I am doing flow cytometry with the mouse breast cancer cultured cell line and using fluorophore tagged primary antibody (recommended dilution 1:50) against two protein, say A & B. A is a well known marker of cancer cell and B is hypothesized to show some changes during oncogenic progression.
To optimize the antibody dilution we still are doing single stained sample, i.e., around 10^6 cells with either APC-A or PE-B. The protocol I follow is-
At first, I am harvesting 1*10^6 cells in versene solution (0.48mM EDTA in PBS). Then after centrifugation at 2000 rpm for 2 min 200ul PBS with 2% FBS is added to pellet and allowed for blocking in rocking condition to 4 degree C for 30min followed by centrifugation. After that, antibodies are added in 1:30 and tubes are kept in 4 degree for 30 min at rocker. after incubation, cells are washed with PBS+1%FBS and viability dye PI (0.5 ul form 10mg/ml stock) is added to check live cell populetion only in PI specific cells and APC single colour cells as PI and PE has spectral overlap. After 5 minutes incubation cells are washed once with PBS+1% FBS and finally resuspended in 1 ml of PBS+1%FBS for cytometry. I am looking for 10000 events in flow cytometry.
I was also facing some problem with PI staining (just to check the % of live cells) as maximum percentage of cells were coming PI + signal and PI high signifies dead cells (as I understood). In a most recent result though I got some improvement in getting some live cells. In the last experiment I only used unstained, PI only, APC+PI samples. In this result I got less PI binding as well as, least APC signal (the least percentage can be non specific also as I heard that APC is very good fluorophore that gives sharp binding signal). It happens when I gate the cluster of cells shown in the picture. But when I shift the gate towards high FSC-high SSC area I get an increase in APC positive cells.
My question is which population should I take for analysis? I can understand that in the cluster with large number of cells APC signal (if any bound) gets divided among many cells as in flow cytometry we can see average fluorescence. But in high FSC -high SSC area cell number is less and is that why I get greater fluorescence gating that population. So, overall the antibody binding is not taking place or what is happening? If anybody can suggest me any change in the protocol please do.
I am attaching the result, where analysis is done with all events (without gate) the order is as follows unstained, PI only, PI+APC in replicate. In PI binding, we can see clearly two population but In APC signal red gate shows a little binding whereas green gate increases the percentage of APC binding.
Since you are only working with 1 type of cell, always take the denser population (in this case, the lower SSC/FSC cluster).
The higher SSC/FSC the cells scatter, the negative population starts to move to the right. Yes, if you use the negative gate for the lower SSC/FSC populations, it will seem like that the higher SSC/FSC populations are "positive" but this is not true.
As to what is the higher SSC/FSC cluster? It could be a great many things. They may be cells that got clumped together, they may have a different morophology, etc. If you look at your cells visually, you probably don't see a uniform population.
Dear Swarnali Kar,
In your analysis, I could gate at least three different groups of events (or cells) according to their FSC and/or SSC levels. Please find attached the sample gating. I would recommend you to attempt to identify these groups (small, middle, and large) with a microscopy. I might find out which group of events you are really interested in (probably the middle?) guessing from the shape and smoothness, liveliness (viability), etc..
If you are worried about doublets, you could gate the events with FSC-A together with FSC-H or FSC-W. For more discussion on this, see below.
https://www.researchgate.net/post/How_does_FSC-A_vs_FSC-H_indicate_the_singlets_in_a_FACS_data
Different groups of events could have different levels of autofluorescence as you find in the difference in APC signal levels. Therefore, it is recommended to only compare cells with similar autofluorescence levels.
Thanks Shen-An Hwang , I would go with the denser population. It seems that I have to optimize antibody binding conditions. Can you suggest the parameters those need standardization in the protocol I shared?
Thanks Tetsuo Tsukamoto for sharing the link, it is really helpful. I would try to look into three populations you have gated. As, I have just started with flow cytometry, can you clarify a bit more the last point (about autofluorescence) you have raised?
Different types of cells absorbs lights with different wavelengths (blue, green, red, or violet) and convert the lights to different colors of lights (for example green, red, violet). Autofluorescence are independent from antibody binding.
For example, macrophages are known to sometimes emit strong autofluorescence with very similar wavelengths to PE. So it is sometimes advisable to avoid using PE for macrophages.
Also, it might be helpful to analyze the autofluorescence levels of unstained cells in FITC, PE, APC, etc. channels by FACS prior to determining the fluorochrome used for analysis. If you find significant autofluorescence, then use the channel for autofluorescence check and use another channel for the antigen of your interest.
For a reference, see below.
https://www.researchgate.net/post/Is_it_possible_to_eliminate_auto-fluorescence_in_flow_cytometry
Thank you Tetsuo Tsukamoto for your valuable suggestions. I will surely check for autofluorescnce.
Increase the voltages for APC and PI.
The bigger cells are likely to be dividing/ or changing cells.
You analyze autofluorence only in the unstained sample. If you think it is a problem analyze the same amount of cells for the unstained and the test sample.
In the test if you plot protein A against B you might expect to see a difference in levels of expression of the 2 proteins.
cheers
Ann
You have to gate both the smaller and larger population. It helps if you have some understanding of the physical changes that have taken place in your sample during the treatment phase.
Thank you very much Ann Atzberger for your reply. As you have said about dividing cells (in the high FSC-high SSC), can I confirm that by Flow cytometry cell cycle analysis? and what do you mean by analyzing same amount of cells in unstained and test samples? For both I set 10000 events in flow cytometer.
I just meant it is better to compare the same number of events between control and sample if you have doubts about defining auto-fluorescence.
Please note that the 10000 events also includes dead cells and debris which can be quite high. The less live cells present the more data you need to collect. However your FSC/SSC dot plot looks good. I think these are adherent cells so you would have removed most of the dead cells with the supernatant. For your sample 10000 events will not be enough: you need to focus on how many events are visible in the population of interest. If the number of cells is low (i.e. if the percentage that is positive is low) then you need to increase the total number of cells you save.
You can do cell cycle or you can also stain for Ki67. You can look up various options in Current Protocols in Cytometry...you just need to google it.
cheers
- Badges
- Science method