Re: CD4 loss with activation

From: Mario Roederer (Roederer@drmr.com)
Date: Thu Jan 13 2000 - 18:37:35 EST


To follow on to this discussion, I decided to provide some data to 
prove that using only two colors (CD3, CD8) to identify T cell 
subsets is prone to significant artefact, especially in measuring 
cytokine profiles.

I looked over our database of recent immunophenotyping (over 800 
visits, most of which are HIV+), and tabulated the relative 
proportions of CD3+ T cell subsets, being CD4 (CD4+CD8-), CD8 
(CD8+CD4-), DN (CD4-CD8-), or DP (CD4+CD8+).  The following are 
expressed as a percentage of CD3+ T cells [given as "mean (SD)"]:

HIV-:
CD4 = 63 (10)
CD8 = 29 (9)
DP = 0.8 (0.5)
DN = 5.1 (3.6)

HIV+:
CD4 = 19 (12)
CD8 = 72 (11)
DP = 1.3 (1.4)
DN = 6.3 (4.2)

Therefore, the use of CD3+CD8- to identify CD4 T cells would include 
7.4% DN T cells (for HIV-), and 25% (!!!!) DN T cells for HIV+.  In 
other words, among CD3+CD8- T cells in HIV-infected adults, 
one-fourth are NOT CD4 T cells.  Therefore, you CANNOT consider the 
DN population to be inconsequential.  Because of the unique cytokine 
profile of this subset, it can significantly impact on cytokine 
measurements.

There is another important consideration when using stimulated cells. 
Especially with PMA + Ionomycin, a lot of cell death can occur (even 
in 6 hours).  Because dead cells tend to be sticky, the dead cells 
will appear to be DP--and they have very high binding of (certain) 
anti-cytokine antibodies.  Therefore, even gating on CD3+CD8+ to 
identify T cells can include a substantial contamination of dead 
cells after stimulation.  If you are uninterested in dead cells, then 
you can start by simply gating out DP cells in your analysis.  Or, 
you can use a live/dead discriminator that works with fixed cells 
(Ethidium monoazide, EMA is the only one I know of--again, see Mitra 
et al., Int Immunol 1999 11: 1801-1810 for details).

So, DN, DP, and dead cells can all introduce serious artefacts into 
FACS cytokine measurements.

And, no, I'm sorry, but we have not published this data or 
interpretation.  (I'd be happy to put together a brief letter if some 
kindly editor requests it, however.)

mr

(PS:  Please note:  we are not a clinical lab and I do not wish to 
characterize these values as "reference ranges".  These values should 
be used only as an instructive example).



This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:55:26 EST