[Histonet] RE: Confirming IHC staining as "real"

[C.M. van der Loos]

Hi All,
Interesting problem we have here. Feels a bit like the chicken and the egg problem!
I will try to shed some light in it.
 
The procedure to complex an antibody with its original antigen or peptide is to my opinion a technique from an old dusty box. A technique from the time we were still dealing with POLYCLONAL antibodies that were not so monospecific and sometimes needed to be absorbed to reduce background staining. Doing such a blocking experiment with a MONOCLONAL antibody is however quite useless. Once the peptide blocks your staining for 100%, you end up with a feeling of a "good antibody" whereas one that isn't blocking that well is considered to be a "bad one". In fact, neither results of the blocking experiment give any information about the "real" specificity of the antibody. 
Three good reasons: 
1. To produce monoclonal antibodies applicable on FFPE sections, antigens are processed through formalin, alcohol, xylene before immunization of mice. This strategy may introduce a conformational change of an antigen/epitope. For that reason those antibodies will not, or just partly, recognize a purified antigen in its original state. 
2. Peptides used for blocking experiments may be recombinant proteins. These aren't containing the characteristic sugar chains being the target of many antibodies.   
3. A monoclonal antibody usually reacts with a very small sequence composed of only a few amino acids or a few sugar moieties. When performing IHC we all hope such a sequence is characteristic for a certain cell type or protein we wish to visualize. However, because the sequence is so small there is quite a good chance that other cell types or proteins contain something a similar sequence. If such a sequence is also present at other cells or proteins, we observe positive staining we don't want to see; let's call it "unwanted specific staining". This is not "background staining" because the reaction between antibody-antigen is after all still "specific".
For further reading on these issues read: Willingham, JHC 47:1233-1235, 1999: "Conditional epitopes: Is your antibody always specific?" 

For the three reasons above I believe that blocking tests doesn't give any insight to the specificity of an antibody. And, that was the initial question to this list.

Apart from the controls as suggested by Tim, you better select a number of good positive and negative control tissues based on literature data. Not necessarily based on IHC, but perhaps with other detection methods like ELISA, FACS, Western blotting.

Chris van der Loos, PhD
Dept. of Pathology
Academical Medical Center
Amsterdam - The Netherlands  

----- Original Message ----- 
>From  "Morken, Tim - Labvision" <tpmorken <@t> labvision.com> 
Date  Thu, 04 Mar 2004 08:36:04 -0800 
To  Histonet <@t> pathology.swmed.edu 
Subject  RE: [Histonet] Confirming IHC staining as "real" 

Probably the best is to complex it with the known antigen before putting it
on the slide and then it should be negative. If you don't have the antigen,
then you need a known control and some info on what the ab is supposed to
stain. Negative controls should include 1) no primary or serum (buffer
only), 2) normal serum from host animal of your primary 3) nonsense
antibodies (aren't supposed to stain the target tissue - use ab's from the
same host and preferably same Ig type) and 4) non-target tissues incubated
with your primary(should be negative). 

Tim Morken
Lab Vision / Neomarkers
www.labvision.com