 |
 |
 |
 |
Laura Dash asks about measuring secreted proteins using agarose microdrops
and refers to our recent paper in Bio/Technology (Kenney et al. 13:787, and
see also Gray et al. J. Imm. Meth. 182:155). What do you really need to do
it... what tricks are lurking around the corner, etc.
We've shown hybridoma secretion assays in some detail, and have shown posters
with transfected cells selected for high secretion. All of our published
work was done using an instrument and some reagents obtained from One Cell
Systems of Cambridge MA. (617 868 2399). They have an exclusive license for
patents in the field issued to MIT, and have done lots of work in the area.
You should certainly speak with them before embarking on any commercial
endeavor in this area.
As we mentioned in both our papers, we have also made drops using a magnetic
stirrer and reagents which are generally available. I don't want to say that
you can't miss, but it's not really very tough. We used 500cs
dimethylpolysiloxane from Sigma in a 50ml erlenmeyer flask, and biotinylated,
low-melt agarose from FMC Bioproducts (it's not in their catalogue, but speak
with John Morgan (207 594 3470). We filter-sterilize the molten agarose
through .2um filters (a hassle but important). A clean cell prep is crucial.
Ficoll any suspension cultures, or wash adherent cells before
trypsinization. In our experience, injecting the molten agarose-cell
suspension slowly into stirred oil is the best way to make uniform, small
drops. The size of the drops is determined by a combination of factors
including viscosity of the oil, temperature (fluidity) of the agarose,
stirring speed (shear) and homogeneity of the cell dispersion.
The most important consideration in my mind is to think about ways to test
your capture and reporter ligands. We always do some "spike" experiments
using drops without cells, and various combinations and titrations of the
soluble components of the assay to make sure things are acting reasonably.
When you are happy with these results, start working with your cells. Before
you worry about fluorescence, run some unstained drops and make sure you can
easily resolve singly occupied drops from empty (~80%) and multiple hits
(~5%). Preparations with small drops are much easier to resolve (and sort)
than big drop preps. In a phase microscope you should find lots of drops
about of ~2 cell diameters, with the cell nicely centered, and when you sort
the prep you should be able to get rid of any double hits. The number of
empty drops in your sorted prep will depend on whether you set the instrument
threshold low enough to detect empties, but they should be only a few percent
of the sort if you run at several hundred events per second. This scatter
resolution is important because the unoccupied drops are great internal
controls for lots of experiments.
For fluorescence, be sure to include a "no capture" control, that is, drops
with cells and all the reagents you will use except for the ligand that
fastens your secreted protein to the agarose. This prep should show only
background (cellular autofluorescence) fluorescence levels, and tests for
lots of problems like nonspecific reporter binding to agarose or cells, or
nonspecific binding of the secreted product to the gel. With good reagents,
and healthy cells this should work fine. Test different secretion periods.
With hybridomas, and some transfected cells we can start to get signal after
only 15 minutes. You know you've gone too long when the unoccupied drops
start picking up signal, indicating that some of the cells have saturated
their drops and started spilling product into the culture.
Sorting works fine with a 100um nozzle. We've done almost all our work on a
FACStar+, with an ACDU to clone. Recently we tried a BD Macrosort and it was
pretty sweet, allowing us to get much shorter delays with the 100um nozzle.
A quick experiment on a Coulter Elite looked fine, good scatter resolution
and sort with the quartz flow cell.
I'm obviously pretty proud of this assay format and would of course like to
hear from folks as they try it. The papers are unfortunately out of date in
terms of author's address, since Syntex is now Roche and our group is now
scattered. Please send comments to this bulletin board or to me at
thatjack@aol.com.
Jack Dunne
415 988 0628
|
|
|
|