In response to Mr Benito and others who are interested in non radioactive
LPAs..
Thanks for all the responses to "Query on LPAs: replacing tritiated thymidine
with flow assay". Recommended assays are mixed lymphocyte, Brdu, PKH or
CSFE. I will continue to post additional information. See below for
responses to date.
Lorrie Epling
Mixed Lymphocyte Assay
1) From: "Barren, Phil" <BarrenP@MedImmune.com>
We have done Mixed Lymphocyte Proliferation Assays (i.e. end point
proliferation of lymphocytes) for several years.
We also look at the effects of drug on proliferation etc.
We find it highly accurate and reproducible.
We use a counting bead standard (6 micron fluorescent) added to the samples
post last wash.
1E6 beads per 1 ml of sample.
Count 1,000 beads and determine the lymphocyte number.
example count 1,000 beads and then get 1,500 lymphocyte counts =
1.0E6 beads per ml and 1.5E6 lymphocytes per ml.
CFSE:
1) From: "Jonni S. Moore" <moorej@mail.med.upenn.edu>
CFSE provides a wonderful alternative that we have used in several
systems..you label the responding cells and measure the proliferative
fraction by flow cytometry.
2) From: "Reed, Doug S Dr USAMRIID" <Doug.Reed@DET.AMEDD.ARMY.MIL>
It depends on what you mean by success or failure! Certainly there are
reports in the literature on the use of CFSE and other dyes to track
proliferation of cells. We've been using the same system here to look at
proliferation of cells from macaques. This works beautifully for basic
research and has been quite helpful in looking at what cells are
proliferating as opposed to the total proliferation as assessed by thymidine.
BUT, if you're talking about clinical application of such work, then there is
one major problem with these flow-based assays. Data analysis. Thymidine is
"easy" - you take your radioactivity values and compare wells. But data
analysis can be huge with CFSE. Suppose, for example, you want to track
proliferation of CD4 and CD8 T cells. Easy enough to gate based on CD4 vs CD8
and then do histograms for CFSE expression in each population. Then you have
to determine where your peaks are (at least 8-9 possible peaks for CFSE) and
the percentage of cells in each peak. You can then print that out and type it
into Excel and use a template for quickly determining the frequency of
responding cells or the division index, depending on which reference you've
read on these assays. I can't as yet decide which measure is more 'accurate'
for what we want to know.
As an example - we've been doing this on 96 well plates using the Multiwell
autosampler on a FACSCalibur. We have typically done duplicate wells with 4-5
dilutions of antigen plus two control (media alone) wells. So for three
antigens we need 30-36 wells. So we can get two, or at most three, animals
per plate. Acquiring the data (30k events) takes half a day for 60-70 wells,
the analysis of which can then take up easily a full day (from printing out
the plots on WinMDI to entering the data into an Excel template). I'm getting
better at the analysis end of things and I'm hoping to get software capable
of batch analysis so I can automate the analysis end of it as well. Either
way, though, it will be time-consuming.
My opinion (for what it's worth!) is that these assays are great for basic
research but until the time required to do the analysis is shortened I don't
see much potential for clinical application, at least not for large numbers
of samples.
3) From: Joanna Roberts <jroberts@malaghan.org.nz>
Doug, to speed up the analysis leg of CFSE proliferation work, you could
consider using FlowJo. When we do CFSE proliferation assays (for mouse work),
data analysis using FlowJo makes the whole business of getting the data into
excel for analysis a breeze. The ‘Table editor’ quickly applies the same
gate logic to all the different batches of samples and then produces a table
that can then be opened in excel, containing all the raw data for a
particular batch of samples. Copying and pasting this into a pre-defined
excel worksheet makes relatively light work of things. So it might not end up
taking that much longer than the work required to turn CPM from thymidine
incorporation assays into pretty proliferation graphs.
4) From: "Mark Munson" <mem@vsh.com>
Subject: CFSE or PKH Proliferation assays by flow cytometry
Organization: Verity Software House
Flow-based assays using either CFSE or the PKH dyes have been in use for
quite a while. Dr. Alice Givan at Dartmouth has published extensively on
this issue. Verity's ModFit LT, for PC or Mac, has a built-in proliferation
analysis "Wizard" that makes setting up the analysis quite easy. ModFit LT
supports batch analysis, and also has a databasing capability in ASCII text
format that can easily be read directly by Excel and other
spreadsheet/statistics software.
If you wish, I'd be happy to send you a demo and video tutorial CD set so you
can see how the proliferation wizard works. By the way, there are already
users of ModFit LT 3.0 at UCSF, so you are eligible for the "additional user
license" price for ModFit LT 3.0 for PC or Mac of only $300.
5) From: Alice.L.Givan@dartmouth.edu (Alice L. Givan)
ModFit software from Verity has a "Proliferation Wizard" that does all the
proliferation
calculations for you from cells stained with CFSE or a PKH dye (for example,
it
gives you the precursor frequencies of the proliferating cells and it also
gives you
a proliferation index that tells you how many more cells you have now than
when you
started the culture).
What ModFit does is model the intensity histogram of the cells --- so it can
use the
separate or quasi-separate peaks found in some cases with CFSE-stained cells.
Or it
models the theoretical positions of the peaks for dividing cells if you are
staining
with the PKH dyes (that don't give such good intensity separation) or if your
CFSE
staining hasn't given separate peaks.
Main problem with ModFit software is that it is using the theoretical
position for
the intensity peaks of the proliferating cells -- it won't be as accurate if
cells
are dividing asymmetrically and the proliferation peaks don't match the
theoretical
positions expected for halving of intensity with each division. We have
found that
it matches the peaks quite well with CFSE -- as long as you use a correction
for the
particular amplifcation factor for your log amplifier (the log decades full
scale
are hardly ever exactly 4). And we have also found that the ModFit software
gave us
values that correlated well with the known values of a model system.
Conflict of interest: After writing our paper on using PKH and flow to
quantitate
precursor frequencies of antigen-specific T cells (J. Immunol. Methods 230:
99, 1999),
I worked with Verity to get them to include the precursor frequency
calculation in
their ModFit software. I have no financial interest at all in this --- but
am somewhat
devoted to the concept of using CFSE or PKH to do precursor frequency
calculations
as I feel this method provides much more complete analysis of proliferative
responses
than does tritiated thymidine (which is only a bulk assay related to the
total number
of proliferating cells at the time of the assay).
Others who, before us, have used CFSE and flow to quantitate precursor
frequencies
are Wells et al (J. Clin. Invest. 100: 3173, 1997) and Song et al (J.
Immunol. 162:
2467, 1999). If I am missing any references on this, please let me know.
6) From: Adrian Smith <A.Smith@centenary.usyd.edu.AU>
I haven't ever used ModFit so I can comment on it directly, but,
FlowJo also has a proliferation platform that allows you fit curves
to CFSE-type proliferation data and generate various stats and gates.
Overview...
http://www.flowjo.com/specproliferation.html
Details...
http://www.flowjo.com/v4/html/proliferation.html
I believe FlowJo's proliferation platform automatically takes care of
irregularities of log amps by including a parameter to the fit which
models this (I sure Mario could supply the mathematical details to
those so inclined). It can also model the contribution of
autofluorescene. Like most FlowJo analyses it all happens easily and
automatically (but you can adjust parameters manually if you need to)
and it is very easy to apply to batches of samples. Exporting to
Excel is also very easy - as already mentioned by Joanna Roberts in
this thread. I haven't quite worked out how to get it fit curves
every time but overall it seems to do pretty good job on my data.
(Disclaimer: - I provided a some input into the way the platform
works and I have been beta testing FlowJo for a while now)
BTW - after a lot of delays FlowJo 4 is coming along really nicely -
especially if you are using Mac OS X. I like the new comparison
platform a lot (it is not quite there yet but it shows lots of
potential). If you have a dual processor Mac you should check out the
latest version which has added support for multi-processors.
If you are worried about upgrading to a new dongle they have just
released v3.7 which works with both v3 and v4 dongles, just in case
you need to go back to version 3 for some reason. (Disclaimer: I also
pushed for this - there are some very conservative people in my lab
who hate software upgrades!).
Note there was another program being developed specifically for
CFSE-modelling (I think it was going to be called "CFSE Modeller").
However, it looks to have disappeared and I'm not sure what it's
status is now. I never really had the chance to test it out because
the author was overly paranoid about piracy, ie there was no way to
test it with your own data. Given that attitude it is not
particularly surprising that is failed to thrive - which is a pity
because Phil Hodgkin was involved in developing it and he has done a
LOT of mathematical modelling of CFSE-data.
BRDU:
1) From: andreas.simm@medizin.uni-halle.de
Incorporation of BrdU and staining with an anti-BrdU antibody is as far as I
know the best replacement of the tritiated tymidine assay.
2) From: "Robert C. Leif" <rleif@rleif.com>
Tritiated thymidine can be replaced by antibodies against 5BrdU. A good
source is Phoenix Flow Systems. Go to Reagents ABSOLUTE-S.
www.phoenixflow.com
I had the original idea, which was implemented by Howard Gratzner, Diane
Ingram, and Al Castro. The combination of the invention of monoclonal
antibodies by Kohler and Milstein and techniques of Darzynkiwicz et al.
made it into a practical assay.
Ancient History:
H. G. Gratzner, R. C. Leif, D. J. Ingram and A. Castro; "The Use of
Antibody Specific for Bromodeoxyuridine for the Immunofluorescent
Determination of DNA Replication in Single Cells". Exper. Cell Res. 95,
p. 88 (1975).
H. G. Gratzner, A. Pollack, D. J. Ingram and R. C. Leif;
"Deoxyribonucleic Acid Replication in Single Cells and Chromosomes by
Immunologic Techniques". J. Histochem. Cytochem. 24, pp. 34-39 (1976).
R. C. Leif, S. P. Clay, H. G. Gratzner, H. G. Haines, K. V. Rao and L.
M. Vallarino; "Markers for Instrumental Evaluation of Cells of the
Female Reproductive Tract: Existing and New Markers". The Automation of
Uterine Cancer Cytology, Edited by G. L. Wied, G. F. Bahr and P. H.
Bartels, Tutorials of Cytology, Chicago, pp. 313-344 (1976).
H. G. Gratzner and R. C. Leif, "An Immunofluorescence Method for
Monitoring DNA Synthesis by Flow Cytometry". Cytometry 1, pp 385-389
(1981).
3) From: Babs Marrone <blm@lanl.gov>
The following 2 papers describe our successful use of a BrDU assay to replace
3HThy assay, in a human clinical detection assay. The 2000 article contains
more methodological details.
Beryllium sensitivity is linked to HLA-DP genotype. Wang ZL, Farris GM,
Newman LS, Shou YL, Maier LA, Smith HN, Marrone BL. TOXICOLOGY. v. 165(#1)
pp. 27-38 AUG 13, 2001
Detection of beryllium sensitivity using a flow cytometric lymphocyte
proliferation test: the Immuno-Be-LPT. Farris GM, Newman LS, Frome EL, Shou
YL, Barker E, Habbersett RC, Maier L, Smith HN, Marrone BL. TOXICOLOGY , v.
143(#2) pp. 125-140 FEB 21, 2000
This archive was generated by hypermail 2b29 : Sun Jan 05 2003 - 19:26:19 EST