>Basically, you need to run at high flow rates, of >50,000 cells/s and then=
=20
>you need to threshold out most RBCs (say 90%, when 0.1% are WBCs), but the=
=20
>remainder comprise WBCs at or below 1/1000 cells. Every white cell has 1, 2=
=20
>, 3, 4,.. RBCs coincident except for a MoFlo MLS system with 5 microsecond=
=20
>dead-times, where you get mainly none, one or two coincident RBCs.
>
>My patent protocols generally use a pulse width and a DNA dye to add an=20
>exclusion gate for non-nucleated cells, but then you must resolve the issue=
=20
>of RBC coincidences and how they affect gates for the WBC subsets and the=
=20
>dispersion in fluorescences of positives and negatives!
>
>If you don't have a MoFlo, then buy one; else forget it!
>
Bob may be correct about needing a MoFlo to sort white cells uncontaminated
by red cells; however, analysis of WBC's in unlysed whole blood can be done
on more conventional machines. The field dates back to the 1960's, when a
group working with Leitz in Germany (Hallermann L, Thom R, Gerhartz H:
Elektronische Differentialz=E4hlung von Granulocyten und Lymphocyten nach
intravitaler Fluochromierung mit Acridinorange. Verh Deutsch Ges Inn Med
70:217, 1964) considered adding a fluorescence channel to the then current
scatter-based blood cell counter. They showed that using acridine orange as
a dye would not only allow WBC counting in unfixed blood using fluorescence
as a trigger signal, but that different levels of red AO fluorescence could
discriminate mononuclear and granulocytic cells. Adams and Kamentsky (Adams
LR, Kamentsky LA: Machine characterization of human leukocytes by acridine
orange fluorescence. Acta Cytol 15:289, 1971; Adams LR, Kamentsky LA:
Fluorimetric characterization of six classes of human leukocytes. Acta
Cytol 18:389, 1974
590) showed that differential leukocyte counting in unlysed whole blood
could be accomplished with AO; the Los Alamos group confirmed cell
identifications by sorting (Steinkamp JA, Romero A, Van Dilla MA:
Multiparameter cell sorting: Identification of human leukocytes by acridine
orange fluorescence. Acta Cytol 17:113, 1973). The prototype "Cytomat"
hematology counters built at Block Engineering in the 1970's analyzed whole
blood at 50,000 cells/second, using a nuclear DNA stain to identify and make
6-parameter, 3-beam measurements on WBC's while counting RBC's and platelets
and measuring hemoglobin on a cell by cell basis (Curbelo R, Schildkraut ER,
Hirschfeld T et al: A generalized machine for automated flow cytology system
design. J Histochem Cytochem 24:388, 1976; Kleinerman M: Differential
counting of leukocytes and other cells. U S Patent 3,916,205, 1975; Shapiro
HM, Schildkraut ER, Curbelo R et al: Combined blood cell counting and
classification with fluorochrome stains and flow instrumentation. J
Histochem Cytochem 24:396, 1976; Shapiro HM, Schildkraut ER, Curbelo R et
al: Cytomat-R: a computer-controlled multiple laser source multiparameter
flow cytophotometer system. J Histochem Cytochem 25:836, 1977; Shapiro HM:
Fluorescent dyes for differential counts by flow cytometry: does
histochemistry tell us much more than cell geometry? J Histochem Cytochem
25:976, 1977). Since then, B-D has obtained patents on counting and
measuring surface markers on lymphocytes using immunofluorescence as a
trigger signal instead of a nucleic acid stain; Terstappen and Loken
(Terstappen LWMM, Loken MR: Five-dimensional flow cytometry as a new
approach to blood and bone marrow differentials. Cytometry 9:548-56, 1988),
while at B-D, also described staining procedures using a combination of
antibodies and the nucleic acid stains thiazole orange and LDS-751. This
wasn't commercialized because, while B-D holds patents on thiazole orange,
Abbott holds patents on LDS-751. However, the use of cyanine dyes (in which
class thiazole orange falls) and styryl dyes (in which class LDS-751 falls)
as nucleic acid stains was described in the literature as far back as the
1930's. Thus, there's a great deal of prior art in this area which might
have some bearing on current (and should have had some bearing on past)
patent applications in this area.
I've done T-cell subset analysis in whole blood using nucleic acid stains
and/or CD3 immunofluorescence for triggering. The use of a signal other
than scatter is critical, as the signals from RBC's will choke the scatter
channel of most instruments at rates higher than a few thousand
cells/second. With a fluorescence trigger, not only don't you need a MoFlo,
you can almost certainly do the job on a FACScan or EPICS XL, and I've done
it myself on Cytomutts and the Ortho Cytoron Absolute as well as on the
Block instruments.
-Howard=20