing something, the entire plate should be coated with your specific protein, and you need a reference serum (or purified standard) containing or consisting of human anti-antigen IgG to use as your standard.

Let us know how you get on, or if you need any more help with this assay

good luck!

#4 elisa2013

elisa2013

    member

  • Active Members
  • Pip
  • 7 posts
0
Neutral

Posted 15 July 2013 - 10:00 AM

Yes, we are coating with 2 different proteins, an IgG capture aby for the STD curve column (a human IgG STD serum w/ known amount of IgG) & specific protein in the other columns- we are only looking for a relative IgG concentration in compared to the amount of expected IgG in the STD serum, not necessarily specific.
Anyway, with the high bind plates, I can get the STD curve to come down half-way decent for the IgG STD with slopes about 1.5 (no range errors given in SMPro, but higher b values than i'm used to seeing), but the serum samples run in the wells coated with specific protein have really high CVs. We are diluting 2-fold & the starting ODs are about 2.5ish, but fall slowly, making the adjusted results in SMPro have a huge range of titers & creating high CVs (40-60%).
I am using a block step before adding primary aby on all wells, 30min. Dilution buffer & secondary aby dilution buffer is the same (1xpbs/0.5%bsa/0.005%tween). Primary inc. is 2hours, secondary aby inc. is 1 hour. I am using hrp/tmb for 30 min substrate-is that too long? could that be causing oversaturation?
When ODs fall slower than the expected 2-fold dilution made, what is going wrong and at which step?
thank you.

#5 Missle

Missle

    Veteran

  • Active Members
  • PipPipPipPipPipPipPipPipPipPip
  • 132 posts
18
Good

Posted 15 July 2013 - 12:27 PM

In my experience 30 minutes is too long for an hrp/tmb incubation. Your strong signals will likely start to precipitate by that time and will give you variability. Also, if the incubation isn't happening in the dark, you'll start to get some non-specific background.

As an aside, the standard curve really isn't valid if you change the capture protein between your sample and the std(especially where the standard curve is relying on the affinity of the capture antibody for antibody in the serum and your sample is relying on the affinity of the antibody in the sample for the specific antigen on the plate) . It may be able to tell you there is little antibody versus a lot of antibody but no form of quantitation will be accurate - not even semi-quantitative. It isn't comparing apples to apples - more like apples to carrots :)

Good luck!

#6 Missle

Missle

    Veteran

  • Active Members
  • PipPipPipPipPipPipPipPipPipPip
  • 132 posts
18
Good

Posted 15 July 2013 - 12:35 PM

Sorry, I didn't answer the slow fall in OD part. I don't think it has to do with the length of incubations but rather the amount of antigen on the plate or the starting concentration of antibody in your sample. The slow fall in OD is similar to the plateau effect where there is an excess of antibody. Either your specific protein isn't coating well on the plate (I don't know if you have a control for that or how much you're coating) or there is too much specific antibody in your sample. If you're confident that your specific protein is binding to the plate well then I would try starting the antibody sample at a higher dilution and see if you can get a better signal drop.

#7 Ben Lomond

Ben Lomond

    Enthusiast

  • Active Members
  • PipPipPipPipPip
  • 75 posts
6
Neutral

Posted 15 July 2013 - 03:18 PM

As both Missile and I are suggesting, you are running two assays on one plate...it makes little sense to compare the results of your assay with the specific protein coat to a curve with a different coat for the reasons we have both indicated. If you require only relative results...create a pool of serum containing your antibodies for your specific protein, use this to prepare your curve with arbitrary units, and use this as a reference for moving forward.

When ODs fall, what is 'slower than expected'? Each assay format will behave differently depending on the binding affinities and avidities of the antibody/antigen interaction and the curve will fall slower with one format than another. When you say high CVs, do you mean high CVs between replicates of the same sample, or CVs generated from back calculated results at different dilutions of the same sample? If your replicate wells give the same ODs when loaded with the same sample at the same dilution, there is hope for the assay, just stick with the specific protein coat and find a way to standardize the curve so you can compare like with like.

For TMB, as Missile suggests, 30 minutes before stopping TMB is a little on the long side from my experience too.

Good luck

#8 almost a doctor

almost a doctor

    Veteran

  • Active Members
  • PipPipPipPipPipPipPipPipPipPip
  • 244 posts
15
Good

Posted 16 July 2013 - 12:19 AM

In addition to what Ben Lomond and Missle said (which I totally agree), I have a question. What is your "IgG STD"? I mean, if you are using this as some form of positive standard for the amount of specific IgG in your samples, then (as said before) you should coat the whole plate with your protein, and use that serum (in a dilution series) to make a standard curve. If, on the contrary, the "IgG human reference serum" is more likely a negative serum you are going about the assay completely wrong. The way to measure antibody titres (although not absolute amount) is to compare to your negative serum. The protocol will be something like:
1) coat plate with antigen (every well same protein same concentration).
2) block.
3) prepare a dilution series of all samples (1:100, 1:200, 1:400, 1:800, 1:1600, 1:3200...) INCLUDING NEGATIVE SERUM (IgG STD)
4) incubate plate with serum dilutions (in replicates)
5) wash
6) incubate with secondary antibody (same dilution for whole plate).
7) wash
8) add substrate (here, just to add that as mentioned before 30min seems too long for TMB)
9) read the plate

The antibody titre for each sample, will be the lowest dilution (for that sample) that gives signal above background (negative serum signal), we use to set 3x background as our cut off (ie, anything lower than 3x background was considered negative).

Found this in a very quick google search: http://www.elisa-ant...SA-applications I'm sure there are plenty more references.


Hope this helps, if this is not what you are trying to do I've completely misread your posts.

#9 sgt4boston

sgt4boston

    Veteran

  • Active Members
  • PipPipPipPipPipPipPipPipPipPip
  • 276 posts
1
Neutral

Posted 19 July 2013 - 11:40 AM

Boy is this topic confusing!
Follow the suggestions above

Method/Test1...capture antibody and optimize that test with standards/controls/calibrators and conjugate. Once you get a curve then run your unknows.

method/Test2....Separately, use your binding protein as capture along with standards/controls/calibrators etc etc and do the same thing.

Once you have these 2 tests optimized then results can be compared.

#10 elisa2013

elisa2013

    member

  • Active Members
  • Pip
  • 7 posts
0
Neutral

Posted 23 July 2013 - 10:50 AM

What if you don't have any standards/cntls/calibrators with a known amount of specific aby to the binding protein? How do you establish a STD when you are starting from scratch & the only thing you have to compare to is a commercial STD with a known amount of IgG (not specific). I'm not sure why the protocol was set up the way is was, never made much sense to me.
We have this: 1) a specific protein (I'll call it "X")
2) serum samples for which we are trying to find an amount of IgG specific for our protein
3) a commercial quantitation kit that has a Reference serum with a known amount of IgG (not specific for "X") , an IgG coating antibody, & a secondary IgG hrp-aby
What we DON'T have is this:
a sample to use as a STD/REF curve that has a known amount of aby specific for "X"
I get what Vetetan is saying for Method/Test1...if I coat with our commercial kit of capture IgG & run the Reference serum that comes with it as the STD curve with an assigned conc., that will give me the general IgG conc. for serums that we run on that plate & from that I can assign one of them as a new STD or control with an assigned conc. if I wish
When I get to step 2 & coat the plate with "X", what am I using as the STD for the curve and what conc. do I assign it for the curve? Am I still using the Reference serum from the kit with it's assigned IgG conc.?
sorry this is so confusing, I'm used to establishing assays with some sort of known STD to work with first.

#11 Ben Lomond

Ben Lomond

    Enthusiast

  • Active Members
  • PipPipPipPipPip
  • 75 posts
6
Neutral

Posted 23 July 2013 - 06:15 PM

So I am presuming that you wish to estimate the quantity of antibodies specific to your protein x, which is either a pathogen/ vaccine / biological therapeutic. Each individual will mount an immune response to protein X in his/her own individual way so the antibodies from each individual will bind to protein X with varying affinity/avidity, so your results will always be relative to whatever standard you adopt and not truly quantitative.

It may be possible to affinity purify specific antibodies from positive serum samples using a protein X affinity matrix followed by acid elution of the specific antibodies and quantify the eluted antibodies and create a standard curve from that. You could use this as your own reference which you will always refer back to. Alternatively, you could generate a polyclonal anti protein X antibody in a different species (and this may be acceptable in certain disciplines but your detection antibody would have to cross react with human and poly species IgG) or generate an engineered human mAb or panel of human mAbs specific for protein x. Several providers can do this for you, one you might look into is HuCal antibodies from Abd Serotec if funds permit.

With that said, there are published guidelines for immunogenicity testing of therapeutics/vaccines etc, which take into consideration many factors including antigen tolerance (your assay won't work if you have antigen X in the sample binding to the anti-antigen X antibodies). Relevant guidelines for anti-therapeutic antibodies can be found by googling 'Immunogenicity testing guidelines - Millipore', I am sure you can find similar guidelines for vaccines/pathogens/or other...time spent becoming familiar with the specific approaches for your particular discipline would be very useful before moving forward.

I think all here are agreed that the human IgG quantification kit is not helpful in this endeavour.

good luck

#12 elisa2013

elisa2013

    member

  • Active Members
  • Pip
  • 7 posts
0
Neutral

Posted 25 July 2013 - 10:53 AM

Thank you, yes, your paragraph 1 statement is what we are trying to accomplish, only a relative measurement, not really quantitative. I will bring your comments up with the scientist to re-think our approach.





Home - About - Terms of Service - Privacy - Contact Us

©1999-2013 Protocol Online, All rights reserved.