Panel titration: staining volume, histone markers

[SinaAng]

Hi everyone!

I'm establishing my own panel for CyTOF and am about to titrate all my antibodies. I have two questions regarding this:

1) Since I will use barcoding, I want to titrate my antibodies on a higher amount of cells (around 5-10 million). For my first couple of trials I just used the recommended antibody volume / staining volume of 1ul in 100ul for 1-2 million cells. At what point do I have to increase the staining volume? Is 100ul still fine for 10 million cells or do I need to increase this?

2) I have a few histone markers in my panel. Does anyone have experience titrating these (e.g. using inhibitors)? I have no experience whatsoever working with these so I would be grateful for any input.

Thanks,
Sina

[mleipold]

Hi Sina,

1. Yes, I would recommend that you titrate your antibodies under the conditions that you're going to be using. In other words, if you're going to have 10M cells in a pooled barcoded sample, your total reagent usage is probably going to be lower for a number of markers than if you had 10 samples of 1M each.
- it's been my experience that antibody staining volume doesn't matter that much....most antibodies are high-enough affinity that they'll be able to "find" their targets. In fact, one time I messed up and didn't remove wash buffer from a well before I added my antibodies, so the final staining volume was 500uL rather than the 50uL I normally do. The staining looked exactly the same as usual.

That said, I do think that you want enough buffer volume to make sure that you have fully resuspended your sample to a true single-cell suspension (ie, no clumps that would cause uneven staining). So, from that practical point of view, I might suggest increasing your total staining volume.


2. On the few projects that I've done with histone markers, I didn't do functional testing (inhibitors, etc). It's my experience that there are enough variations in signal across various cell types that I could use that as the check (ie, Naive CD4+ would have a different staining intensity for Histone Marker X compared to B cells or CD56bright NK cells). There are only a couple markers like total H3 or total H4 that are *relatively* constant across most cell types.....most other markers have a lot more cell-type variation.


Mike

[JamesW]

Hi Sina,

On point two we used the EZH2 inhibitor Tazemezostat for H3K27Me3. Some details in figure 5 and the methods of this preprint. https://www.researchsquare.com/article/rs-4327429/v1
We see no change on non-proliferating cells but H3K27Me3 drops in close correlation to number of cell divisions. Since we are blocking methylation rather than actively removing its likely driven by proliferative dilution. Main point there is that combining it with CFSE will make it easier to interpret the results and that if your cells are not dividing you may see no effect from that particular inhibitor.
It's not the experiment in the paper but we also confirmed that H3K27Me1 is slightly reduced while H3K27Ac is increased on the same cells. Since EZH2 is known to be primarily responsible for Me3 and partly for Me1 on H3K27 and that Ac is in competition for binding with them that effect on staining lined up with the biology. So I was personally convinced that the staining was legitimate.

Best,
James

[SinaAng]

Hi Mike, hi James,

thank you for the quick response, that helps. I think I'll try without any inhibitors first then.

As a follow up question, since Mike, you mentioned titrating under the conditions you will also later perform the staining. Would you also recommend titrating the entire pool together then? I have read conflicting opinions about whether to separate conjugates with m+1 or m+16 into separate pools or titrating everything together, so I'm unsure which is best.

Thanks a lot for the help!

Sina

[mleipold]

Hi Sina,

Some of that comes down to your initial panel design:

1. Do you know ahead of time (previous literature, etc) which markers are likely to be really bright? If so, did you put them on dimmer metals, and dimmer markers on brighter metals?
- note: this includes SBio surface marker conjugates. Do not assume that SBio has done a perfect job of matching marker expression with metal brightness.....you always have to check in your case.

2. Even if you have several brighter markers, you can usually find ways to mitigate spillover.
- one side comment: the "M+1" and "M+16" nomenclature that a lot of people use is *not* completely correct, and can lead to misconceptions.
- M+16 is always related to M: M+16 is the oxide spillover of M. The magnitude of M+16 spillover is directly related to the signal at M, factoring in how easily oxidized M is. The trend is that low lanthanides (La and Pr, but also Nd and Sm) are more easily oxidized than high lanthanides (Er, Yb). This is why the Tuning procedure checked the oxidation signal due to La: if you keep LaO below 3% of La, then you're automatically keeping all the other M+O oxides below 3% of M.

* however: "M+1" (related to isotopic impurity) should really be thought of as "the next highest isotope in that element", which is *not* always the same as the next increment in mass. For example: M=150 is an Nd isotope, while M=151 is an Eu isotope. Nature doesn't make a 151Nd, so 150 should never spill into 151 due to isotopic impurity. And in this case, vice versa, since nature doesn't make a 150Eu either. In the case of 151Eu, the isotopic spillover you would be looking for would be 153Eu (=M+2)

So, you can use Element (Pr, Nd, Sm, Eu, Gd, etc) choice to help with some of the spillovers. Similarly, the monoisotopic elements (Pr, Tb, Ho, Tm) can also be used to help control this: since they're 100% that channel, they won't spill into neighboring channels due to isotopic impurity. And, there's nothing else that will directly spill into them either: eg, 159Tb is the *only* naturally occurring 159 mass.


3. You may be able to use some mutually exclusive marker in adjacent channels. For example, you may have prior knowledge that a particular methylation marker is mutually exclusive of a particular acetylation marker (eg, H3K27me1 vs H3K27ac).


** To directly answer your question: panel design is iterative. I personally tend to spend a lot of time in initial panel design, choosing marker expression levels and appropriate metal brightness. I feel that gives me the best chance of making a good panel the first time around, which then allows me to titrate my full panel at once.

Once I feel I have a good titer on everything, *then* I would start checking on some spills (next isotope above and below, M+16, etc) for things that should be biologically negative. The results of that would potentially cause me to change my titer a bit to reduce spill, or potentially require me to make a new conjugate or two and retiter them.

Ultimately, a Metal Minus One (MMO) is the surest way to be check on *total* (all kinds of) spill into a channel. But yes, if you want to do some carefully chosen Metal Minus Many (MMM), then this paper from Bill O'Gorman's group has some good tips: https://doi.org/10.1007/978-1-4939-9454-0_3


Mike

[SinaAng]

Hi Mike,

thank you so much for the extensive reply, I really appreciate the time and effort you put into this.

I mostly didn't do the panel design myself but relied on the help of our contact person at Standard Biotools. A lot of the extracellular markers are similar to an existing panel for immune cell profiling so I assume that they followed that basic principles there.

However I have quite a few Sm and Nd isotopes so those could definitely create issue. Some combination are mutually exclusive but I do have some where the an extracellular marker (e.g. CD14) could potentially spill over into one of my intracellular channels (e.g. pSTATs, histone markers). I think those might be a problem.

I guess it would therefore be smart to do the initial titration with antibodies divided into pools that keep those separate and then put them all together once I have the optimal titration and check for spillover the way you mentioned.

Best,
Sina