RE: Antibody titrations (2004...2005...2006... and beyond)

With this posting, I absolutely agree!	One additional recommendation to 
Mario's I would add is for antibody titrations and subsequent staining, do 
not ever stain with less than 10ul of volume of your diluted antibody.	Do 
serial dilutions in a plate or tubes and then transfer 10ul (or more) of 
the appropriately diluted antibody to the tube containing cells - unless, 
of course, you use an extremely accurate and repeatable pipetteman.  Most 
standard bench pipettemen are not routinely calibrated and not very 
accurate under 10ul not to mention what volume might be on the outside of 
the tip.

julie


At 11:55 AM 4/27/2004 -0400, Mario Roederer wrote:
>Joanne's comments reinforce the requirement that you must titrate your 
>antibodies (whether you buy them, make them, or steal them), under your 
>conditions .... ALWAYS.  Don't assume that the manufacturer's recommended 
>concentration is going to work best for you.  Don't assume that you need 
>to use 100 times as much antibody for 100 million cells (that's clearly 
>wrong), but you can't assume that you can use the same amount 
>either.  Joanne is correct that time and temperature are important 
>variables.  Volume is not an important variable as long as you think in 
>terms of concentration  or dilution of antibody as opposed to microliters 
>per test.
>
>I agree with Joern's comment that it will be hard to change people's way 
>of thinking about these until the manufacturers start making their 
>suggested titres as dilution rather than "microliters per million 
>cells"--which of all of the possible units of concentration is the 
>worst....)  Who among us has asked the manufacturers how they determined 
>the titre value--was it 10 minutes on ice?  30 minutes at room 
>temp?	Fixed? Unfixed?  Full moon?
>
>Note that even changing conditions like staining fixed & permed cells will 
>dramatically affect the titration.  A common anti-CD3 antibody that we use 
>at 1:5 dilution on unfixed cells can be used at less than 1:50 if you 
>stain already-fixed and permed cells!	(On the other hand, our anti-CD8 
>antibodies titrate to the same level whether cells are fixed or not).
>
>A quick calculation serves to show that most antibodies are already in 
>vast excess so that cell number is a relativley unimportant 
>variable.  Let's consider anti-CD4.  There are ~50,000 CD4 molecules per T 
>cell, so staining 1 million PBMC (50% = CD4) means that there are about 
>2.5 x 10 ^10 CD4 molecules.  A typical anti-CD4 antibody titres out to 10 
>ug / ml (1 ug per 100 ul test volume); this is equivalent to about 3.6 x 
>10^12 antibody molecules.  Therefore, the antibody is in 150-fold 
>excess.  Changing the cell number from 1 million to 5 million will be 
>irrelevant to the final staining intensity.  Changing to 100 million may 
>make a mild difference, so use 5x as much antibody (not 100x).
>
>Bottom line:  always titrate EVERY reagent, and do it under every 
>condition that you use in the laboratory!
>
>mr
>
>At 11:33 AM -0400 4/27/04, Joanne Lannigan wrote:
>>Mario:
>>I have to agree with Julie on this based on personal experiences. We have 
>>noticed that some manufacturers' recommendations are in 20-50 fold excess 
>>while others are at the peak of the titration curve where a five fold 
>>lower concentration affects the staining intensity significantly. And as 
>>you mentioned, the antibody affinity plays a large role.  The other 
>>factor which some forget about is the staining volume; adding 5ul of 
>>antibody to 10ul of cells yields a different final concentration than 
>>does 5ul of antibody in 200ul.  Incubation temperatures and times are 
>>contributory factors as well.
>>Another point I would like to bring up is epitope density. Too often I 
>>see people titrate their antibodies on whole bone marrow or spleen and 
>>then use these concentrations to label pre-enriched populations of cells, 
>>i.e. 90% enriched T or B cells and wonder why they don't get the same 
>>intensity of staining. I remember back to the 80's when people in the 
>>clinical labs were moving to whole blood labeling and discovered that 
>>patients with CLL and very high white counts (primarily clonal B cells) 
>>had staining issues when using the standard 100ul of blood to the 
>>recommended ab concentration recommended by the manufacturer for 100ul of 
>>blood (based on normal WBC counts). It was particularly a problem when 
>>trying to demonstrate light chain clonality. The consensus group for 
>>leukemia/lymphoma immunophenotyping subsequently recommended diluting the 
>>whole blood in patients whose white count was greater than 20,000/ul. The 
>>bottom line is that there needs to be an excess of antibody over bindable 
>>epitopes in order to achieve saturable binding. Using too large of an 
>>excess, especially with low affinity antibodies, leads to decreased 
>>signal/noise due to nonspecific binding of the "free" antibody.
>>I think in some circumstances your statement is correct but there are 
>>other factors to consider in addition to cell numbers.
>>I am sure there will be further discussions ensuingŠ.
>>
>>Regards,
>>Joanne
>>
>>Joanne Lannigan, MS
>>Director, Flow Cytometry Core Facility
>>University of Virginia
>>Jordan Hall, Room 7067
>>P.O. Box 800734
>>Charlottesville, VA 22908-0734
>>Office: 434-924-0274
>>Lab: 434-243-2695
>>Fax: 434-982-1071
>>email: joannelannigan@virginia.edu
>>-----Original Message-----
>>From: Mario Roederer [mailto:roederer@drmr.com]
>>Sent: Thursday, April 22, 2004 3:35 PM
>>To: Cytometry Mailing List
>>Subject: Antibody titrations (2004)
>>
>>It is time for my annual reminder that antibody amounts for staining are 
>>based on the concentration of the antibody, not the number of cells being 
>>stained*.
>>
>>I.e., the antibody:antigen ratio is such that, until you stain more than 
>>50 million cells, in general, you are staining in vast antibody excess 
>>and the cell number is irrelevant.
>>
>>So if you stain 2 million or 5 million (or 100,000) cells, do NOT change 
>>the amount of antibody you are using, you simply waste reagent and might 
>>increase background!
>>
>>mr
>>
>>(* PS, as a side note:  very high affinity antibodies are actually "bad" 
>>in this regard.  The optimal titre for these reagents is so low that in 
>>fact they can become cell-number (antigen-amount) dependent at typical 
>>cell numbers, about 1 million per stain.  On the other hand, low affinity 
>>antibodies are nice in that you could stain 1 billion cells with the same 
>>amount that you stain 1 million cells and have no decrease in fluorescence!
>>
>>For a full explanation of these issues, and a brief description of how to 
>>titrate antibodies properly, see Kantor, A. and Roederer, M. (1997). FACS 
>>analysis of lymphocytes. In: Handbook of Experimental Immunology (Fifth 
>>Edition), Herzenberg, L. A., Weir, D. M., Herzenberg, L. A. and 
>>Blackwell, C. (ed.), Blackwell Science, Cambridge, pp 49.1-.13.
>>
>>2.4MByte PDF available upon request)
>>
>>
>>At 12:59 PM +0800 4/22/04, Michael Wong wrote:
>>>Dear Noosheen,
>>>
>>>Our lab has just purchased the eBioscience ZAP-70 FITC antibody. I 
>>>wonder how much antibody is needed to incubate with 1x10^6 cells. 
>>>According to the NEJM article, they are using 1.5ug antibody per 500,00 
>>>cells. But they are using Upstate unconjugated antibody. Is this the 
>>>same for this eBioscience antibody?
>>>
>>>Regards,
>>>Michael