Greetings to Cliff, Jim, and all,
I agree with Jim Houston, and offer this what its worth: "High
speed" issues at the pressures/frequencies commonly used in the popular
commercial sorters isn't new. Recall that Fulwyler's 1965 system ran at
72KHz, and the high-speed systems at developed at Lawrence Livermore and
Los Alamos in the mid 80's for sorting chromosomes ran at over 200KHz,
for example. What we're talking about is figuring out what cells can be
run under what conditions on what instrument.
Having said that, at Scripps we have been dealing with a variety of
samples on our sorters that are equipped with what is commonly called
the high speed option, which can easily double or triple the throughput
over the standard configuration. As a practical matter, we prepare the
sorters to run in one of three general speed ranges based on nozzle size
and drop frequency. Each range is associated with a different recharge
rate for the sort. In that way, very low speed (<30KHz) - medium speed
(<40KHz) - high speed (>40KHz) setups allow us to charge based on how we
expect the samples to run.
It didn't seem like a good idea to have a user pay twice as much for
a sort just because his cells may be more fragile and need to be run at
half the rate for twice as long, for example. Our base rate goes up
$10.00/hr to the maximum at high speed. We also considered charging the
user by the number of cells run, but that took away any incentive to run
quickly if you can - and we wanted to encourage users to run a fast as
reasonable without damaging their cells. Typically, we try to maximize
recovery by not exceeding one cell every 5 or 6 drops accross the
frequency ranges and use a 1.3 or 1.4 drop envelope. A lower speed sort
at 30KHz, for example, will have an upper limit near 5000 to 6000
events/second. A high speed sort at 66KHz will usually have an upper
limit near 13000 events/second. Exceptions: we do enrichment sorts and
sorts where recovery is not the main concern much faster, and start
slowing down the event rate if the investigator appears to be getting
too pale. In other words, we adjust the event rate based on abort rate
to create tolerable losses no matter what the setup is in terms of
nozzle, frequency, and pressure.
We decided 40KHz was a reasonable drop rate above which we would
call the setup high speed, and have been doing this since 1996. In our
facility, we go "high speed" using a 70um nozzle (> 40KHz) about 50% of
the time. The remainder of the time we're using a 80um, 90um, or 100um
nozzles at medium or low pressures to deal with samples that seem to
require it due to cell size or vulnerability.
Currently, we only put healthy, robust, smaller cells on the sorters
at high speed, which is typically 34PSI at 66KHz on our Vantage SE
systems. We have found that if you're not sure how well a valuable
sample may survive the vigors of a high speed sort, yet the sample seems
healthy, you can do well going near 22PSI and 44KHz - cells that don't
make cytokines after a 34PSI sort often will after 22PSI, for example.
Fragile cells (apoptotic, dendritic cells, some activated cells, etc.)
seem to require pressures under 25PSI for good performance after
sorting. I think we have found exceptions to nearly every rule we have
come up with, so consider these very general guidelines as to what we
are doing. Fragile sorts (some dentritic cell samples, microglia, etc.)
are usually done with an 80um nozzle under 20PSI, but some investigators
are "pressure shy" and don't even want to try to run extremely valuable
or fragile samples higher speeds - so we've still not sure exactly what
the limits might be for some samples. Some sensitive samples seem to
require < 15PSI to perform as expected after sorting. To complicate
matters, post-sort viability can vary among laboratories for reasons
over which we have no control. Milage on the MoFlo and Coulter systems
may differ somewhat from ours, but in a stream-in-air sorter you cannot
change the Poisson characteristics of sorting cells within discrete
droplets. That is the same on all the systems.
So what is high speed? I would say, "it depends." Sure, that the
cells be "alive" after sorting is not enough. They do need to still
makes cytokines, etc., to be useful to the investigator, and titrating
setup conditions for optimal sorts and post-sort cell performance isn't
trivial. Buffers are critical, I suspect a lot of the cell "death" some
have reported after sorting has more to due with their sample buffer not
containing HEPES, or mixing significant volumes of PBS containing cells
and a carbonate buffer such as media, for example, than the sorter
setup. Sure, it is fun to wind up the drop drive frequency, light your
hair on fire, and go as fast as you can - but there are a lot of
variables, and pressure/frequency is only one of them.
Regards,
Joe
This archive was generated by hypermail 2b29 : Sat Mar 10 2001 - 19:31:11 EST