Titration of antibodies means finding the optimal amount to use in order to maximize signal (specific binding) while minimizing noise (non-specific binding). This is essential for high parameter flow because not only will an untitrated antibody produce poor or incorrect results for that marker, large amounts of non-specific staining may overwhelm signals throughout the panel. Optimally titrated antibodies give more reproducible results and also reduce the lab’s costs because the amount used is usually less than expected.
Testing in this case means determining the optimal conditions for staining, and checking that the antibody produces specific signal on the cells you want to stain.
In order to get the most use out of this process, try to first answer these questions:
Where is my antigen? Is it on the surface, in the cytoplasm, in the nucleus or an organelle?
Which cells express my antigen? Do they express it always, or only under certain conditions (e.g., specific tissues or after PMA/ionomycin stimulation)? Similarly, which cells don’t express it and thus might serve as negative controls? See ImmGen, RNA-Seq or flow examples from companies or papers.
Is the expression high, low, unimodal, bimodal or multimodal? This affects your choice of fluorophore. You should have a good idea of what the expression is supposed to look like in order to know if you are seeing the right pattern in your testing.
What concentration is your antibody? Oftentimes, fluorophore-conjugated antibodies are supplied at 0.2mg/ml. If your antibody says something else (e.g., 1mg/ml or “tests), you may need to test more concentrations than in the examples below.
I always titrate my antibodies using other markers present to define the cell populations that both do and do not express the marker detected by the antibody undergoing titration. Why? This helps me know whether the signal is specific or not. In the absence of a known genetic knockout of the target, the best option for determining what the negative should look like is a known negative population based on our knowledge of biology. What about isotypes or FMOs? Well, an isotype control tells you about the non-specific binding of an antibody conjugate of the same isotype, but those isotype controls have been selected for lack of binding. Not exactly a fair comparison. FMOs are useful to know the background without the antibody, so you can include that, but to be honest, they're more useful for understanding spreading error in large panels.
In this example, KLRG1 is expressed on activated Tregs, but not on naïve Tconv so we can better determine the correct dilution to use. We wouldn’t have this information if we had tested the antibody without the markers for CD4, Foxp3 and CD62L.
How to titrate
You can directly start titrating--skipping testing--if your antigen is well known, you know what it should look like and where it is expressed, and you know the staining conditions required to produce the best staining. If not, you probably need to do some testing or research first. Anything expressed on the cell surface can go straight to titration.
Titrate your antibodies using the same number of cells and the same staining conditions that you intend to use in your panel/experiment. This means using the appropriate staining protocol and markers. For instance, if you want to look at Ly-6C on Tregs, you need to include the markers that allow you to gate on Tregs, not just Ly-6C, which is expressed at high levels on monocytes and low levels on Tregs. Always perform your titrations with a dead cell exclusion marker.
Example A: Titrating CD25
CD25 is expressed highly on Tregs, with minimal expression on resting Tconv or B cells. For this, you would want to include Foxp3, CD3, CD4 and viability. Prepare a backbone mix of anti-CD4, CD3 and viability stain in the appropriate buffer. Pipette into 6 wells of a 96-well plate as follows:
A) 240ul
B) 120ul
C) 180ul
D) 120ul
E) 120ul
F) 180ul
Add 2.4ul of anti-CD25 to Well A. Well A now contains anti-CD25 at 1:100. Pipette up and down to mix, and transfer 120ul to Well B. Repeat down to Well F. You should now have anti-CD25 diluted in a mix with anti-CD3, CD4 and viability stain at 1:100, 200, 500, 1000, 2000 and 5000. Store in the dark until after the next step.
Plate a fixed number of cells into the same wells of a 96-well plate. Block for 20min. Spin down and discard supernatant. Collect your mix from before and transfer 100ul to the cells, resuspending carefully. Stain for the desired time, then wash. To get Foxp3 staining, then fix, permeabilise and stain for Foxp3. Wash and acquire.
Here's the result:
In this example with CD25 titration, merely looking at histograms of CD25 expression on all live cells makes it hard to determine where the positives and negatives are. We also would have a mix of specific binding on Tregs plus non-specific binding to macrophages. Since we’ll be gating on T cells in the final panel, it makes sense to determine the optimum dilution for specific signals on our pre-gated population.
In this example, we can also see that the CD25 staining is similar across a range of concentrations. This is a good indication that this antibody will be robust to small changes in cell number or staining conditions. If we don’t see this, we might consider looking for a different conjugate to improve the robustness of our panel and limit batch effects.
Example B: Titrating Eomes
Eomes is a transcription factor expressed in NK cells and a bit in certain activated CD8 populations. For this, you would want to include NK1.1 or CD56, viability and probably an exclusion marker (F4/80 or CD14) since NK cells are not common and macrophage/monocyte binding might interfere.
Plate your cells in a 96-well plate and block. Perform surface staining for the NK and macrophage/monocyte markers, plus viability. Wash and fix with eBioscience Foxp3 Fix/Perm buffer. Wash twice with eBioscience perm buffer and discard the supernatant. In another set of wells on the same plate you may now prepare your intracellular staining titration, using eBioscience perm buffer as the base (with Fc blocking if desired):
A) 240ul
B) 120ul
C) 180ul
D) 120ul
E) 120ul
F) 180ul
Add 2.4ul of anti-Eomes to Well A. Mix and serially dilute by transferring 120ul from A into B, then C, etc. Add 100ul of the diluted antibody to the cells, mix and incubate overnight at 4C. The next morning, wash and acquire. Note: Most transcription factors are optimally diluted at 50-200, so you may need to start with 2x or 4x the amount of antibody.
Assessing results:
You can assess the results of your titration by measuring the MFI of the positive and negative populations and looking at the signal:noise ratio (positive/negative). A higher value is better. You can also just look at the population that should not express the marker and determine which dilution factor provides the least staining there while preserving positive signal. If the negative cells move a lot, you probably need to try again with more blocking or with a greater dilution series.
In most cases, the same antibody clone conjugated to fluorophores of similar brightness will require similar dilution factors. If you are titrating anti-CD8-PE-Cy7 and you see a big difference in your optimal concentration compared to data you've previously obtained using CD8-PE, consider that something may be wrong.
Testing
Testing is most important if you are looking at intracellular markers. With intracellular staining, you use fixative, which can degrade antigens by crosslinking them, permeabilization, which can release cytosolic antigens, and you open the cell, exposing antibodies to a far greater variety of antigens for non-specific interactions.
Testing is also important to determine if the new antibody you have purchased from antibodies-r-us.com is specific. For this, you need to have positive and negative controls, which can be cell types or conditions (+/- simulation). Alternatively, you might have a blocking peptide or you might use soluble antigen (e.g., blocking staining with anti-IL-2 by including recombinant IL-2).
One standard test you might run is comparing fixatives such as the eBioscience kit versus 2% formalin vs. PFA+methanol.
If you are looking at a cytokine, you might test various stimulation conditions, e.g., PMA vs PdBU.
You may also need to test blocking conditions, such as using serum from the same species as your antibody, or reproducibility, such as how sensitive your antibody is to the presence of debris or non-hematopoeitic cells.
Notes on testing unconjugated antibodies:
If you’re using an unconjugated antibody, it probably is against a less characterized antigen and the antibody probably hasn’t really been validated for flow. First, determine the species of your primary antibody, and select an appropriate secondary. Titrate the secondary using a primary that is clearly bimodal (e.g., CD4, CD8).
Alternatively, conjugate your antibody using a kit or reactive fluorophore, after which you will be able to use serum from the host species (e.g., rabbit) to block some of the non-specific binding.