Antibody's specifity...

Bryan Heit bjheit at
Wed Jul 23 17:54:26 EST 2003

Comments within:

Mike Clark wrote:

>Antibodies may or may not be "specific" depending upon how they
>are selected and how in practice you define specificity. Specificity is not
>an absolute universal concept. In most cases antibody specifity is defined
>for a narrow application based on emprical observation. e.g. "In ELISA
>assays my antibody only binds to antigen A but not to control antigens B,
>C, or D."
But how many times have you run a cell lysate blot with an antibody and 
seen greater then 1 line?  It's rare, and when it does occur it's 
usually human error or a closely related protein.  In all of the blots 
and elisa's I've done (that's a lot) I've never once encountered an 
antibody with cross reactivity for a non-related protein, and that 
includes some "dirty" polyclonal serums I derived myself.

>The physiological specificity of an antibody usually results from both a
>positive selection of B-cells on one antigen and negative selection on
>others. The most commonly encountered negative selection in-vivo is
>negative selection for autoreactivity.
You are mixing up B-cell development with germinal centre formation. 
 During B-cells development the pre-B-cells undergo immunoglobin gene 
rearrangement.  Cells which fail rearrangement die, a process known as 
positive selection.  Once the gene rearrangement is complete the cells 
then undergo negative selection (how this occurs is not clear).  During 
negative selection B-cells which react with self-proteins are 
eliminated.  But this process has almost nothing to do with the 
maturation of an antibody you would buy or isolate.

When a B-cell becomes activated by an antigen (i.e. when you immunize 
the animal to raise your antibody) you are only activating the cells 
which have undergone negative and positive selection.  Chances are that 
some of these cells will weakly recognize the antigen you inject and 
start to divide.  As these cells divide they deliberately mutate their 
immunoglobin gene to make "new" versions of their original immunoglobin. 
 Occasionally this process results in an immunoglobin with higher 
affinity, and there are mechanisms within the lymphoid tissues to select 
for these higher affinity cells (a process called affinity maturation). 
 The end effect of this is after we immunize an animal a limited number 
of B-cells will be formed.  All will express antibodies with high 
affinity and specificity to the antigen you immunize the animal with.

>So the answer to the original question posed by Nikolai above is, yes,
>under some circumstances.
Can you cite a single occurrence where antibodies show cross-specificity 
with a protein only 50% identical, unless it is recognizing a shared 
epitope?  I did a quick check of pubmed and could not (not to say there 
isn't, I didn't look too closely).

>Also all antibodies have the potential to bind a large number of different
>antigens and thus specificity can easily be seen to be  a "relative"
Specificity is not a relative concept, at least not in any of the lab's 
I've worked with.  Antibody specificity is well demonstrated in the 
literature - it is very rare for an antibody raised against an antigen 
to identify other antigens, even when closely related.  For example, in 
my lab we do a lot of work looking at the adhesion and signalling 
molecules involved in leukocyte recruitment.  We use both murine and 
human models, and we need separate antibodies for both systems, even 
though it's not uncommon for the murine and human forms of these 
proteins to be greater then 95% at the amino acid level, and almost 
identical structurally.  

>You can illustrate this principle with a simple thought
>experiment. For example if you take an antibody to any given antigen, and
>then you raise a series of different anti-idiotype antibodies to that first
>antibody, you essentially have a whole family of different antigens all
>with binding affinity to the same antibody.
Or you could look at it as a group of antibodies all recognizing the 
same antigen.  Idiotype antibodies are a special case, as it is hard to 
identify what is the "antigen" and what is the "antibody".  As I said 
previously, I am unaware of any study showing cross reactivity between 
proteins with little/no homology.


More information about the Immuno mailing list