Many thanks to those of you who sent advice on our CFSE-based MLR assay.
Here is a summary of the advice, plus my comments. The original query is
at the end.
"TC media often has indicators that fluoresce, and thus the longer they
grow, the more 'autofluorescent' they may become. A brief wash may not
allow for the dye to diffuse out. " ? Indeed, when I was doing a BCEF-based
cytotoxicity assay, the protocol called for RPMI without phenol red. We?ll
give this suggestion a try and let you know if it helps.
Several replies dealt with the possibility that the CFSE is leaking out of
the stained cells and is being picked up by the unstained cells:
"If I understand correctly, you're using "feeder" cells (unstained)
with CFSE-stained cells, over a 72-hr period?"[ actually, cocultured
unlabelled speen cells added as putative regulatory cells or as control
cells] " I would say your CFSE-stainedcells are leaking their CF, and your
"unstained" cells are picking it up.
"I would bet that the much higher dilution factor in vivo is causing a
loss of available CFSE and labelled cellular debris and protein and thus
giving the better results."
"Does ?reutilization of your CFSE label? mean that you're staining
more cells with the same medium, which contains CFSE? If the staining is
just as bright as was the staining for the first set of cells, it sounds
like you're
using too high a concentration of CFSE. It is very bright, and should be
titrated."?The unstained cells are only giving a very pale signal in the
fluorescein channel, in channels wherein we?re hoping to find divisons 5 or
6 through 8 of the stained responder cells. Also, to get that many
divisions, one needs a reasonably bright CFSE stain. Based on comparison
with the literature, I think we?re not overdoing it.
"In our hands, CSFE tends to bit leak from the cells, and the dye
contaminated other cells adjacent to it. "
Whether or not there is any dye uptake by unstained cells cultured with
CFSE-stained cells is an important issue,and may well be the source of our
problem, although none of the literature that I?ve seen suggests the stuff
leaks. As a succinimidyl ester, CFSE is not fluorescent, but becomes so
when it diffuses into cells and is cleaved by cytoplasmic esterases. The
resulting molecule is highly reactive and reportedly forms
membrane-impermeable adducts with proteins, etc in the cytoplasm, hence its
utility as a tracer. I suppose some still uncleaved molecules could
diffuse right out again, though I would hope our washes would take care of
most of them. (cleaved but not-yet reacted molecules would probably bind to
the protein in our wash medium).
"We use PKH26 from sigma with very good results. This dye is 'Locked' into
the membrane of the cells so no dye leakage." This sounds like a great
solution to potential leaking, tho Parish (Immunol Cell Biol 1999 77:499)
points out that lymphocytes are not ?homogenously? labelled with PKH26.
Other replies:
"Any chance you can gate out your stimulating cell population (using an
MHC I antibody specific for the stimulator allotype)"?unfortunately, the
regulatory cells we?re testing are of the responder strain. "or perhaps
the ?disturbing? cells are mainly dead, you can do propidium iodide
staining." We do FSC/SSC gating for lymphocytes/blasts, but need to reserve
our other two channells for CD4/8 and other markers.
"Are your stimulators irradiated or otherwise inactivated? We found that it
was really important to "clean up" the cells after culture by spinning over
a Ficoll-Paque gradient. This removed most of the dead/dying stimulators
and solved the "autofluorescence" problem."?This sounds like a good idea,
but many of the divided cells, whose presence we are trying to detect, may
be undergoing apoptosis. I believe it was Turka et al. that showed that
you can still detect these dying progeny during the CFSE assay. In
particular, an effect of our regulatory cells may be to induced "premature"
apoptosis in the responding population . I am afraid that a density
gradient removal may selectively remove cells that we needed to be counting
in the assay.
"do you really need the number of cell division information or could you
live with proliferation vs non-proliferation: use BrdU and stain
intracellularly for the incorporated BrdU." Yes, we do really need all the
wonderful information the CFSE assay could theeroretically provide:
precursor frequency, # mitoses/dividing cells, phenotypes of the responding
cells. It?s a really great technique?if we can just adapt it to our
system!!
Many thanks to you all.
Original query:
>We are attempting to use CFSE to monitor in vitro MLR?s for a study of
>alloregulatory cells in mice. Our model involves a relatively weak
>response (single class plus minor antigens).
>
>Our responder cells are labelled with CFSE and, when harvested, stained
>with either CD4- or CD8-Tricolor (Fl3). We take great care with
>compensation, as the bright CFSE fluorescence spills into Fl2 and Fl3. We
>live gate on lymphocytes/blasts using FSC/SSC.
>
>Our problem is with autofluorescence of the cells that are not labelled
>with CFSE, stimulator cells, and particularly, the cells we are adding as
>co-cultured cells: responder strain regulatory cells or responder strain
>naïve cells (controls). The co-cultured cells are added as unfractionated
>or fractionated splenocytes.
>
>The autofluorescence of the non-CFSE labelled cells increases during 72 hr
>in culture until it spans 2 decades on the Fl1 log scale. The brighter
>autofluorescence thus interferes with measuring the number of responder
>cells that have undergone 4 or 5 to 8 divisions and are then CFSE dim.
>Thus the usefulness of the CFSE measurement is severely limited.
>
>Our responder cells are washed well before culture and reutilization of
>the CFSE label does not appear to be a problem.
>
>In contrast to the in vitro MLR, in vivo MLRs performed with CFSE give us
>beautiful data (but unfortunately are not suitable for our regulatory
cell
>study).
>
>We would be grateful for any information or thoughts on this problem you
>may have.
>
This archive was generated by hypermail 2b29 : Sun Jan 05 2003 - 19:01:21 EST