Thu Dec 05, 2013 7:58 pm by mleipold
Hi Ian,
It *does* take a fiendishly long time to develop a robust panel. Ofir and I are trying to find ways to set up a database of working conjugates for general knowledge, but we're not there yet.
To answer your question more directly: When you say it doesn't work, what do you mean by that? No significant signal on your expected cells?
This could come from a number of things.
1. The prep itself is bad.
a. Poor antibody recovery (low Nanodrop A280): if the recovery is less than 20% or so, I throw out the prep, usually untried. Even if there is some activity, there are usually antibody aggregates or other problems that will affect staining. Not to mention, with a really low yield, you're going to be remaking the conjugate again soon anyway. It could be a bad lot of antibody, a hole in the spin filter, or overreduction causing aggregation.
b. Poor metal loading. If your protein concentration is OK but you don't see signal, try making a dilution of your antibody stock and running it in liquid mode (like tuning solution). I would recommend serial dilutions down to 1/100,000x and then work your way up. Most antibodies will give you detectable signal at about 1/10,000x.
This could be caused by a bad metal stock (or pipetting error), a bad polymer stock (if the maleimide has been hydrolyzed, it can't covalently link to the antibody free thiol after reduction), or under-reduction (too dilute or expired TCEP; though TCEP should be good for at least a year, it *does* go bad eventually). All of these could lead to the antibody just not having much metal.
c. Overreduced antibody, due to too high [TCEP] or too long spent with TCEP.
In all these cases, trying the particular prep *again* is what I usually do (ideally, with a different lot number of antibody). Only if it fails a second time do I move on to trying other clones. I personally haven't seen much difference in quality between the same clone from different vendors.
2. Try another clone. I had to try at least 4 CD85j antibodies before I found one that worked well enough in CyTOF.
3. Make sure that the epitope you're testing for survives experimental conditions. CD62L doesn't survive freeze/thaw during cryopreservation, but does come back at least somewhat after resting. Fixation drastically affects the binding of most anti-CD16 clones (drops to almost nothing) and many anti-CD56 clones (often get a lot of false-positives). Healthy resting cells may not make a lot of an activation marker: you may need to stim them, or use a cell line for validation (proteinatlas.org can help you find a cell line).
These questions can be worked out on standard flow, *as long as you exactly follow the CyTOF staining protocol* for consistency.
4. There are some epitopes we just haven't been able to find a good CyTOF clone for yet. Using a FITC/APC/PE/biotin-labeled primary and doing secondary staining with a MAXPAR-labeled anti-FITC/APC/PE/biotin/etc adds a step, but has been shown to work in many cases where the primary antibody lost activity upon direct MAXPAR labeling. This also has the advantage of repurposing flow antibodies you *know* work for you. Of course, validate the MAXPAR-labeled secondary with something specific and abundant like FITC/APC/PE/biotin-CD3/CD4/CD8/CD20 beforehand.
This is generally the last thing that I try.
Mike