Re: dead cell descrimination with PI

From: Mario Roederer (Roederer@drmr.com)
Date: Thu Feb 08 2001 - 13:41:17 EST


Absolutely necessary.  Dead cells can cause all sorts of artefacts.
When staining freshly isolated PBMC off of a Ficoll, it's not so
important (unless you are analyzing subsets whose frequencies are
<1%).  But if you have cultured cells, or, even "worse", stimulated
cells, you must include PI.

Dead cells have a naughty ability to nonspecifically binding
antibodies.  Not only that, but not all antibodies will bind with the
same "nonspecific avidity".  (Yes, you can quote me on that new term).

It isn't too hard to find published flow data which are obviously
dead cells--usually, these appear on a strict diagonal when plotting
two immunofluorescence channels against each other.  Even worse is
when people use quadrant gates (or, for that matter, any gate), and
call these cells "positive", "double positive", or "looky here I've
found something new".

The problem is significantly exacerbated when looking at relatively
rare populations.  For most of us, our cell preparations usually
contain only a few percent dead cells.  (Even straight off of a
ficol, we can see 1% dead cells).  If you have a 1% population that
is nonspecifically binding (some) of your antibodies, you will
completely confound your results.

PI can be measured in a variety of different detectors, including the
PE, Cy5PE, and other red detectors.  Use one of them to gate out the
cells (usually, PI is squashed against the top of the Cy5PE axis),
and then you needn't worry about it's contribution in other channels
anymore.  If you have particularly bright stain in that channel, you
can gate out PI-positive cells in a different channel.   There is no
need to devote a separate detector to PI.

PI is easy to use--just at it during the final incubation step, you
can (but don't have to) wash it out.  You can fix the cells in
0.5-1.0% paraformaldehyde if you analyze them within 1-2 hours (after
that the PI leaks out).

If you need to do live/dead discrimination on cells which will be
fixed for longer than a few hours, or on cells that will be fixed and
permeabilized (for intracellular staining), then use Ethidium
Monozide (EMA).  The EMA technique has been published; I won't go
into details here.  Suffice it to say, however, that EMA does require
that you dedicate a channel to its measurements; it's not as bright
as PI and has more background.  Therefore, you can't distinguish it
from a typical immunofluorescence marker.  Note that doing live/dead
is particularly important for stimulation assays--especially
intracellular cytokine--and EMA is the only way I know of to do this.

As an example, some years ago we were approached by a fine reagent
manufacturer who had two different clones of anti-Human IL4.  They
were getting very different results--one gave consistently about 4-5x
as many IL4+ cells as the other.  We compared the clones, and found
that the one that gave higher binding bound very well to dead cells;
the other clone did not.  Thus, in our stimulations (PMA+Ionomycin)
where typically 10% of the cells die, we would find about 12% IL-4
positive (10% dead + 2% making IL4), but the other clone gave only
the 2%...

In my erstwhile lectures to students on flow cytometry, I would hand
out a list of artefacts to look our for:  cells on the diagonal was a
dead-on indication of dead cells--and nonspecific staining.  Of
course, looking at univariate histograms, one would never know.
Thus, I indicated to them that should the manuscript not specifically
detail that dead cells were excluded by PI-staining, and the authors
were looking at relatively small subsets, then you'd better take it
with a grain of salt.  (Wait, that'll only raise your blood pressure;
take it with a grain of sugar).

mr


At 4:56 PM -0500 2/7/01, Derek Schulze wrote:
>As a general question, how many people out there are using PI to gate out
>dead cells when immunophenotyping?  Is this really necessary?
>
>Derek Schulze
>
>Cancer Research Labs
>Queen's University
>353 Botterell Hall
>Kingston, ON, K7L 3N6
>(613) 533-6635
>
>http://meds.queensu.ca/medicine/crl/



This archive was generated by hypermail 2b29 : Sun Jan 05 2003 - 19:01:05 EST