Help with CFSE staining of splenocytes

26 posts / 0 new
Last post
coltifer
coltifer's picture
Help with CFSE staining of splenocytes

I'm having lots of trouble getting a CFSE protocol to work for cells. I harvest splenocytes and then put them in 1%FBS in PBS with 2.5micromole of CFSE at 15million cells per mL. After 9minutes in the incubator I add 5x the volume with cold 20% FBS in DMEM. I then put the tube in ice for 10 minutes and then wash three times with serum containing media. I then plate the cells at .5million/ ml at 4mL total in a 6 well plate with 5microgram/mL of Con-A.

I rarely see nice sharp peaks of the CFSE stain after collecting on Day 1, Day 2, Day 3, Day 4. The cells might be dieing...But why I'm not sure. The FITC peaks I see are very broad and don't show the nice sharp divisions. I've been working on this protocol for months now. Any ideas??

marcus muench
marcus muench's picture
I use a minor modification of

I use a minor modification of your protocol, not that I really think that is the problem. I incubate the cells with 5 micromolar CFSE in just PBS for 10 minutes at room temperature. Minor difference and I'm not sure that it accounts for your troubles.

I'm tempted to believe that the mixture of cell types you have is contributing to the spread in the peaks. Can you counter stain with another marker like CD3 to identify the sub-population your interested in in the final analysis? Staining PBMC from human blood I find that the surviving non-T cell population (monocytes mostly) after an MLR has a lower CFSE fluorescence than the undivided T-cells. If I didn't have CD3 in my analysis I would also have one big sloppy peak of CFSE expression. You might get some idea if I'm right by looking at your forward vs. side scatter profile and see if you can identify suspopulations of cells. Gate on the subpopulations and see if that sharpens your peaks.

coltifer
coltifer's picture
yeah, i've tried double stain

yeah, i've tried double stain with CD3 PE and still get a blur of populations.

marcus muench
marcus muench's picture
Are you discriminating live

Are you discriminating live and dead cells with PI or some other such method?

Can you post a picture?

coltifer
coltifer's picture
I haven't been discriminating

I haven't been discriminating dead cells. I will try that next though with 7-AAD.

I've tried both splenocytes and a T cell cell line. I'm thinking that the splenocytes have several cell types and divide at several rates. But this still doesn't explain the smear I see when I double stain with CD3 or when I stained a cell line with CFSE.

Here are some pictures of the results i get.

marcus muench
marcus muench's picture
Your staining looks pretty

Your staining looks pretty good at Day 0, although I'm a bit surprised by how quickly it declines. I assume your showing me the results from the cell line? In our hands, even after 7 days in a murine MLR we still had a tight peak of undivided cells. It looks like all your cells are dividing, so that makes me think this is your cell line. Even with a tighter peak of initial staining, discriminating subsequent populations is less than perfect, as you can see below. The data is from our publication: Chen JC, Chang ML, Muench MO. A kinetic study of the murine mixed lymphocyte reaction by 5,6-carboxyfluorescein diacetate succinimidyl ester labeling. J Immunol Methods 2003;279:123-33.

At this point my best guess is that the spread of your populations is due to a combination of three things. One, your cells vary in size (FSC), probably due to their position in the cell cycle. Do your peaks get sharper if you gate only on the main cell population seen in your FSC vs SSC plot? Do the slightly larger cells have a higher autofluorescence and thus are broadening the peaks? Second, your day 1 results may actually represent two populations, those that have divided once and those that have divided twice. Again, your cells seem to be dividing at a rapid pace given how fast they are loosing staining. Third, even cell lines have a population of dead cells and discriminating these cells should help clean things up, but this won't change your results that much. I think in MLR it is very important as the stimulator cells and even the responder cells are likely to die at a high rate depending on the day of analysis.

(Edit: my figure didn't seem to attach, so if you look up the paper it is figure 3).

coltifer
coltifer's picture
Gating on just the lymphocyte

Gating on just the lymphocyte population doesn't change the peaks. They're virtually identical to the peaks I've shown you.
I'm going to go back to staining spleen cells and LN cells and double staining with PE-CD3 or something.

This is frustrating...I only want to hammer out this protocol, so that we can use it in the future for in vivo proliferation studies. Any ideas on how to go about this? I really appreciate your help.

marcus muench
marcus muench's picture
I must admit I'm also running

I must admit I'm also running low on bright ideas.

Try staining spleen or LN cells as you stated and do an in vitro proliferation assay and stain for dead cells when you analyze. At this point I would avoid the cells lines if that is not what you want to do in the end anyway.

Although your initial CFSE staining looked OK, is there any chance your CFSE is old? We keep our aliquots at -20°C and with desiccant.

edit: I meant to say in vitro, not in vivo. Also, when you test this on splenocytes, culture some without stimulus. Unstimulated T cells should survive a number of days and stay CFSE bright.

I have not performed in vivo proliferation assays, although it sound relatively straight forward once you get it all working in vitro. I remember a paper being posted on these boards some time ago regarding the technique, but your probably aware of all the relevant papers before you started the project.

Gibacht
Gibacht's picture
I have noticed that the

I have noticed that the amount of ConA you are using is 5microgramms per ml, we have found that 1 microgramm per ml is enough. And also it might help to vortex the cells 2 times during those 9 minutes labeling time. But these are only minor "cosmetics" but it might help.

Gibacht

samm
samm's picture
actually, the 5 ug v/s 1 ug

actually, the 5 ug v/s 1 ug may not just be 'cosmetic' - 5 ug/ml Con A actually shows greater cell death than 1 ug/ml for primary cells (JLeuk Biol,2002,2005).
Your protocol is very similar to what I use, except mine is 2 uM final, and culture in 96-well plates at 100kcells/well. Your cells do seem to stain well initially, so no problem there. (I'm not very surprised if a cell line shows no/low 0 div peak)
Try reducing ConA conc or using ViViD Red/PI or other live v/s dead discriminitor.
Ensure you have a DDM module on (FSC-A v/s H),before FSC-A v/s SSC.
Your primary culture optimal profile should look like this (~70 h activation), with a peak ratio ~0.5. The fig is for just the basic DDM-gated cells, with no other SSC/live-dead gates applied.

coltifer
coltifer's picture
Thanks, I'll try the 1ug/ml

Thanks, I'll try the 1ug/ml Con-A on splenoycytes and see if it helps.

coltifer
coltifer's picture
Okay, so here are some plots

Okay, so here are some plots from Day 4 of the CFSE stain at 1.25uM. I also lowered the con-A concentration to 1ug/ml. Still not pretty.

Having a lot of trouble attaching images...no formats work for me.

coltifer
coltifer's picture
images

images

samm
samm's picture
If I see it right, you seem

If I see it right, you seem to have picked the wrong gate from your FSC-SSC plot - its straddling the lo-high FSC population.
I don't know which cytometer and program you are using, but if it supports it, try acquiring/analysing cells in this order:
Set threshold to ~5000
Gate cells on the diagonal in the FSC-A/FSC-H plot (DDM) - use as count/collection gate, but acquire ALL cells
Gate the lymphocyte population from FSC-A/SSC-A
Apply compensation matrix.
Gate CD3+ cells
Apply CFSE analysis.

How is your compensation matrix like? It appears as if you do not have the proper compensation set up for spillover from FITC to PE.
CFSE is a rather terrible dye in that respect - it spills over in PE, PerCP-Cy5 (FL2, FL-3) to a considerable degree. Once that is straightened out, you should be able to see the division peaks in your CD3+ lymphocyte population much more clearly. Also, you will need to use a curve-fitting program - most times it cannot be done by eyeballing it - see the example I'd posted earlier.

coltifer
coltifer's picture
The red population shown in

The red population shown in PE vs FITC is the smaller gate from the scatter plot. I think the compensation is fairly good (PE and FITC is fairly distinct and no weird curvature is there.) I do admit that the FSC/SSC plot looks odd, but I seem to get that when using Con-A in cultures. I did do the FSC-H/FSC-A but it is not shown.

MaI)Max
MaI)Max's picture
While CFSE staining can be

While CFSE staining can be pretty finicky and I am not an expert, I can make several suggestions. First,  I would suggest purifying your splenocytes prior to staining with CFSE. I noticed that different cell populations stain differently, and may cause a "blurring" of the peaks if populations are responding at different rates or times.

I resuspend purified B-cells in 1 ml PBS buffer at a concentration of about 50-100 million cells per ml.  The CFSE stock is made fresh immediately prior to staining and covered with aluminum foil to prevent bleaching of the dye (I cover all of the cells after staining as well).

In addition, I stain for 5 minutes of staining with 4uM CFSE at 37 C, followed by immediate quenching with media. In addition, I mix the cells every minute while staining.

After staining, I found that it helps to wash the cells 3 times with fresh media to remove any residual CFSE in the solution.

Before running the samples through the Cytometer, it helps to remove the media in which the cells were living, and replace with fresh buffer (I use MACS buffer with PI added to screen for cell viability).

When running the samples, I keep them on ice and in the dark for as long as possible. This may help as well.

Also, when running the sample, make sure your gating is appropriate: I found that my cells of interest, even after purification, only represent a fraction of the total events that are recorded. I have attached a sample picture of FSC-H/SSC-H, PI gating, and CFSE flourescence to show you what I mean.

Hope that helps!

marcus muench
marcus muench's picture
 Good post MaI)MAx.  One

 Good post MaI)MAx.  One question, why do you first gate on FSC vs SSC rather than on live cells?  Seems it would be easier to identify cells of interest based on light scatter characteristics after you remove the many dead cells.

Todeh
Todeh's picture
Im also having some trouble

Im also having some trouble with the CFSE protocol, here is my procedure

  1. I use a CFSE concentration of 5µM )
  2. Cells were incubated in the dark for 5minutes
  3. Washed twice with cold RPMI (Serum and hank's solution)
  4. Re-suspended in 1ml RPMI
  5. I added 15µl of PHA
  6. Cells where then incubated in a 96 well plate in a co2 incubator for 5 days at 37 degrees.

here are some of my readings:

marcus muench
marcus muench's picture
 Todeh,

 Todeh,

Do you anything to remove dead cells, like propidium iodide?

Todeh
Todeh's picture
I have the 7AAD viability

I have the 7AAD viability stain,i will try that. Thanx

marcus muench
marcus muench's picture
 A big factor in getting the

 A big factor in getting the individual peaks to discriminate is defining your cell population of interest as best as possible.  A live T-cell has different autofluorescent and CFSE-staining properties from dead cells, monocytes etc.  The overlapping peaks from all these cells can make finding your live proliferating cells among the bunch difficult if not just plain impossible.  

Todeh
Todeh's picture
Thank you very much for your

Thank you very much for your advice,
I have been using the viability stain and i have changed the CFSE i was using to the cell trace cfse, i started to get promising results. But for some reason i keep getting two populations from the unstimulated cells, i have tried staining them with B cell markers and other markers to try to identify what they are but i havent found anything. They stain as CD3 positive cells!!!
I dont currently have the dot plot on file so i will attach the histogram, the dot plot basically has a cluster of cells all to the far right of the plot and to the left of this population there is a smaller but distinct population! Does anyone have any idea what this could be?

marcus muench
marcus muench's picture
 Todeh,

 Todeh,

I went back and looked at my lab's old publication on using CFSE for murine MLR and found very similar results as yours.  Both CD3+ and CD3- cells were represented mostly by a single population (peak) of CFSE staining, but there is a small population (or two for CD3-) that are brighter.  I don't know what these cells represent.  My guess is that it may either be some other cell population sticking to CD3 non-specifically, or perhaps cell-cycle or activation state differences in the cells at the time of CFSE staining that result in a higher level of staining.  You could try looking at forward vs side scatter profile of cells in the two peaks to see if they are larger cells in the brighter peak.  In our analysis we mostly just ignored the events and gated around the main peak population.

See Fig. 1 in:

Chen, J. C., M. L. Chang, and M. O. Muench. 2003. A kinetic study of the murine mixed lymphocyte reaction by 5,6-carboxyfluorescein diacetate succinimidyl ester labeling. J Immunol Methods 279:123-133.

If you don't have access to a copy of the paper, then you can send me a message and I cam supply it for you.

Happy New Year and good luck!
 

Todeh
Todeh's picture
I'm still having trouble with

I'm still having trouble with my unstimulated cells, im getting two populations but sometimes the population on the left is bigger than the right and vice verca. but the interesting thing is that the histogram only shows a single peak. does anyone know why this is the case?
i tried gating on both populations seperatly and i get the same picutre for both.
I have also tried constantly vortexing the sample after the addition of CFSE, but i found no difference.

Todeh
Todeh's picture
And also i noticed that when

And also i noticed that when i add Candida Albicans to the cell prior to culture the two populations move further apart!

Todeh
Todeh's picture
I've just recently started to

I've just recently started to use con-A, but im having trouble. Im getting a very high percentage of dead cells and the cell count is extremely low compared to my results using the PHA!  any ideas?