Question for the lab geeks out there regarding general methods for antibody detection of specific analytes in cells, serum, and urine.
Do you have a favorite book, chapter, or any other reading material that you would use to guide students through designing and validating their own ELISA for a serum or urine protein? If you have any lecture slides you use and could share, I'd be happy to credit you. Some famous guy(s) gave me his H1N1 slides last year and helped me look like a genius.
I know that it takes a mighty, mighty fine antibody to do this relative to one for an immunoblot or even an in-cell western.
Of course, I know that one of my advisors would tell me to just buy a damn kit but that is the wrong answer for my current needs.
And no, I don't have the money for a Luminex xMAP thingie just yet.
For those of you who wonder what the heck I'm talking about, ELISA stands for enzyme-linked immunosorbent assay. It's a method that uses a highly-specific antibody to measure stuff in any biological matrix - applications include HIV testing, PSA levels, and even home pregnancy tests. Most of the time ELISAs are used to measure proteins but they can also sometimes be used to measure small drug molecules such as the cardiac glycoside, digoxin (a drug taken by someone with congestive heart failure).
In fact, here is a link to a nice animation I found on how a home pregnancy test works.
The Virtual Immunology Lab at the Howard Hughes Medical Institute (HHMI) site is neat but I'm wanting to present more technical detail on assay development and validation.
In the meantime, thank you for any advice.
Photo credit: New England Biolabs
- Log in to post comments
Seriously, just buy the pregnancy test from the store it's much easier and cheaper and.... oh... never mind: http://scienceblogs.com/terrasig/2008/02/liveblogging_the_vasectomy_chr…
So are you counting making antibodies as the first step, or are commercial ones available? Where are you getting antigen from, to make the antibody and then later to use for a standard?
If the antibody is not one you're experienced with, you'd probably see how it performs on a Western before getting down to the ELISA. You want to see the specificity in the particular biological sample you want to test. I think you're right on needing a better antibody than for immunoblotting, but I haven't the faintest idea what the minimum affinities for each method actually are (it obviously could depend a bit on washing conditions and such). I know with commercial antibodies those that are advertised as suitable for ELISA are much rarer than those that work for WB.
@Chemgeek - You are a very humorous man. Slightly off-topic, you will not be surprised to know that the positive pregnancy test that gave rise to the PharmKid sits in her baby book together with her mitotic chromosome spread when we were screened for Down's syndrome.
@becca - We've already talked about generating monoclonals and polyclonals and how antibody quality for ELISAs must be much, much better than for Westerns. The capture technique (using two antibodies that recognize different epitopes on the same target) seems to allow one to get away with using substandard antibodies - kind of the ELISA equivalent of an IP-Western.
The NIH Chemical Genomics Center has some good protocols. Their protocol for in-cell westerns has been helpful for us at times. Here's their one for ELISAs:
http://www.ncgc.nih.gov/guidance/section10.html
The best place to start is with the Ligand Binding Society Recommendations http://www.springerlink.com/content/g04w745374114736/
The recommendations provide a complete pathway for determining accuracy and precision for a given ELISA as well as freeze thaw stability of analyte and dilutional linearity. If you are working in development for a CRO you will most likely be given a monoclonal and a polyclonal from the sponsor. Your job is to then do some prelim development experiments to determine orientation and appropriate standard curve range and then apply the recommendations to the assay (validation).
Helpful Hint: if you are running a sandwich with mouse or anti-mouse Ab s, always use 1% mouse serum in the sample and conjugate diluent. Your signal to noise ratio will thank me.
Sorry dude. I'd love to help with ELISAs but I doolittle of them.
Check your mail.
For a book recommendation: Immunochemical Techniques Laboratory Manual by John Goers. Older, but very very solid instructions. Can be successfully used by even the stupidest student.
You are detecting a protein, yes? Not a small molecule? Proteins are easier.
In fact the antibodies for ELISA no longer have to be so perfect. The fluorescent tags have improved mightily over the past decade, you can get AlexaFluor conjugated on just about anything. Much better than AP. Do you have any guesses on how much of this protein you might normally find? That will help determine what the detection tag needs to be. I'd use normal fluorescent tags down to, oh, no less than high nM, lower than that you can use fluorescent beads down to about femtomolar. Lower than that and you will have to attach an oligo to the detection Ab and RT-PCR for quantitation. That can be done down to about 10 particles/well.
If you are working in development for a CRO you will most likely be given a monoclonal and a polyclonal from the sponsor