[Cytometry] Antibody titration & staining index

Thomas Ashhurst thomas.ashhurst at sydney.edu.au
Thu Feb 20 21:37:18 EST 2014

Hi Ina Laura,

In terms of how to analyse titration data, this is the approach I take (adapted from Mario Roederer's online protocol and other sources). I have included more detailed workflows at the end of my email.

The images listed below can be found at this link: https://www.dropbox.com/sh/a9uartva5pm2iu0/XKtU7QWF32

I will try to address these three questions:
1. What is the best optimal concentration
2. How do I determine this concentration
3. How do I do this with homogenous samples

I hope this helps! Please feel free to ask for clarification or some more examples if necessary :).



1. What is the best optimal concentration
Titrations are described as antibody concentration (i.e. ab volume / total volume), not ug of antibody per number of cells.

Firstly, are you interested in "saturating" or "separating" concentrations.
1. "Saturating" is the concentration where the antigen density per cell can be accurately measured (i.e. you can detect low vs high expression on individual cells). At saturation, the number of cells shouldn't impact your results. This is important for markers where the level is indicative of identity or function. This is chosen as the lowest concentration where the staining index of the positive population is unchanged from those more concentrated.

2. "Separating" is the concentration where negative and positive are separate, but the level of antigen expression on a cell cannot be quantified. This is typically used for markers that are on/off (i.e. + or -) and are used as identity markers. For example, CD8+ T cells are identified as CD8+, not CD8lo or CD8hi. There is some relevance for achieving saturation in this case, as the level of CD4 or CD8 on T cells can be indicative of activation status (from memory it downregulates slightly upon activation, I might be wrong about this). "Separating" would not be sufficient for markers such as Ly6C (in the mouse), as Ly6Chi, int, and lo are all relevant measurements on monocytes or other granulocytes, and so saturation needs to be used.

My lab almost always works with saturating concentrations, particularly as we look at different tissues in a range of settings. This means that any increase or decrease in expression level is due to antigen up or downregulation, not because it is under-saturating.

2. How do I determine this concentration
Calculating staining index is a good approach, but I like to ALSO create a meaningful visualisation of my titrations as well. This helps for homogenous samples especially, I find.

A neat way to visualise titrations, suggested to me by Ben Roediger is to concatenate your FCS files (see protocol BELOW). This essentially combines all of your titration FCS files into one. Here is an example using FITC-Ly6C (on mouse spleen). This example is partially homogenous, although there is a clear positive "group". In this case, 1/400 was the "optimal" concentration. The staining index of the positive population is at a plateau from 1/100-1/400. Therefore, 1/400 is the lowest concentration which still achieves "saturation". At this level, the negative/background is still not too high (although it is higher than 1/6400 for example).


Note: in another example below, sometimes the background can be very high at saturating concentration, and might be difficult to use.


3. How do I do this with homogenous samples
Homogenous samples are more difficult, but can still be done. Below I have included an example of APC-CD11c (mouse) titrated on spleen cells. 1/100-1/400 appears to have a consistent positive group, and 1/50 probably has some excess staining. The saturating curve then starts to drop off at 1/800 down to 1/6400.

Another trick that can help is to also stain your cells with another marker that would help to identify the population you are interested in. For example, if you were interested in titrating human CD14, then you could co-stain with human CD11b at a fixed concentration. CD11b+ cells (expressed on the myeloid lineage, including monocytes ) will help separate out the CD14+ cells from the background, clearing up your results. I don't have an on-hand example of this.



How to concatenate FCS files
- If you have separate FCS files for your different concentrations, then you can add a keyword in flowjo.
- Click EDIT, create a new keyword call it "titration"
- On the workspace list enter the concentration of antibody (i.e. 1/100 = "100").


- Then select all your files for that particular titration set
- Click workspace > "concatenate samples"
- Under "create additional parameters" chose "titration"
- Ensure "concatenate successive files together" is selected (or greyed out if you have already selected your files)


- This will create a NEW FCS file, that merges all of your titration FCS files
- Open this FCS file in a NEW WORKSPACE, change the x-axis to "titration" and the y-axis to whatever fluorophore you are looking at.

- You should get an image like this (mine is FITC-Ly6C -- mouse). Left to right is: 1/100, 1/200, 1/400, 1/800 (ab uL / total vol uL) etc


How to perform titrations:
Serial dilution
- Create a 2-fold serial dilution of antibody concentrations, in 8 wells of a 96-well plate, each with 100uL final volume.
- Put 200uL in the first well, and 10uL in the other 7
- Put 8uL of chosen antibody into the first well and mix
- Take 10uL and add to second well -- continue until the last well, and discard the remaining 50uL

You will end up with 8 wells with 100uL in each
- Well 1: 1/25
- Well 2: 1/50
- Well 3: 1/100
- Well 4: 1/200
- Well 5: 1/400
- Well 6: 1/800
- Well 7: 1/1600
- Well 8: 1/3200
(ab volume/total volume)

This is measured in uL of antibody / uL total volume. It is important to know the concentration of your antibody stocks, but functionally when you are making antibody mixes you will do it in uL.

Apply to cells
- 8 wells of a separate 96-well plate
- Add 1 million leukocytes each (or whatever cell type you are using, however many you intend to use)
- Fc block the cells
- Add 50uL of each antibody dilution to the appropriate well

- INCUBATE: 20-30min, 4*C (whatever conditions you will be using)
- SPIN and wash 2-3 times

- Record on cytometer

Thomas Ashhurst, B.Sc (hons) | Ph.D. Scholar
Viral Immunopathology Laboratory

Discipline of Pathology, School of Medical Sciences and
Sydney Institute of Emerging Infectious Diseases and Biosecurity (SEIB)
Sydney Medical School
Rm 260, Blackburn Building D06 | The University of Sydney | NSW | 2006
T  +61 2 9351 6157 | F +61 2 9351 3429 | M +61 401 050 583
E  thomas.ashhurst at sydney.edu.au<mailto:thomas.ashhurst at sydney.edu.au>

On 18/02/2014, at 2:40 AM, Barsky, Lora wrote:

I've always thought that homogenous populations like monocytes could be properly titrated by preparing two groups of tubes, one group stained with a serial dilution of CD14 (8 tubes) and another group stained with a serial dilution of isotope control (8 tubes).   It is critical that the two antibodies, CD14 and isotope control, be matched isotype class and have an identical beginning concentration.  The tubes are stained and washed separately.

After processing, the matched pairs are combined (16 tubes reduce to 8 tubes) and acquired.  Also acquire an unstained sample.  The unstained sample should be processed as the others were, except no antibody is added.  This will introduce the same artifact to the negative which may now be present in the antibody labeled tubes.  Artifact may come from processing reagents.

Now your data set contains positive and negative populations and a stain index or  alternatively a signal to noise ratio be calculated. The negatives now show at what concentration the cells bind to excess antibody concentration while the positives will show decreasing intensity of label.   The point at which the negatives and positives show the largest separation is chosen as correct titer.

Remember not to use lymphocytes as your internal negative.  They and monocytes have differing autofluorescence intensities, therefore become inadequate pairs for determining stain index or signal to noise ratio.   Also, monocytes have been shown to bind with higher background fluors containing Cy- fluors, but I figure you know this already.  Choosing to correct this issue by finding staining tigers.

Best wishes,
Lora Barsky
Director, USC flow cytometry core facility
University of Southern California

Sent from my iPhone

Sent from my iPhone

On Feb 17, 2014, at 6:41 AM, "Ina Laura Pieper" <inalaurapieper at gmail.com<mailto:inalaurapieper at gmail.com>> wrote:

Hello flowers,

What's the objective way of deciding on the optimal antibody concentration
during antibody titrations? I've searched and found this description on the
UC Flow blog of plotting the stain index against antibody concentration,
However, to calculate stain index it seems I need a positive and negative
population in my sample, probably of the same autofluorescence, which isn't
always achievable.

What do I do if I want to calculate the optimal concentration in a
titration of a CD14-antibody in whole blood? I've done overlay histograms
of stained and unstained samples gated on monocytes (positive) and
granulocytes (negative), so that i can visually identify the concentration
at which the granulocytes are no longer staining unspecifically. Is this
the right way to do it? Seems very subjective to me. From the attached
figure I think a 1:5 dilution is the best (no unspecific binding on
granulocytes whilst still a clear monocyte signal).

What should I do when staining a homogenous sample, and there is no
negative population?

Best regards,
Ina Laura
Cytometry mailing list
Cytometry at lists.purdue.edu<mailto:Cytometry at lists.purdue.edu>
Search the list archive at  http://tinyurl.com/cytometry

Cytometry mailing list
Cytometry at lists.purdue.edu<mailto:Cytometry at lists.purdue.edu>
Search the list archive at  http://tinyurl.com/cytometry

More information about the Cytometry mailing list