N2 Assay

Hi Flowers,

I just wanted to thank all of you that responded to my "dot plot"
questions.  I especially want to thank Kathy Schell for assisting me
through an actual assay.  My hope is to keep improving from this point
forward.  Thanks to all of you.  Some of the responses are listed below.

Sincerely Grateful,
Israel


Hi Israel,
Given the plots you show it is very difficult to make any conclusions.
In
particular there are a couple of technical shortcomings which I will run

down for you.
First a few words what you are supposed to do (although I can only
guess)
Your PI wants to see whether the expression of CD54, 58, 80, 86 changes
after the treatment with IL4 and TNF. CD54 (ICAM-1) and CD58 (LFA-3) are

adhesion molecules which are expressed on many cell types (I do not know

whether on melanomas). CD80 and CD86 (formerly known as B7.1 and B7.2)
are
activation markers for leukocytes (ligand for CD28 and CTLA4 (=CD152)).
I
guess your PI expects upregulation of this marker on the melanomas.

Analysis:
FSC-SSC plot. Set a region 1 around the cells of interest.
Make a histogram FL1, gated on region 1. Analyze the unstained cells.
(your
negative ctrl). Make an overlay of the isotype ctrl. It should run
nearly
identical with the unstained otherwise you have non specific staining.
Make
a third overlay with the FL1 stained cells. If the curve runs like
isotype
control, the marker is not expressed, if the curve runs shifted to the
right, the cells express the marker. If the curve has two modes (aka
peaks)
you have two populations (one positive one negative). In a new
histogram,
plot the stained cells and determine the % of positive cells with a
marker.
Do the same with FL2
Finally make histograms with the stained cells from treated cells with
an
overlay from the same staining with untreated cells. If the curves are
different, the treatment had an effect.
For analysis set a marker region in a way only about 0.1% of the
untreated
cells would fall in. Then determine the % of cells in the treated
sample.
This approach is only correct if you have positive AND negative cells in

your treated sample. Otherwise you have to show the increase of marker
expression.
To do this: make a histogram with stained untreated cells. Put a marker
over
the stained cells.
Do the same with the treated cells. If the marker has an increased
expression after treatment, the MFI (mean fluorescence intensity) should

increase. Use the median value from the statistics (if there is no
median
(have a close look, sometime you have to activate this) use the
geometric
mean (Gmean, but only if your data are in log). If your data are in lin
(which we should not do) only then use the mean, otherwise it will be
wrong.

Good luck
Ulrich

Hello Israel,

First of all, for a person who had no training in flow cytometry you are

certainly doing good.

Second, before even talking of flow, few comments about your cell line.
From
my own experience, when you activate cells with all these growth
factors,
you change cells drastically. So do not expect to see your cells in the
same
window on your flow plots. As you mentioned cell morphology is changed
after
the activation: it means that if the cell size is changed, your FSC is
changed; if the internal complexity is changed, your SSC is changed; and

auto fluorescence may become higher after the activation. Therefore, it
might
be impossible to compare your activated cells versus your control
non treated cells.

But you certainly can give it a try by doing several things:
You may try to grow your cells, collect them and run without all these
growth factors involvement, choose the best (well presented) markers for

these cells and run 2-color FITC-PE flow. You should be interested in
looking   FSC/SSC, FITC/SSC and FITC/PE results. Also, instead or in
parallel with isotype controls, run just unstained cells and look at the

results in the FITC/PE window. This way you will be able to set up your
quadrants and gates and find your "double negatives" .
 Then take your cells and activate them according to the instructions
and
repeat the same procedure. And you will see right away, where to look
for
your cells of interest. If  the scatter is moved, you will move your
gate
and still might be able to look at your cells.
 And again, sometimes especially for cell lines isotype controls are
useless
(I am not using them too much even in my clinical settings), so run
instead
unstained cells, look for your double negatives, set up your quadrants
and
this may help your to see your specific staining (or lack of it) upon
activation.

As for clumping, I may suggest couple things:
1) to add up to 2% of fetal bovine serum into your staining buffer,
2) increase from 1 to 2mM concentration of EDTA,
3) if your procedure allows, fix cells with 1-2% of formalin in PBS
before
running, keep them for 20 min. at +4 degrees C and then run,
4) buy falcon tubes with the nylon mesh on it to filter your cells
before
staining.

If you are have any questions, feel free to contact me through e-mail or

phone and I would be glad to help you as much as I can.

Good luck. Irina

Hi Israel-

Without knowing your settings, I would suggest the following:

1.  Decrease the FSC setting.  Your cells of interest are off-scale.

2.  To compensate properly, set your control cells in the first log
decade (the lower left corner) of an FL1 vs FL3 dot
plot.  Then run single-stained samples (stained only for ONE antibody of
interest -- when running multiple markers with the
same fluorochrome, you can use the brightest one to set compensation).
Then compensate FL1 and FL3 so that their
respective X or Y median values are the same as for the negative cells.

3.  Confluent cells tend to clump more than non confluent.  If the
biology demands that you culture to 100% confluence,
then you must; however, if possible, grow to 70% or 80% confluent.

4.  Wash and stain cells in Calcium- and Magnesium Free medium (dPBS).
This should help prevent clumping.  Many people
also use some small % BSA (0.5 to 2%).

5.  For data presentation, especially when asking for trouble-shooting
assistance, we need to see your controls:
experimental (biological), compensation (single stained, with positive
and negative populations), and ESPECIALLY your
negative populations!  I might have missed something, but it seems that
your PMT voltages are too high (settings for each
detector), but I can't tell for sure, because there is no negative or
unstained population for reference (the other
possibility is that your staining is REALLY bright).

6.  You need to titrate your antibodies.  Gross excess of antibody can
stain your negative population positive, to the
point of actually having both appear to be the same, i.e. no apparently
negative cells.

7.  Take some cells BEFORE fixation, but after all other manipulations.
Use a dye like Propidium Iodide (for membrane
integrity) or CMXRos (for mitochondrial membrane potential), and run
them on the cytometer, looking for Propidium positive
cells in the FL2 channel on a log scale.  This will give you an idea of
how your cells were doing before and during their
processing.  This can help you troubleshoot your cell culture
conditions, as well as your protocol.  If you are not fixing
your samples, you can use Propidium in your samples to show compromised
cells.

8.  Back to cell culture:  different culture conditions can dramatically
change receptor/chemokine expression.  If you want
reproducible data, split your cells before confluence, do not allow
medium to become depleted, and ALWAYS seed them by
density, NOT BY VOLUME (i.e. seed 5e5 cells in a 10 cm dish; don't just
split a 10 cm dish 1:4).  Believe me, this is
critical when looking for reproducible biology.


I hope this helps.  Please let the list know how you resolved your
situation.

Andrew