0196-476~/83/0304-0227$00.

00/0 Pages 227-242

CYTOMETRY Vol. 3, No. 4, 1983
Copyright 0by the Society for Analytical Cytology Printed in 1J.S.A.

REVIEW ARTICLE
Multistation Multiparameter Flow Cytometry:
A Critical Review and Rationale'
Howard M. Shapiro
Flow Cytometry Laboratory, Sidney Farber Cancer Institute, and Department of Pathology, Harvard Medical School,
Boston, Massachusetts 02115

The capacity for fluorescence excitation cusses the history of the instrumentation,
by beams of different wavelengths at sepa- the parameters now measurable and the
rate points along the sample stream, and the probes used for their measurement, and the
capacity for computer analysis of multipar- methods for data analysis. Required sensi-
ameter data thus obtained, are now avail- tivity and precision are discussed, leading to
able in flow cytometer/sorter systems from the conclusion that many of the advantages
commercial producers. It is now readily ap- of multistation, multiparameter flow cytom-
parent to most experienced users of flow etry can be made available in less complex
cytometers that such multiparameter anal- and less costly instruments using less pow-
ysis offers the most convenient solution to erful sources and less elaborate computer
the problem of characterizing subpopula- hardware than are presently incorporated
tions of cells within a mixed population. The in commercial apparatus.
use of multiple beams facilitates resolution
of fluorescence signals from several probes
within or upon a single cell and widens the Key Terms: Flow cytometry, fluorescent
range of analytical alternatives available to dyes, multiparameter analysis, spectropho-
experimenters. This critical review dis- tometry

The development of flow cytometric apparatus, and of bodies; descendants of this apparatus are now marketed by
related methodology for cell preparation and data analysis, Becton-Dickinson (B-D FACS Systems, Sunnyvale, CA). The
has largely occurred in response to the specific needs of Coulter EPICS cell sorter and its predecessors (Coulter Elec-
researchers in a relatively small number of academic and tronics, Hialeah, FL) were developed by a group originally led
industrial laboratories. As a result, much of the design of the by Fulwyler, who, with Mullaney, Steinkamp, Van Dilla, and
first generation of instruments was dictated by the particular others, built instruments at Los Alamos Scientific Laboratory
tasks to which the apparatus was to be applied. T h e Cytofluo- which were used for studies of cancer cytology and tumor cell
rograf series of instruments now made by Ortho (Ortho Di- kinetics (55, 150, 179). T h e Impulscytophotometer of Dittrich
agnostics Systems, a division of Johnson & Johnson, West- and Gohde (48, 56), originally commercialized by Phywe AG
wood, MA) descends from the rapid cell spectrophotometer (Gottingen, FRG), and now sold by Ortho, was also intended
built by Kamentsky and his colleagues a t IBM's Watson primarily for cell kinetic studies.
Research Laboratory and used for studies on the automation By late 1979, there were approximately 500 flow cytometers
of cancer cytology (75, 76, 84). The Fluorescence Activated and sorters in use in research laboratories worldwide. Most of
Cell Sorter (FACS) developed by Herzenberg et al. a t Stan- these instruments were early model Cytofluorografs and Im-
ford (24, 69, 70) was designed primarily to count and isolate pulscytophotometers. They lacked sorting capability and
lymphocyte subpopulations stained with fluorescent anti- were, for the most part, used for studies of cell kinetics based
on analysis of DNA content (56, 60). The majority of sorters
' Portions of this work were supported by National Institutes of in use were B-D instruments equipped with 2-watt argon ion
Health Grants AM21094 and CA 19589, and by National Science lasers; the 488 nm line emitted by these sources was used to
Foundation Grant PCM-7923472.
Portions of this work were presented at the VIII Conference on excite fluorescence of fluoresceinated antibodies (98). FACS
Analytical Cytology and Cytometry, Wentworth-by-the-Sea, New systems with 4- to 5-watt argon lasers emitting 40-120 mW in
Hampshire, May 19-25, 1981. the ultraviolet were used for analysis and sorting of cells
227
228 SHAPIRO

vitally stained for DNA with one of several bisbenzimidazole as many as eight parameters were recorded with a dedicated
compounds developed by Hoechst AG (Frankfurt, FRG). minicomputer system (42, 133) in what later came to be called
While most of the Impulscytophotometers could measure list mode. The apparatus also had the capacity for making
only a single fluorescence parameter, the Cytofluorografs real-time classification decisions although sorting hardware
could measure light scattered forward and two fluorescence was not included. The first of these Cytomat instruments
parameters. The FACS and Coulter’s TPS sorter could meas- derived its five illuminating beams from a single mercury arc
ure fluorescence and forward scatter, and could be equipped lamp (42); a later version (133) used low power argon and
to measure either orthogonal scatter or fluorescence at two helium-cadmium laser sources for UV, blue and green beams.
wavelengths or in two planes of polarization. Flow cytometers The multiple illumination beam principle was subsequently
at this level of complexity proved entirely adequate for count- incorporated into instruments built at Heidelberg (157), Han-
ing and sizing cells, for detection and sorting of cells bearing nover (50), Livermore (47), and Los Alamos (152); several of
a single surface antigen, and for characterization of cell kinet- these systems used more powerful lasers and could sort cells.
ics by DNA content analysis of otherwise homogeneous cell Fluorescence excitation by beams of different wavelengths
populations. For many other potential applications, however, at separated measurement stations became available in com-
the speed and precision with which the instruments could mercially produced flow cytometers in mid-1979. The princi-
make measurements of cells were irrelevant in the face of pal use of the dual-laser FACS at the National Institutes of
experimenters’ inability to determine which of the measured Health, Bethesda, MD, equipped with argon and krypton ion
cells belonged to a particular subpopulation of interest. lasers, has been for analysis and sorting of cells labeled with
In principle, one could study the kinetics of a subpopulation fluorescein- and rhodamine-conjugated antibodies (7,67,130).
defined by its antigenic characteristic by sorting cells bearing The Stanford group, which initially attempted to excite both
the identifying antigen, and then restaining them with a DNA fluorescent tags using a multiline visible beam from an argon
fluorochrome and running them through the machine a second laser (97), now uses separated beams from an argon laser and
time. This is impractical; sorting runs typically take hours, a dye laser for dual-antibody studies (62a, 117). By the end of
while analyses of tens of thousands of cells can be done in a 1981, multistation instruments for multiparameter flow cy-
few minutes. It is now readily apparent to most experienced tometry and sorting, incorporating dedicated computer sys-
users of flow cytometers and sorters that multiparameter tems, had been delivered by Becton-Dickinson, Coulter, and
analysis offers the most convenient solution to the problem of Ortho, attesting to the emergence of a demand for sophisti-
characterizing subpopulations of cells within a mixed popula- cated instruments.
tion. In the present era of shrinking budgets, we can expect that
Until recently, most of the flow cytometers in use were a decreasing number of potential users of expensive apparatus
extremely limited in their capacity for two-parameter analysis will be able to afford multiparameter flow cytometers as they
and totally unequipped for analysis of three or more param- are now built and priced. It is my strong opinion that those of
eters. us who develop instruments, who have helped to nurture the
Multichannel pulse height analyzers, developed for nuclear demand for them, and who have the most experience with the
research, were first mated to flow cytometers at Los Alamos, associated techniques for sample preparation and data anal-
and are still widely used and generally satisfactory for record- ysis, are now obliged to expend some effort in making the
ing single-parameter data. Multichannel analyzers can also be technology more generally affordable, in order to best realize
used for rudimentary multiparameter analyses by using one its wide range of potential applications.
or more additional parameters with single channel analyzers In this, the first of a series of papers (140, 141), I will
and gating the multichannel analyzer to accumulate data only consider the existing range of cellular parameters and of
when the additional parameters fall within preselected ranges. probes which can be used for their measurement in the multi-
The Los Alamos group also used early, hardwired two-param- station, multiparameter flow cytometers now available com-
eter pulse height analyzers (37, 150, 151). Even in nuclear mercially. I will also give some attention to the sensitivity and
research, however, computers came to be recognized as offer- precision required for measurement of these parameters, and
ing greater con*:enience for analysis of two or more param- to factors which may influence selection of parameters and/
eters. Kamentsky’s original instrument incorporated a dedi- or probes where alternatives exist. The second paper (140)
cated minicomputer (75-77). By the late 1970’s, workers in will describe some experiments on the influence of various
several laboratories had described hardware and software for instrumental factors on overall performance; the third (141)
multiparameter analysis of data obtained from flow cytome- will describe construction of relatively inexpensive apparatus
ters (3, 79, 107, 126, 145, 156). for multistation multiparameter flow cytometry.
As long as only a single light beam was used for fluorescence
excitation, rather severe limitations were placed on the dyes
and dye combinations which could be used for multiparameter Parameters and Probes
flow cytometry. In order to escape from this constraint, my I have classified cellular parameters as “intrinsic” or
former colleagues and I developed a series of instruments in “extrinsic,” depending upon whether they can or cannot be
which cells stained with mixtures of fluorescent dyes passed measured without the use of reagents; I have also referred to
through as many as five illuminating beams of different wave- “structural” and “functional” parameters (138). A two-way
lengths, separated in space. Cells were tracked through the classification of parameters measurable by flow cytometry
system by hybrid electronics and correlated measurements of appears in Table 1.
 MULTISTATION MULTIPARAMETER FLOW CYTOMETRY 229
Table 1
Cellular Parameters Measurable b y Flow Cvtometrv urement, high resolution slit scanning (188), multiangle light
- scattering measurement with sectoral detectors (124), or
Parameters
~ _ _ _Structural _ _ _ _ Functional ~ polychromatic emission detection (184), I will not discuss
Intrinsic Cell size Redox state these analytical techniques further.
Cell shape
Cytoplasmic granular-
Simultaneous measurement of several extrinsic parameters
ity in a single cell requires the use of combinations of reagents. A
Pigment content (e.g. number of fluorescent dyes described for flow cytometry of
hemoglohin, chloro- extrinsic parameters in fixed and in intact cells are listed in
phyll) Table 2. The spectral characteristics of these dyes and the
Extrinsic Surface antigens Membrane integrity
excitation wavelengths available from various sources are
Lectin binding (surface Surface charge shown in Figure 1. Consideration of the data in this Table and
sugars) Figure makes apparent the rationale for the use of multiple,
Total protein Intracellular pH spatially separated fluorescence excitation beams. Even when
Basic protein Cytoplasmic/mitochon- one can choose from a number of probes to select those with
drial membrane poten-
tial desired spectral characteristics, the use of several beams fa-
Sulfhydryl groups Surface/cytoplasrnic re- cilitates resolution of fluorescence signals from multiple
ceptors probes. When such a choice is not available, multiple excita-
DNA content Enzyme activity tion beams may provide the only means of making correlated
RNA content Structuredness of cyto-
measurements of two parameters.
plasmic matrix (30)
Chromatin structure Membrane fluidity or Light scattering and extinction measurements of in-
microviscosity trinsic parameters: In most multiparameter instruments
Membrane-bound cal- Membrane permeability now in use, forward (small-angle) light scattering measure-
cium ion Endocytosis ments are used to obtain an indication of cell size. A substan-
DNA synthesis
tial body of literature documents the influence of many factors
other than cell size on small angle light scattering by particles
Depending upon the composition of a cellular sample and in the size range which includes most cells (81, 95, 96, 100,
upon the experimenter's goals, the same characteristic of a 124, 125, 127, 128, 134, 142, 143, 144, 186, 187).
cell can be used both to discriminate it from other cells in a Among the cellular properties other than size which influ-
mixed population and to define its physiologic state. In some ence forward scatter measurements are refractive index dif-
cases, the same molecular structure within or on a cell may be ferences between cells and the suspending medium, internal
defined by structural and functional approaches. For example, structure, and the presence within cells of material which
it was recently found that a cell surface glycoprotein com- absorbs strongly a t the illumination wavelength used. The
monly present on activated and leukemic lymphocytes, and difference in refractive index between intact cells and cells
first detected by antibody binding, is a receptor for transferrin with damaged membranes is commonly exploited in immu-
(161). nofluorescence analysis to exclude dead cells based on their
Intrinsic parameters have been used in combination in smaller scatter signals. This discriminates these cells, which
single-beam as well as in multistation instruments to aid in tend to stain nonspecifically with fluorescent antibodies, from
identification of cell subpopulations. After the demonstration intact cells bearing the surface antigen under study (95, 98),
by the Los Alamos group (125) that blood lymphocytes, but such discrimination is less than perfect (43).
monocytes, and granulocytes could be discriminated by a Forward scatter measurements are also strongly influenced
combination of forward (small-angle) and orthogonal (90" or by the wavelength used and by the range of angles over which
wide-angle) light scattering measurements, my colleagues and scattered light is collected, the latter being determined by the
I (132-134) used these parameters for automated differential focal lengths and numerical apertures of collecting lenses and
leukocyte counting. More recently, Hoffman et al. (68) incor- the size, shape and position of irises, slits, obscuration bars,
porated these measurements into an instrument system which and beam stops being used (96, 128, 142, 143, 144). Since the
does immunofluorescence analysis of lymphocytes in buffy design of optical systems used for forward scatter measure-
coat or whole blood samples without the customary prelimi- ment differs considerably from manufacturer to manufacturer,
nary step of Hypaque-Ficoll gradient centrifugation. Forward one cannot assume that the same results will be obtained from
and orthogonal scatter measurements have also been used, such measurements in different instruments.
alone or in combination with lectin binding or autofluores- Theory predicts that the relationship between forward scat-
cence measurements, to enrich hematopoietic cell populations tering amplitude and particle size will not be monotonic;
by sorting (110, 129, 178, 186). experiments with particles of known sizes produce results in
All of the multiple illumination beam flow cytometers de- good agreement with this prediction (e.g., (128)). Thus, while
scribed to date have employed the same basic optical geome- it is now common practice to assume that forward scatter
try used in all commercial flow sorters, with the axes of signals produced by intact cells of a single type in any given
illumination and collection optics orthogonal to each other instrument are proportional to cell size, this assumption may
and to the direction of sample flow. Since the commercial be unwarranted. Forward scatter measurements are not a t all
instruments now in widespread use do not have measurement reliable for estimating relative cell sizes of cells of different
or data analysis capabilities for electronic cell volume meas- types, particularly if there are substantial differences in inter-
230 SHAPIRO

Table 2
Fluorescent Probes for Extrinsic Parameters
Spectrum: Parameters Measured

Surface Membrane Enzyme ac-

Emis- Total DNA con- RNA con- Membrane
Excitation structures/proper- tent integrity PH potential tivity
sion protein tent
ties
uv Blue SITS SITS Hoechst ADB’. ’ coumarin
dyes; based sub-
DAPI strates
uv Green DANS DANS ADB’ naphthol
based sub-
strates
Violet-blue Green FITC FITC olivomycin, FDA, FDA’, oxacarbocy- fluorescein
mithra- COFDA COFDA’ anines based sub-
mycin, strates
chromo-
mycin
Blue-green Green FITC FITC A02 FDA, FDA’, oxacarbocy- fluorescein
COFDA COFDA? anines; based sub-
rhodamine strates
123
UV to yellow Red SRlOl ethidium ethidiurn,
propi- propidium
dium
Blue-green Red TRITC; phycoery- TRITC AO’, pyronin indocarbo-
thrin cyanines
Green-yellow Red XRITC, Texas thiacarbocy- resorufin
Red; phycocy- anines, ox- based sub-
anin adicarbo- strates
cyanines
Red Red allophycocyanin oxazine I indodicarbo-
cyanines,
thiadicar-
bocyanines
’ Abbreviations: ADB, 1,4-diacetoxy-2,3-dicyanobenzene(177); AO, acridine orange; COFDA, carboxyfluorescein diacetate; DANS, dansyl
chloride; DAPI, 4’,G-diamidino-2-phenylindole; FDA, fluorescein diacetate; FITC, fluorescein isothiocyanate; SITS, 4-acetamido-4’-isothio-
cyanatostilbene-2,2’-disulfonicacid; SR101, sulphorhodamine 101; TRITC, tetramethylrhodamine isothiocyanate; XRITC; rhodamine 101
isotbiocyanate. Probes listed under “Surface Structures/Properties” are covalently bound to antigens, ligands, etc.
Requires emission measurement at two wavelengths.
:’Requires excitation at two wavelengths.

nal structure. For example, peripheral blood granulocytes, measurement space by slight alterations in optical geometry
which are larger than lymphocytes based on hematocrit or and concluded that slight changes in the range of collection
electronic cell volume measurement, give smaller forward angles occasioned by our modification of the system to make
scatter measurements in at least some instruments (134). measurements a t different wavelengths were as much respon-
Using Cytomat flow cytometers (42, 133), my colleagues and sible for the different results as were changes in wavelength.
I (unpublished results) found we could alter the relative This conclusion was reinforced by a report by Loken et al.
amplitudes of forward scatter signals obtained from these two (96), in which measurements were made using a modified
cell types by making relatively small changes in optical ge- FACS. Cell cluster positions were affected by small changes
ometry. in collection angle when forward scatter measurements were
Recent published work by others (100, 186) has addressed made at a fixed wavelength.
the use of forward scatter measurements at different wave- The discordant observations discussed above and personal
lengths for discrimination of blood cell subpopulations. The experience suggest to me that the apparent utility of multiple
results obtained by others, using B-D FACS systems, differ angle or multiple wavelength forward scatter measurements
from those we obtained with the Cytomats. Using the FACS, in discriminating cell populations may well be offset by the
better resolution of cell subpopulations was obtained when difficulty of achieving comparable results using different in-
488 or 514 nm illumination was used than when ultraviolet struments. The combination of forward and orthogonal scatter
(351/363 nm) illumination was derived from the same argon measurements, mentioned previously, produces more con-
ion laser source. In contrast, we consistently obtained better sistent results which appear less instrument dependent and is
resolution of blood cell types in unstained samples using therefore coming into wider use in the study of mixed cell
ultraviolet light (325 nm from a He-Cd laser) than was ob- populations.
tained using the 441 nm blue line from the same laser or the While forward scatter pulse amplitudes have numerous
514 nm line from an argon laser. We were able to change the limitations as indicators of cell size, time-of-flight or pulse
relative positions of cell clusters in a two-dimensional width measurements derived from forward scatter signals
 231

ISSION nm 300 400 500 600 800

FIG.1. Spectral characteristics of some fluorescent dyes and excitation sources usable for flow cytometry. Excitat,ion (halftone) and emission
spectra (solid lines) are compressed; the top line of each defines a 15 nm region around the peak, the middEe line connects the half-maximum
points, and the bottom line connects the points at which excitation or emission falls to one-tenth maximum. Abbreviations are explained in the
legend for Table 2. Hoechst 33342, DAPI, mithramycin, ethidium, and acridine orange spectra are given for DNA-bound dye; spectra for
for dye in octanol, simulating characteristics of membrane-bound dye. Hoechst 33258, 33378, and 33662 have spectra similar to
2. Olivomycin and chromomycin spectra are similar to that of mithramycin. Propidium exhibits absorption and emission maxima
approximately 15 nm above the respective peaks for ethidium. Sulforhodamine 101, Texas Red, and XRITC, all rhodamine 101 derivatives,
have essentially the same spectrum. Source emission intensities are indicated in schematic form; a mercury arc lamp, with appropriate filters,
can deliver 10-20 mW a t each of the lines indicated, while the argon and krypton lasers typically supplied with flow cytometers emit 40-100
mW a t all lines indicated and considerably more at strong lines such as the 488 nm argon line. Helium-cadmium lasers currently available emit
2-10 mW a t 325 nni and 3-40 mW a t 441 nm; 2-10 mW helium-neon lasers, emitting a t 633 nm, are inexpensive and widely available. Spectral
data was obtained from References 28, 31, 34, 39, 72, 114a, 119, 120, 148, 153, 154, and 158.

have been shown to be useful for cell size and shape measure- from a cell; while relatively little advantage has been taken of
ment by several groups of investigators working with different this effect in research flow cytometers, Technicon (Tarrytown,
instruments (90, 142, 143, 144). The capability for making NY) has incorporated scatter measurements in different wave-
such measurements is presently offered by Coulter and Ortho. length bands obtained from a quartz-halogen source in the
The presence of material which strongly absorbs the inci- Hemalog D differential leukocyte counter (101, 115). This
dent light will generally diminish the forward scatter signal instrument classifies leukocytes based upon the development
232 SHAPIRO

of cytoplasmic staining following the action of cellular en- be detectable in a flow cytometer with proper excitation, e.g.,
zymes on chromogenic substrates. Kaplow and Lerner (79,80) the violet lines from a krypton laser.
have achieved similar results by analysis of forward scatter Autofluorescent pigments such as lipofuscins accumulate
and extinction signals from leukocytes treated with chromo- within cells with age. Jongkind et al. have used such autoflu-
genic substrates, using an Ortho instrument with 633 nm orescence to sort cells of presumptively different ages for
illumination from a He-Ne laser. At present, Ortho’s systems subsequent biochemical analyses (74a).
are the only commercially available cell sorters with extinction
measurement capability. In the original multibeam instru- Fluorescent Probes for Measurement of Extrinsic
ment my colleagues and I described (42), we could discrimi- Parameters
nate erythroid cells from leukocytes by measurement of he- Surface structures and properties: Structural param-
moglobin absorption in the Soret band, using high aperture eters such as surface antigens and lectin-binding sites (which
optics and an arc source. If appropriate lenses were used, the are specific sugar moieties), as well as functional parameters
system could also measure absorption of nucleic acids a t 260 such as surface charge and surface receptors for hormones,
nm. In the later, laser-equipped instrument (133), hemoglobin growth factors, neurotransmitters, viruses, etc. can be ana-
measurements were attempted by integrating the extinction lyzed using specific ligands fluorescently tagged with one of a
signals at 325 nm; this proved unsatisfactory in a brief clinical number of compounds which form covalent bonds with the
trial. A more definitive assessment of the comparative accu- ligand. As can be seen from Table 2, suitable tags can now be
racy of high resolution absorption measurements and extinc- chosen for use with excitation in the ultraviolet, blue, green,
tion measurements using the orthogonal configuration com- and yellow regions of the spectrum. Within a short time,
mon to commercial cell sorters could be obtained by using the reactive cyanine dyes and other labels suitable for use with
krypton laser lines near the Soret band, rather than the red excitation from helium-neon lasers should become avail-
admittedly suboptimal 325 nm line which we used for hemo- able as well.
globin measurements. Fluorescein, usually conjugated as the isothiocyanate
Fluorescence measurements of intrinsic parameters: (FITC), is by far the most popular fluorescent tag. Comment
The autofluorescence of most mammalian cells is due to the on flow cytometry of cells stained with fluoresceinated anti-
presence of pyridine and flavin nucleotides, which, respec- bodies is scarcely necessary since it has probably motivated
tively, impart UV-excited blue and blue-excited green fluores- the purchase of the majority of cell sorters now in use (98).
cence to cells (6,17). Autofluorescence of neutrophil(l86)and Studies of fluoresceinated lectin binding to cell surfaces are
eosinophil (187) blood granulocytes has been used as an subject to most of the same constraints as apply to work with
identifying parameter, in combination with forward and or- fluoresceinated antibodies (151).
thogonal scatter, to permit isolation of these cell types in high Valet et al. have used fluoresceinated polycations for
yield and purity by sorting. measurement of cell surface charge (175, 176). This method is
The fluorescence of the nucleotides is a function of oxida- somewhat more informative than is measurement of electro-
tion-reduction or redox state, and measurements of NADH phoretic mobility because, by varying the size of the tagged
fluorescence have been used to monitor redox states of cells polycation used, one can define the accessibility of various
and organs since the late 1950’s (32, 33, 83, 85, 131). In 1979, charged sites as well as the total number.
Thorell reported the use of a modified Ortho instrument with Fluorescently tagged hormones and other ligands have been
an argon ion laser operating in the ultraviolet for flow cyto- used for the study of cell surface receptors for growth factors
metric measurements of redox state in yeast and liver cells (146), chemotactic peptides ( l l n ) , insulin (108, 146), low-den-
(165). The instrument has since been refined to allow simul- sity lipoprotein (57), and histamine (116), among others. In
taneous measurement of scatter, absorption, and fluorescence general, attention has been paid to demonstrating the pres-
of extrinsic fluorochromes (166). More recently, Hafeman et ence of, and quantitating the number of, such receptors. In
al. used flow cytometry of redox state in blood neutrophils to principle, studies of the kinetics and thermodynamics of spe-
demonstrate that the respiratory burst induced by chemotac- cific ligand binding to cell surface receptors, such as are now
tic stimuli and phorbol esters is an all-or-none event (61). routinely done using radiolabeled ligands, can be done using
Plant photosynthetic pigments also have a characteristic fluorescently tagged ligands and flow cytometry. Bohn (21,22,
autofluorescence, typically showing excitation and emission 23) has presented the details and advantages of this approach.
maxima a t longer wavelengths than those of pyridine and Tetramethylrhodamine isothiocyanate (TRITC) is com-
flavin nucleotides. Trask et al. (168) have used the spectral monly used as a second label for two-color immunofluores-
characteristics and amplitudes of autofluorescence with for- cence studies done with the microscope; blue light is used to
ward and orthogonal scatter signals for analysis of mixed excite the green fluorescence of FITC-conjugated antibodies,
populations of phytoplankton. while green light is used to excite the red-orange fluorescence
In early erythroid development of bone marrow cells, por- of the TRITC-conjugated material. TRITC can be used for
phyrins may accumulate in cells faster than does heme. While flow cytometry as a single label with excitation from the 515
heme is nonfluorescent due to the quenching of the intrinsic nm line of an argon laser, but the simultaneous analysis of
fluorescence of porphyrin by iron, both the porphyrin precur- cells stained with FITC- and TRITC-conjugated antibodies
sors of heme and the zinc protoporphyrin formed when iron or ligands presents several problems. As can be seen from
accumulation is impaired (as in iron deficiency and lead Figure 1, TRITC is very poorly excited at 488 nm, while this
intoxication) fluoresce. One would expect such fluorescence to wavelength is nearly optimal for FITC excitation. FITC emis-
 MULTISTATION MULTIPARAMETER FLOW CYTOMETRY 233
sion extends toward the red end of the spectrum sufficiently 488 nm is several times more intense than the fluorescence of
that a detector equipped with filters which will pass only the an equimolar solution of fluorescein excited at the same
longer wavelengths of TRITC emission will still be sensitive wavelength.
to FITC emission. If one can be sure a priori that cells to be In a multibeam instrument, the yellow-excited fluorescence
analyzed will be more intensely stained with the TRITC- of conjugates of phycocyanin and/or the red-excited fluores-
conjugate than with the FITC-conjugate, it is feasible to do cence of allophycocyanin conjugates could be discriminated
two-color immunofluorescence work with this dye combina- from fluorescein and phycoerythrin fluorescence, allowing si-
tion in a flow cytometer, using 488 nm excitation or broadband multaneous analyses of four antigens, receptors, etc. Allophy-
visible excitation from an argon laser and electronic compen- cocyanin, derived from Anabaena uariabilis, appears well
sation to reduce crosstalk between the two fluorescence signals suited (absorption maximum 650 nm, emission maximum 660
(97). The FITC/TRITC combination is spectrally well nm) for use with He-Ne laser excitation at 633 nm. Phycobi-
matched for studies of energy transfer between the fluoro- liprotein immunofluorescent labels should be available soon
phores, and analysis of the spatial distribution of lectin binding from at least one commercial source (Becton-Dickinson).
sites on cell surfaces has been done by Chan et al. using these Total protein content: Total protein content of fixed cells
tags to label concanavalin A (31). can be estimated by staining with a variety of acid dyes which
In general, however, two color immunofluorescence studies bind ionically or covalently to positively charged groups on
are more readily accomplished when FITC is combined with proteins. Freeman and Crissman (52) examined several dyes
a label other than TRITC. The other labels described to date suitable for argon ion laser excitation at 488 nm and found
have excitation maxima sufficiently far from that of FITC to FITC most suitable as a total protein stain; FITC can be
necessitate the use of a second excitation beam. Two com- combined with propidium iodide for simultaneous protein/
pounds most frequently used are both derivatives of rhoda- DNA staining (37, 41). The combination of sulforhodamine
mine 101; XRITC (Research Organics, Cleveland, OH) is 101 (SR101) and DAPI was selected by Stohr et al. (158) for
conjugated by an isothiocyanate linkage, while Texas Red use with a dual source system in which the DNA stain DAPI
(166a) (Molecular Probes, Junction City, OR) incorporates a was excited by an UV beam, and SRlOl was excited by a blue
sulfonyl chloride as the reactive group. The spectral charac- or green beam. The SRlOl/DAPI combination can also be
teristics of both labels are essentially those of rhodamine 101; used with single-beam UV excitation (54). Both FITC and
they can be excited by the 568 nm line of a krypton laser or SRlOl staining procedures are rapid; FITC is covalently
by a dye laser nearer the optimal wavelength of about 590 bound while SRlOl binds ionically. Neither these dyes nor
nm. An illustration of the rapidity with which the field of other acid dyes conventionally used for protein staining pen-
multistation flow cytometry is developing comes from the fact etrate the membranes of unfixed cells to any significant extent.
that papers describing conjugation procedures for XRITC and Some uncharged dyes, such as rhodamine B, enter live cells
Texas Red had not appeared in print by the end of 1981; and stain cytoplasmic structures but their specificity as pro-
word-of-mouth and the manufacturers' product information tein stains has not been demonstrated under these circum-
have provided a reasonably large community of users with the stances.
only data available. Staining of total protein in fixed cells with covalently bound
Two color immunofluorescence measurements require dye probably became popular because cells could be washed
staining procedures different from the indirect staining pro- to eliminate background fluorescence from unbound dye.
cedure most commonly used for studies of a single surface Staining with sulfonated dyes is an equilibrium process; the
antigen. The direct labeling of two antibodies with FITC and dye, which is ionically bound to protein, would be removed
XRITC or Texas Red offers one alternative. A second alter- during a wash step. In experiments with glutaraldehyde-fixed
native is the use of XRITC- or Texas Red-avidin and biotin- cells, my colleagues and I (132, 134) were able to achieve
conjugated antibodies (16) or of hapten-conjugated antibodies relatively low background by working with the sulfonated acid
and differently tagged antibodies against the two haptens dyes LN (132) and brilliant sulfaflavine (BSF) in buffered
(190). Further discussion of antibody staining procedures is solutions of relatively low ionic strength (10 d), in which
outside the scope of this paper. I should, however, note that fewer anions competed with the dyes for binding sites on
there is now a clear trend toward the development of faster protein.
staining protocols which require few or no wash steps (68, Protein staining by acid dyes is dependent on the ionization
68a). constants of the dyes and of charged groups on protein. BSF
A new class of fluorescent probes applicable for immuno- for example is used as a stain for basic protein at pH 8 and as
fluorescence analysis and studies of other cell surface struc- a total protein stain a t pH 2.8 (92). Anilinonaphthyl and
tures has recently been described by Oi et al. (114a). The fluorescein compounds can be used to demonstrate sulfhydryl
authors prepared conjugates of naturally occurring, highly groups in proteins (36, 78).
fluorescent phycobiliprotein photosynthetic pigments from There is some evidence that, at least in fixed cells, meas-
red and blue-green algae and used the conjugates to label urements of orthogonal (90") light scattering may provide an
lymphocyte surface antigens. Using electronic compensation acceptable estimate of total protein content. In glutaralde-
(97), good resolution of signals was obtained when fluorescein hyde-fixed peripheral blood cells, total protein, as estimated
antibodies and antibodies linked to the algal protein phycoer- from the fluorescence of the UV-excited dye LN, and orthog-
ythrin were excited simultaneously by a single 488 nm beam. onal light scatter are tightly correlated, on a cell by cell basis,
The red (-575 nm) fluorescence of phycoerythrin excited at in the several classes ofleukocytes. This is illustrated in Figure
234 SHAPIRO

U MONOCYTES Hoechst 33258 and chromomycin has been used to differen-

0 ' tiate chromosomes and bacteria based on differences in DNA
NEUTROPHILS base composition (27, 180). Mithramycin has also been used
w in combination with ethidium bromide to increase fluores-
cence yield and improve specificity of staining (56, 163).
Under carefully controlled conditions, the green fluores-
cence of acridine orange (AO) yields a precise estimate of
DNA content in a t least some cell types (44,111,167). In some
circumstances, A 0 can be used in this manner to stain live
cells (111).The advantage of the method lies in the simulta-
neous indication of RNA content given by the metachromatic
orange fluorescence of AO. Principal disadvantages of proce-
dures using A 0 are the difficulty of achieving proper staining
: conditions and the overlap of A 0 fluorescence with that of
4 4 1 nm ORTHOGONAL SCATTER most other fluorescent tags, restricting use of A 0 in dye
FIG.2. Joint frequency function of 441 nm orthogonal scatter and combinations.
UV-excited blue fluorescence of the stilbene disulfonic acid protein While chromomycin, mithramycin, and olivomycin may
stain LN in a population of approximately 2000 leukocytes stained enter some living cells (26, 74), Hoechst compounds 33342,
with LN, ethidium bromide, and brilliant sulfaflavine after glutaral- 33378, and 33662 (but not 33258) are, at present, the only dyes
dehyde fixation and measured in the Cytomat laser source instrument which have been widely used for estimation of DNA content
(132-134). The strong correlation between orthogonal scatter meas-
urements and total protein estimates made with fluorescent dyes in living cells with retention of cell viability (5, 62, 82, 99, 102,
apparently exists in other cell types (see text). 137, 160, 189). From a practical point of view, this means that
an instrument designed to measure DNA content in living
cells must include an UV light source for excitation of Hoechst
2. While publications on the topic have not yet appeared, dyes.
others (H. Crissman, personal communication; M. Pallavicini, RNA content: The use of A 0 for simultaneous estimation
personal communication) have noted that fluorescent protein of DNA and RNA content has been extensively described (44,
stains and orthogonal scatter measurements yield similar dis- 111, 167). Tricyclic heteroaromatic homologues of AO, such
tributions. In beginning studies on a new cell population, it as pyronin Y and oxazine 1, can also be used for RNA content
would make sense to examine the correlation between orthog- estimation in combination with dyes such as methyl green
onal scatter and total protein as determined with fluorescent and Hoechst 33342 (118, 137, 162). The Hoechst 33342/py-
dyes; if the orthogonal scatter measurement proved usable, a ronin Y stain can be used with fluoresceinated antibodies for
fluorescence channel would be made unnecessary or could be DNA/RNA/immunofluorescence measurement in a dual-
used for measurement of a more specific parameter such as a beam (UV and 488 nm) flow cytometer using appropriate
surface antigen or intracellular enzyme. emission filters (137). A UV-excited, blue fluorescent dye
DNA content: Little need be said about stains for DNA related to DAPI is said to be specific for RNA (183) but its
since they are familiar to most users of flow cytometers and use in flow cytometry has not been described.
sorters. The acriflavine Feulgen procedure originally used for Chromatin structure: Darzynkiewicz et al. (44) describe
quantitative fluorimetric DNA content analysis (169, 191) has the use of A 0 metachromatic fluorescence to demonstrate
been supplanted by simpler staining procedures using fluoro- native (green) and denatured (red) DNA following acid or
chromes with varying degrees of specificity for DNA (38, 39, heat treatment. This procedure allows detection of chromatin
40, 44, 53, 89, 137, 163). Ethidium bromide and propidium structural differences in lability between proliferating and
iodide, popular because they can be used with relatively quiescent cells and unambiguous identification of mitotic cells.
simple staining procedures (53, 163), bind to double stranded Dye combinations such as Hoechst 33342/pyronin Y may be
RNA as well as to DNA. More specific staining of DNA with usable for similar studies. If fixation does not substantially
these dyes has been achieved by treatment of fixed cells with alter chromatin structure, combinations of A-T and G-C bind-
RNAse (56, 163); the combination of propidium iodide with ing dyes, such as Hoechst 33258 and chromomycin A I (27)
fluoresceinated antibodies allows simultaneous analysis of might give differential staining of cells with different degrees
DNA content and immunofluorescence in a single beam (488 of chromatin compaction.
nm) flow cytometer (25, 182). Membrane-bound calcium ion: Chlorotetracycline
Two groups of compounds which stain DNA more specifi- (CTC) has been used as a fluorescent probe for detection of
cally than ethidium and propidium are widely used. The first release of Ca" from intracellular membranous structures
of these includes DAPI and Hoechst compounds 33258,33342, which occurs almost immediately after activation of a number
33378, and 33662 which are all UV-excited blue fluorescent of cell types by cell surface ligand-receptor interactions or by
dyes that bind preferentially to adenine-thymine rich regions ionophores (28, 34, 105, 109, 136). The CTC-Ca" chelate is
of DNA (89). The second group includes chromomycin, mith- considerably more fluorescent than CTC. The fluorescence of
ramycin, and olivomycin, which give yellow-orange to green the chelate is further enhanced in the hydrophobic environ-
fluorescence when excited with violet or blue light and which ment of a lipid membrane, and spectral shifts occur in addition
bind preferentially to guanine-cytosine rich regions of DNA to fluorescence enhancement. CTC fluorescence in intact cells
(38, 39, 89). Staining with combinations of dyes such as can be measured flow cytometrically using UV or violet exci-
 MULTISTATION MULTIPARAMETER FLOW CYTOMETRY 235
tation (136). More specific probes for cytoplasmic free calcium tion of pH. Calibration is done by measuring cells in buffers
have been developed by Tsien (170-172) and may be usable of known pH after addition of gramicidin to abolish the
for flow cytometry with 325 nm excitation from He-Cd lasers. transmembrane pH gradient. Alabaster (2) has also recently
Membrane integrity: A crude indication of viability is reported on results obtained using ADB for pH measurement.
given by the cells’ capacity t,o exclude dyes such as trypan Gerson (55a) has used a fluorescence emission ratio tech-
blue, eosin, or ethidium. Although some cells which have lost nique similar to that described by Valet et al. (177) for pH
reproductive viability following exposure to drugs may not measurement. 4-Methylumbelliferone, as the free base or as
admit dye for several days after treatment (122), a cell which liberated from the acetate, rather than ADB, is employed as
will admit dye is almost certainly dead. An equivalent test the probe. Using other methods, it was shown (55b) that
involves the use of nonfluorescent esters such as fluorescein exposure of lymphocytes to mitogenic lectins is followed
diacetate (FDA) and carboxyfluorescein diacetate (COFDA). within minutes by a rise in intracellular pH. With the flow
These esters are uncharged and highly lipophilic and pass cytometric technique, it was shown that the cell interior is
readily across cell membranes. Once inside cells, they are more alkaline in proliferating than in quiescent cells. Since
rapidly hydrolyzed by nonspecific esterases to yield the both ADB and methylumbelliferone fluorescence spectra sub-
strongly fluorescent fluorescein and carboxyfluorescein anions stantially overlap the fluorescence spectra of Hoechst 33342
(123). Cells with intact membranes retain the anions. Addition and related dyes, the relatively cumbersome three-beam ap-
of a mixture of FDA and propidium to presumably viable cells proach mentioned above remains the only method presently
produces green cytoplasmic fluorescence in cells with intact practical for direct correlation of intracellular pH and cell
membranes, and red nuclear fluorescence in cells with dam- cycle position.
aged membranes. Numerous variations of the dye exclusion Structuredness of cytoplasmic matrix (SCM): This
test have been described for use in flow cytometry (21-23,43, term is used by Cercek and Cercek (29, 30) to describe a
46, 51, 62, 155, 160, 177). The use of propidium iodide in parameter obtained by fluorescence polarization measure-
immunofluorescence measurements on intact cells would ap- ments of cells incubated with FDA. Changes in SCM have
pear to be preferable to scatter gating for dead cell discrimi- been observed to occur within minutes following exposure of
nation (43). lymphocytes to mitogenic lectins and antigens (29,30,63,155,
Intracellular pH: FDA and COFDA can be used to derive 174). The physiocochemical bases of SCM and SCM changes
an estimate of the pH inside cells or subcellular organelles (64, are unclear. They could conceivably relate to changes in
108, 114, 164, 173, 181).The fluorescence excitation spectra of intracellular and/or membrane-bound calcium ion concentra-
fluorescein and carboxyfluorescein are pH-sensitive. As pH tion, pH, and membrane potential, all of which also occur
rises from 5.0 to 8.0, the efficiency of excitation at 430-450 nm shortly after lymphocytes are exposed to mitogenic stimuli
decreases while the efficiency of excitation at 480-490 nm (55b, 105, 106, 135, 136, 172). Parallel studies of FDA and
increases. The ratio of (fluorescence excited at 488 nm)/(flu- COFDA fluorescence polarization using dual wavelength ex-
orescence excited at 441 nm) thus varies monotonically with citation might clarify the mechanism of SCM changes.
pH over the range 5.0 to 8.0. The use of the ratio eliminates Membrane potential: Permeant lipophilic ionic dyes may
the influence of cell-to-cell differences in the amount of dye be used for flow cytometric estimation of membrane potential
present upon the estimate of pH. When this factor can be as has been described (45, 46, 51, 71, 93, 135, 136, 138, 139,
controlled, or when pH shifts are large, fluorescence measure- 176). Lipophilic cations such as cyanine dyes are distributed
ments with single wavelength excitation can be used for pH across membranes as a function of transmembrane potential;
estimation (108,181).I recently examined pH variation during the more negative the inside of the membrane-bounded com-
the cell cycle of CCRF-CEM cells stained with Hoechst 33342 partment, the more dye is taken up (135, 136, 148).
and FDA with UV excitation from an arc lamp, 441 nm from In cells with energized mitochondria, a potential gradient,
a He-Cd laser, and 488 nm from an argon laser. The cell interior negative, across the mitochondrial membrane further
interior appears to become more alkaline during progression favors dye uptake by cells. The lipophilic quality of the dyes
from G, through S to GZ. results in their binding to membranous or lipid-rich structures
The fluorescein anion produced by FDA hydrolysis enters in cells with high affinity. Therefore, dye concentration within
mitochondria; therefore the pH estimated by using FDA cells can vary considerably with modest changes in membrane
represents a weighted average of cytosol and mitochondrial potential. If cells are equilibrated with low concentrations of
pH. Carboxyfluorescein produced by COFDA hydrolysis does dye and examined without washing, intracellular fluorescence
not enter mitochondria; therefore, COFDA can be used to is a function of cytoplasmic and mitochondrial membrane
estimate pH in the cytosol (164). The pH of intracellular potential, provided dye concentrations are kept low to prevent
organelles can be estimated using fluorescein-tagged mole- quenching. If cells are washed, potential gradients favor reten-
cules which are sequestered within those organelles. FITC- tion of dye by the mitochondria, and intracellular fluorescence
dextran and protein conjugates have been used to study pH responds primarily to mitochondrial membrane potential (73,
in endocytic vesicles (114, 173). 136). Mitochondrial and cytoplasmic membrane potential ap-
A flow cytometric probe for pH which does not require dual pear to be increased in proliferating as opposed to quiescent
beam excitation has been described by Valet et al. (177). 1,4- cells (G,, S, and GZversus Go) (18a, 35, 45, 46, 58, 71, 93, 135,
Diacetoxy-2,3-dicyanobenzene (ADB) is an ester which is 136, 138, 139). Cells treated with some anticancer drugs show
cleaved by intracellular esterases to 2,3-dicyanohydroquinone; diminished mitochondrial membrane potential long before
this substance is excited by UV light. The ratio of fluorescence they appear nonviable by dye exclusion (18, 18a). Conversely,
at 500-580 nm to fluorescence a t 420-440 nm gives an indica- in cells killed by rupture of the cytoplasmic membrane, the
236 SHAPIRO

potential gradient across the mitochondrial membrane may DNA synthesis: Flow cytometry has become an indispen-
be preserved (46).The mitochondria-specific character of the sable technique for cell kinetic studies because it permits rapid
dye rhodamine 123 (45, 71, 154) is based on its potential- and precise measurement of DNA content in cell populations.
dependent uptake into these organelles (73, 136). Such measurements do not, however, in themselves yield
Oxonols (172) or dyes which undergo electric field or poten- information about the rate of DNA synthesis. Relatively new
tial-dependent spectral shifts without permeating the mem- flow cytometric procedures permit detection and quantitation
brane (94) may possibly be useful for estimating cytoplasmic of the incorporation of bromodeoxyuridine (BrdUrd) into
membrane potential subject to less influence from mitochon- DNA for estimation of synthetic rate. BrdUrd can be detected
dria than occurs with cyanines. Under normal circumstances, with antibodies (59), by its quenching of the fluorescence of
however, there is some interaction between cytoplasmic and Hoechst dyes (19, 20, 113) or (and perhaps and) by enhance-
mitochondrial membrane potentials, and the cyanines and ment of mithramycin fluorescence (161a), or by its effects on
rhodamine 123 remain useful, although neither as quantitative denaturability of DNA (44). None of the BrdUrd measure-
nor as predictable as could be desired. ment procedures described to date can be used with viable
Enzyme activity: Flow cytometric determination of en- cells.
zyme content and kinetics in intact and fixed cells is possible Vital staining: The availability of flow sorters has in-
using a variety of fluorogenic substrates (49, 185). Derivatives creased interest in development and use of stains for charac-
of coumarins, naphthols, and fluorescein yield UV-excited terizing intact cells. Among the parameters just discussed,
blue and blue-excited green fluorescent products, respectively. membrane-bound Ca2+,redox state, membrane integrity, pH,
Derivatives of resorufin yield green-excited orange fluorescent membrane potential, structuredness of cytoplasmic matrix,
products. It seems likely that oxazine or thiazine dyes con- membrane fluidity and permeability, and endocytosis are
taining phenolic groups could be esterified to produce fluoro- meaningful only as applied to intact cells. High-intensity laser
genic substrates usable with He-Ne (red) laser excitation. T h e light and fluorescent probes, alone or together, may injure
author is not aware of literature on this subject. cells. Even relatively safe reagents such as antibodies can
Membrane fluidity or microviscosity: Measurements of perturb cell behavior, e.g., by acting as mitogens (180a). It
the fluorescence emission anisotropy of the uncharged hydro- would seem good practice to minimize dye concentrations,
phobic probe diphenylhexatriene (DPH) bound to membranes exposure times, and excitation power, and to carry out as
yield an estimate of the mobility of the lipid bilayer compo- many procedures as possible in the cold, in order to keep
nents (147). Arndt-Jovin et al. described flow cytometric damage to viable cells a t a minimum. It remains up to the
measurements of this parameter in differentiating cells (4). experimenter, however, to determine the effects of the selected
Slow changes in membrane fluidity often reflect alterations in reagents and techniques on the behavior of the cell type(s)
membrane lipid composition, as in cholesterol/phospholipid under study. It may also be necessary, when working with
ratio, but this cannot explain the more rapid changes seen in fluorescent analogues of drugs, hormones, etc., to establish
cell activation (136). In principle, uncharged hydrophobic the effects of the fluorescent tag on the affinity and biological
probes other than D P H could be used for microfluidity meas- activity of the effector portion of the molecule. On the positive
urements if, for example, different spectral characteristics side, flow cytometry offers an unmatched resource for inves-
were desired. In practice, any difference in structure between tigations of the mechanisms of vital staining.
another probe and D P H would be likely to alter binding
characteristics making reconciliation of results with the two Data Analysis
probes difficult. Data obtained by flow cytometry are almost always dis-
Membrane permeability: Intact cell membranes do not, played in the form of one-dimensional (histogram) and two-
in general, offer unrestricted access to small molecules. Flow dimensional (contour plot, cytogram, raster or scatter plot)
cytometry has been used to quantitate uptake by cells of frequency distributions, either of measured variables or of
intrinsically fluorescent drugs (86) and is more generally ap- parameters derived by algebraic manipulation of measured
plicable for studies involving fluorescent analogues of drugs variables. The simplest example of multiparameter analysis
(102) and hormones which are transported across the mem- in flow cytometry is probably the use of strong signals, such
brane. Differences in permeability of morphologically similar as forward scatter signals, for “gating” parameters to improve
cell types may be exploited for cell classification; for example, detection of weak signals, e.g., immunofluorescence, above the
Loken showed that resting mouse B lymphocytes take up background noise level. Most commerical flow cytometers
Hoechst 33342 faster than do resting T cells (99). Lalande et incorporate this capability.
al. (87, 88) demonstrated increased permeability of activated Two-dimensional analysis also facilitates determination, a t
mouse T cells to this dye within a few hours after exposure to a glance, of the degree to which two variables are correlated.
mitogenic lectins or antigens. Hoechst 33342 uptake and im- If the same data were displayed on the X and Y axes of a
munofluorescence has been used to study activation of lym- scatter plot, all the displayed points would fall on a line at 45’
phocyte subpopulations in rats (189). The technique may also from both axes; a correlation coefficient of 1.0 would be
be applicable in man (J. Williams, personal communication). obtained by calculation. If the X and Y values of each data
Endocytosis: Fluorescently tagged macromolecules (104, point instead represented measurements of the same param-
108) and plastic microparticles (153, 176) have been used to eter made in different beams of a dual-beam instrument, we
study endocytosis or phagocytosis. Particularly when the plas- would expect to see deviations of the data points from the 45”
tic microparticle approach is used, it is possible to select line due to the influences of different, presumably uncorre-
reagents with almost any desired spectral characteristics. lated, sets of instrumental sources of variation on the two
 MULTISTATION MULTIPARAMETER FLOW CYTOMETRY 237

measurements made of each cell. The data of Figure 2, which taken for a cell to traverse the measurement stations. Addi-
shows measurements of LN fluorescence and orthogonal scat- tional logic must be incorporated in order to deal with coin-
ter in fixed human peripheral blood cells, in the author’s cident events, e.g., the arrival of a second cell a t the first
experience are about as closely correlated as are any set of station before the first cell has passed through all stations.
data points obtained from measurements of two ostensibly The control electronics for multistation instruments have
different cellular parameters. The general principle which been discussed by Hiebert et al. (66, 154). Commercial instru-
emerges from this example defines the author’s philosophy of ments employ similar circuitry which, while generally inac-
multiparameter analysis: wherever possible, the most expedi- cessible to the user, serves the purpose of making correlated
tious comparison of two candidate parameters as to suitability data from the several measurement stations available for the
for a particular task can be obtained by measuring both in the short-term analysis required for sorting decisions, as well as
same cells a t the same time. When both parameters provide for long-term storage and more elaborate computational ma-
approximately the same information, it is obviously preferable nipulation.
to use that which is easiest to measure. Kamentsky’s original instrument incorporated essentially
It is relatively easy to conceptualize the extension of two- as much computer capability as is now available or needed,
dimensional scatter plot analysis to three-dimensional spaces, and Kamentsky et al. early publications (76, 77) anticipated
in which ellipsoidal clusters are separated by plane-sided solid most later work on multiparameter analysis. The addition of
figures. Stiihr and Futterman (159) have described hardware multiparameter and of multistation measurement capability
and software for such three-dimensional graphical analysis. It to the present generation of commercial apparatus has been
is more difficult to envision analysis in spaces of more than significant because it has enabled users primarily concerned
three dimensions. Detailed expositions of the application of with biological applications to implement techniques previ-
multivariate statistical analysis to cytologic data can be found ously accessible only in institutions engaged in the develop-
in articles by Kamentsky (77) and Prewitt (121) and in a ment of flow cytometric instrumentation.
recent series of papers by Bartels (8-15).
Where multivariate statistics are applied for the purpose of Precision and Sensitivity in Flow Cytometry
identifying subpopulations of cells in a mixed population, the The designers of commerical flow cytometers must meet
objective is typically the definition of a discriminant function, the requirements of immunologists who want the capability
i.e., a function of all of the measured variables with a multi- of detecting a few hundred molecules of fluorescent antibody
modal distribution such that values of the function for all cells on a cell surface, and also the requirements of oncologists who
of the selected type are separated from values of the function want to measure DNA content in cells with a coefficient of
for all other cells in the population. Despite the fact that variation of less than 1%in the GI peak. Many of the practical
several parameters have to be measured to determine and means of increasing sensitivity sacrifice precision and vice
calculate values of the discriminant function, the distribution uersa. In some respects, multistation instruments become
of the discriminant function is one-dimensional. Given a linear easier to design because it is logical to tailor the DNA meas-
discriminant function, i.e., a linear function of several param- urement station for precision and the immunofluorescence
eters which discriminates one cell type from others, one could measurement station for sensitivity. A consideration of factors
use simple analog circuitry to compute the values of the responsible for these performance characteristics facilitates
discriminant and display them on a multichannel pulse height instrument selection as well as experimental design.
analyzer, or input the values to a single channel analyzer to McCutcheon and Miller studied the signal-to-noise behav-
control the counting and/or sorting of the selected cell type. ior of an orthogonal laser source flow cytometer using cali-
As there exist defined statistical procedures for calculating brated light pulses from an LED and actual signals from
discriminant functions, there also exist related procedures for fluorescent particles (103). They defined three regimes of
determining the extent to which any given parameter influ- signal-to-noise behavior based on signal intensity. At the
ences the overall discrimination between classes. There are lowest signal levels, relatively constant background noise from
also reasonably straightforward criteria for adding or substi- filter fluorescence, stray scattered light, and fluorescence of
tuting parameters based upon whether the change increases materials in the sample stream limits the resolution of signals.
the distance between clusters in hyperspace or between modal At signal levels typical of cells stained with fluorescent anti-
peaks of the discriminant function histogram. bodies, the precision of measurement improves as more pho-
In practice, however, the application of multivariate analy- toelectrons are collected from the stained cells. Sensitivity,
sis to a new cell population typically begins with minutes of obviously, can only improve if the signal from the cells is
data collection followed by hours of work using interactive increased relative to the signal from the background. The high
graphics programs for one- and two-dimensional data display intensity signals usually obtained from cells stained with
to partition multiparameter data files. The data analysis in- fluorochromes for DNA content measurement are far enough
volved could be done without computers. Digital computer above the background level for background and photoelectron
hardware, however, allows analysis to be done a t the investi- statistics to become negligible as sources of instrumental
gator’s convenience using stored data. This provides the pri- variance. Instead, spatial and temporal inhomogeneities of
mary rationale for the addition of computer capability to illumination caused by flow instability, cellular asymmetry,
commercial flow cytometers. optical misalignment, or poor design limit the precision with
In a multistation system, short-term analog storage, in the which measurements can be made. Detector dark current will
form of peak detectors and/or integrators controlled by timing not limit sensitivity under any conditions. As a practical
logic, is required in order to preserve data during the time matter, the use of expensive low noise photomultipliers, cooled
238 SHAPIRO

detector housings, etc. in flow cytometers adds cost to instru- and heights but with the same integrated intensity. Decreasing
ments without the possibility of attendant benefits and should h from 100 pm to 50 pm will halve the volume from which
be avoided. stray light is collected without necessitating the inclusion of
In a typical orthogonal laser source flow cytometer, the integrators in the processing electronics. A further increase in
illumination beam is focused to a round or elliptical spot at scatter and fluorescence sensitivity, without loss of precision,
the point of intersection with the core stream. For purposes of can be obtained if an image is formed of the intersection of
discussion, I will assume the vertical axis of an elliptical focal the beam and the core stream a t some point in the collection
spot to be coaxial with the axis of the core stream, and define system. An aperture placed in the plane of this image can be
the height of the spot as h, and the width of the spot (axis used to limit light reaching the detector from outside the
dimension perpendicular to the direction of flow) as w. If the observation volume thus defined, which will be approximately
radius of the core is r, the lateral position of cells in the core ar’h, or 3927 fl. Such an aperture also prevents excess scat-
stream can vary by this amount in either direction from the tered light from reaching a fluorescence detector.
axis of flow. The elliptical focal spot does not define a profiie In multibeam systems, particular attention must be paid to
of uniform illumination intensity; the intensity profile of the the optical filters used to limit fluorescence detector response
unfocused laser beam, in TEM,”, mode, is Gaussian, and the because interference is apt to come from excitation beams at
diameter of the beam is the distance between the l/e’ points wavelengths above and below the bandpass region of at least
within which roughly 87% of the beam intensity is contained. one detector. The use of image plane apertures with combi-
The elliptical contour of the focal spot similarly defines a nations of color glass and reflective filters is, in my experience,
region which contains this fraction of the total luminous flux the best way to minimize cross-talk between fluorescence
in the beam. In order to limit variations in the intensity with channels.
which cells are illuminated as a function of deviations in In a system designed for the study of objects smaller than
trajectory from the center of the core stream, it is necessary cells, e.g., bacteria, chromosomes, or viruses, sensitivity can
to make U I relatively large compared to r; it can be calculated, be increased by focusing the beam to a very small spot. In one
from tables of the Gaussian distribution, that, if w = 20r, there such instrument (65), a focal spot of 5 pm diameter and a 2
will be less than 1% variation in illumination intensity over pm core were used with a water immersed optical system and
the width of the core stream. Thus, for a 10 pm core diameter an image plane aperture to produce an observation volume of
( r = 5 pm), the focal spot must be 100 pm wide or wider to 15 fl; scatter signals from individual virus particles could be
limit signal variances due to this source of illumination inho- detected using less than 100 mW of laser power for illumina-
mogeneity to 1%). tion. In this configuration, the precision of the system, when
The collection optics in an orthogonal system will be ar- used to measure DNA content in cultured cells, was not equal
ranged with the optical axis perpendicular to both the axis of to that of commercial systems. With the optics adjusted to
flow and the axis of the illuminating beam. If the collecting increase spot size, a coefficient of variation (cv) of 0.7% for
lens is placed a t its focal length from the observation point, to the G1 peak was achieved.
collect light from cells without forming an image of the inter- Increasing the laser power would increase neither the sen-
section of the illuminating beam and core stream, scattered sitivity nor the precision of the system just described. Sensi-
and stray light from a volume considerably larger than that tivity would not be increased because background would go
occupied by the core will be directed toward the detector. If up in proportion to signal; precision would not be increased
the beam is round, i.e., h = w = 100 pm, and the sheath is 70 because the illumination profile and noise characteristics of
pm in diameter, with 10 pm core diameter ( r = 5 pm), this the laser source would not be changed by raising the power
volume will be greater than 270,000 femtol. Light scattered level. It is likely that, in instances in which precision is
from it will interfere with orthogonal scatter signals and must improved by increasing illumination intensity, the improve-
be filtered without producing filter fluorescence in order not ment results from changes in laser mode or noise, or from the
to interfere with fluorescence measurements as well. achievement of power levels high enough to produce complete
In order to increase the sensitivity of the instrument for bleaching of fluorochromed cells during transit of the beam,
measurement of weak fluorescence or for detection of small eliminating effects of illumination inhomogeneities. The latter
particles by orthogonal light scattering, it is obviously desir- technique, while perhaps acceptable for analysis, is not well
able to decrease the volume from which scatter signals reach adapted for vital cell sorting because of the increased like-
the detector. The beam width w must be kept a t 100 pm in lihood of cell damage as illumination intensity increases (82).
order to maintain precision. It is not necessary to keep h at None of the instrumental factors just described bear
100 pm; while the beam profile is Gaussian in the direction of strongly on the sensitivity of flow cytometers for detection of
flow, cells will pass through the entire profile as they transit fluoresceinated antibodies on cell surfaces. This is because
the beam. If h is a few times larger than a cell diameter, and cellular autofluorescence, from flavins and from substances
if the cells are relatively homogeneous in size, fluorescence which emit in the spectral region of fluorescein emission, when
signals obtained as cells transit the beam will be roughly excited at 488 nm, is both considerable and variable. An
Gaussian with peak height proportional to integrated inten- average lymphocyte behaves as if it had approximately 10,000
sity. If h is made smaller than a cell diameter, a slit scan will molecules of fluoresceinated antibody bound to its surface;
be obtained as cells pass through the beam. Large, weakly the cv of the autofluorescence distribution is perhaps 20% (91,
stained cells and small, brightly stained cells containing the D. Parks, personal communication). While typical commerical
same amount of dye will produce signals of different widths cell sorters can detect fewer than 3,000 molecules of fluores-
 MULTISTATION MULTIPARAMETER FLOW CYTOMETRY 239

cein (95),they cannot differentiate an antibody-stained weakly my colleagues and I set out to determine how many of the
autofluorescent cell from an unstained strongly autofluores- advantages of such systems could be incorporated into appa-
cent cell when fluorescein is used as the label. This problem ratus using light sources with more modest cost, power and
would best be solved by use of tags with longer excitation and cooling requirements and data analysis systems based on
emission wavelengths; workers with Raman microprobes have microcomputers and/or hardwired logic. The results of our
reported that interference from autofluorescence is greatly studies, which will be reported in subsequent papers (140,141)
reduced when red instead of green light is used (1). suggest that users of commercial flow cytometers, as well as
In instruments designed for high sensitivity such as the those who build their own instruments, may find our alter-
system used for virus studies (65), strong signals from lym- native technology affordable as well as desirable.
phocyte autofluorescence could be obtained using less than 10 I once wrote that “The true benefits of a multiple illumi-
mW excitation a t 488 nm. Experience with the Cytomat nation wavelength multiparameter flow cytophotometer can
instruments, using arc lamp (42) and low power laser sources perhaps be best appreciated only by those who have worked
(133), and more recent reports on studies of bacterial immu- with such an instrument” (133). I hope the number of those
nofluorescence using an arc source system (149) suggest that who appreciate the benefits by that criterion will continue to
10-20 mW lasers will be usable for most immunofluorescence increase as it has in recent years.
work if apparatus is equipped with efficient (N.A. z 0.5)
collection optics and good light shielding and if fiiters are Literature Cited
properly selected and used. A red-excited tag, such as an 1. Abraham JL, Etz ES: Molecular microanalysis of pathological
indodicarbocyanine- or allophycocyanin conjugate, used with specimens in sztu with a laser-Raman microprobe. Science
a helium-neon laser could potentially provide better sensitiv- 206:716, 1979
2. Alabaster 0, Spooner, C, Clagett K: Therapeutic implications of
ity, because of reduced autofluorescence, than is now achieved metabolic heterogeneity: The role of intercellular pH measure-
using fluoresceinated antibodies excited by multiwatt argon ment. Abstracts of the Conference on Cytometry in Clinical
ion lasers. In addition the reduction in cost and in power and Laboratories 4:5, 1982
cooling water requirements might in itself revolutionize im- 3. Arndt-Jovin DJ, Jovin TM: Computer-controlled multiparame-
ter analysis and sorting of cells and particles. J Histochem
munofluorescence flow cytometry. Cytochem 22522, 1974
Among all of the cellular parameters now measurable, DNA 4. Arndt-Jovin DJ, Ostertag W, Eisen H, Klimek F, Jovin TM:
content is the only cellular characteristic which is so constant Studies of cellular differentiation by automated cell separation.
as to require measurement to a precision of 1-2%. Measure- Two model systems: Friend virus-transformed cells and Hydra
ments of DNA content to this precision have, to date, been attenuata. J Histochem Cytochem 24:332, 1976
5. Arndt-Jovin DJ, Jovin TM: Analysis and sortmg of living cells
more readily achievable in arc source instruments than in according to deoxyribonucleic acid content. J Histochem Cyto-
systems with ion laser sources (56,137 149). The few milliwatts chem 25:585, 1977
of power in the ultraviolet needed to excite the Hoechst dyes 6. Aubin J : Autofluorescence of viable cultured mammalian cells.
and DAPI can also be obtained from He-Cd lasers operating J Histochem Cytochem 27:36, 1979
7. Bach M-A, Phan-Dinh-Tuy F, Bach J-F, Wallach D, Biddison
a t 325 nm. A measurement precision of 1 4 % is entirely WE, Sharrow SO, Goldstein G , Kung PC: Unusual phenotypes
compatible with 256-channel pulse height analysis. Most data of human inducer T cells as measured by OKT4 and related
from flow cytometers therefore need not be digitized to more monoclonal antibodies. J Immunol 127:980, 1981
than 8 bits’ resolution and can be collected, analyzed, stored 8. Bartels PH: Numerical evaluation of cytologic data. I. Descrip-
and displayed in systems based on 8-bit microprocessors with- tion of profiles. Anal Quant Cytol 1:20, 1979
9. Bartels PH: Numerical evaluation of cytologic data. 11. Compar-
out loss of information. ison of profiles. Anal Quant Cytol 1:77, 1979
10. Bartels PH: Numerical evaluation of cytologic data. 111. Selec-
Conclusions tion of features for discrimination. Anal Quant Cytol 1:153, 1979
The capability for multiple wavelength excitation of several 11. Bartels PH: Numerical evaluation of cytologic data. IV. Discrim-
fluorescent probes bound to the same cell provides experi- ination and classification. Anal Quant Cytol 219, 1980
12. Bartels PH: Numerical evaluation of cytologic data. V. Bivariate
menters with a wider range of analytical alternatives than distributions and the Bayesian decision boundary. Anal Quant
exist otherwise with only a moderate increase in complexity Cytol 2:77, 1980
of the instrumentation. The benefits of multiple wavelength 13 Bartels PH: Numerical evaluation of cytologic data. VI. Multi-
excitation apply to fluorescence microscopy and to static variate distributions and matrix notation. Anal Quant Cytol
2:155, 1980
microspectrophotometry as well as to flow cytometry. In the 14 Bartels PH: Numerical evaluation of cytologic data. VII. Multi-
field of flow cytometry, the large increase in cost typically variate significance tests. Anal Quant Cytol 3 1 , 1981
associated with configuration or upgrading of a commerical 15 Bartels PH: Numerical evaluation of cytologic data. VIII. Com-
instrument for multistation measurement represents the prin- putation of the principal components. Anal Quant Cytol 3:83,
cipal impediment to more widespread use of the technique. 1981
16 Bayer E, Wilchek M: T h e avidin-biotin complex as a tool in
The cost increase results from the present practice of adding molecular biology. Trends Biochem Sci 3:N257, 1978
an ion or dye laser as a second fluorescence excitation source 17 Benson HC, Meyer RA, Zaruba ME, McKhann GM: Cellular
and a fairly sophisticated computer system for data analysis. and autofluorescence-is it due to flavins? J Histochem Cyto-
Both the laser power and the computer power thus acquired chem 27:44, 1979
18 Bernal S, Shapiro H, Chen LB: Measurement of mitochondrial
may be more than necessary for the tasks at hand. membrane potential as an assay for cytotoxicity of anticancer
Having reached the above conclusions several years ago and drugs. Proceedings of the American Association of Cancer Re-
remaining committed to the use of multibeam instruments, search 22:68, 1981
240 SHAPIRO

18a. Bernal SD, Shapiro HM, Chen LB: Monitoring the effect of 41. Crissman HA, Van Egmond d, Holdrinet RS, Pennings A, Haa-
anticancer drugs on L1210 cells by a mitochondrial probe, rho- nen C: Simplified method for DNA and protein staining of
damine 123. Intl J Cancer 30: 219, 1982 human hematopoietic cell samples. Cytometry 2:59, 1981
19. Biihmer R-M, Ellwart J : Combination of BUdR-quenched 42. Curbelo R, Schildkraut ER, Hirschfeld T, Webb RH, Block MJ,
Hoechst fluorescence with DNA-specific ethidium bromide flu- Shapiro HM: A generalized machine for automated flow cytology
orescence for cell cycle analysis with a two-parametrical flow system design. J Histochem Cytochem 24:388, 1976
cytometer. Cell Tissue Kinet 14:653, 1981 43. Dangl J , Parks D, Oi V, Herzenberg L: Rapid isolation of cloned
20. Bohmer R-M, Ellwart J : Cell cycle analysis by combining the 5- isotype switch variants using fluorescence activated cell sorting.
bromodeoxyuridine/33258 Hoechst technique with DNA-spe- Cytometry 2395, 1982
cific ethidium bromide staining. Cytometry 231, 1981 44. Darzynkiewicz Z, Traganos F, Melamed MR: New cell cycle
21. Bohn B: High sensitivity cytofluorometric quantitation of lectin compartments identified by multiparameter flow cytometry. Cy-
and hormone binding to surfaces of living cells. Exp Cell Res tometry 1:98, 1980
103:39, 1976 45. Darzynkiewicz 2, Staiano-Coico L, Melamed MR. Increased
22. Bohn B: Flow cytometry: a novel approach for the quantitative mitochondrial uptake of rhodamine 123 during lymphocyte stim-
analysis of receptor-ligand interactions on surfaces of living cells. ulation. Proc Natl Acad Sci USA 78:2383, 1981
Mol Cell Endocrinol 20:1, 1980 46. Darzynkiewicz Z,Traganos F, Staiano-Coico L, Kapuscinski J,
23. Bohn B, Manske W: Application of flow cytofluorometry to Melamed MR: Interactions of rhodamine 123 with living cells
ligand binding studies on living cells: Practical aspects and studied by flow cytometry. Cancer Res 42:799, 1982
recommendations for calibration and data processing. Acta Pa- 47. Dean PN, Pinkel D: High resolution dual laser flow cytometry.
thol Microbiol Scand, Sect A, Suppl 274: 227, 1980 J Histochem Cytochem 26622, 1978
24. Bonner WA, Hulett HR, Sweet RG, Herzenberg L A Fluores- 48. Dittrich W, Gohde W: Impulsfluorometrie bei einzelzellen in
cence activated cell sorting. Rev Sci Inst 43:404, 1972 suspensionen. Z Naturforsch 24b:360, 1969
25. Braylan RC, Benson NA, Nourse V, Kruth HS: Correlated 49. Dolbeare FA, Smith RE: Flow cytoenzymology: Rapid enzyme
analysis of cellular DNA, membrane antigens, and light scatter analysis of single cells. In: Flow Cytometry and Sorting, Me-
of human lymphoid cells. Cytometry 2337, 1982 lamed MR, Mullaney PF, Mendelsohn ML (eds.) John Wiley &
26. Burns VV: Studies with a fluorescent vital probe for DNA in Sons, New York, 1979, p. 317-333
mammalian cells. Exp Cell Res 107:459, 1977 50. Eisert WG, Nezel M: Internal calibration to absolute values in
27. Carrano AV, Van Dilla, MA, Gray JW: Flow cytogenetics: a new flow through particle size analysis. Rev Sci Instrum 49:1617, 1978
approach to chromosome analysis. In: Flow Cytometry and 51. Evenson DP, Darzynkiewicz Z, Melamed MR: Simultaneous
Sorting, Melamed MR, Mullaney PF, Mendelsohn ML (eds.) measurement by flow cytometry of sperm cell viability and
John Wiley & Sons, New York, 1979, p. 421-451 mitochondrial membrane potential related to cell motility. J
28. Caswell AH, Hutchison JD: Selectivity of cation chelation to Histochem Cytochem 30:279, 1982
tetracyclines: evidence for special conformation of calcium che- 52. Freeman DA, Crissman HA: Evaluation of six fluorescent protein
late. Biochem Biophys Res Commun 43525, 1971 stains for use in flow microfluorometry. Stain Technol 50:279,
29. Cercek L, Cercek B: Changes in the structuredness of cytoplas- 1975
mic matrix (SCM) in human lymphocytes induced by phytohae- 53. Fried J , Perez A, Clarkson BD: Rapid hypotonic method for flow
magglutinin and cancer basic protein as measured on single cells. cytofluorometry of monolayer cell cultures. J Histochem Cyto-
Br J Cancer 33359, 1976 chem 26:921, 1978
30. Cercek L, Cercek B: Application of the phenomenon of changes 54. Fu CT, Johnston DA, Thompson D, Barlogie B: II)NA/Protein
in the structuredness of cytoplasmic matrix (SCM) in the diag- measurements by flow cytometry. Comparison of several tech-
nosis of malignant disorders: a review. Eur J Cancer 13:903, 1977 niques. Cell Tissue Kinet 13:683, 1980
31. Chan SS, Arndt-Jovin D J , Jovin TM: Proximity of lectin recep- 55. Fulwyler M: Electronic separation of biological cells by volume.
tors on the cell surface measured by fluorescence energy transfer Science 150:910, 1965
in a flow system. J Histochem Cytochem 27:56, 1979 55a. Gerson D F Determination of intracellular p H changes in lym-
32. Chance B, Thorell B: Localization and kinetics of reduced pyri- phocytes with 4-methylumbelliferone by flow microfluorometry.
dine nucleotide in living cells by microfluorometry. J Biol Chem In: Intracellular pH: Its Measurement, Regulation, and Utiliza-
234:3044, 1959 tion in Cellular Functions, Nuccitelli R, Deamer DW (eds.),Alan
33. Chance B, Cohen P, Jobsis F, Schoener B: Intracellular oxida- R. Liss, New York, 1982, p. 125-33
tion-reduction states in uzuo. Science 137:499, 1962 55b. Gerson DF, Kiefer H, Eufe W: Intracellular pH of mitogen-
34. Chandler DE;, Williams JA: Intracellular divalent cation release stimulated lymphocytes. Science 216:1009, 1982
in pancreatic acinar cells during stimulus-secretion coupling. I. 56. Gohde W, Schumann J , Biichner i', Otto F, Barlogie B: Pulse
Use of chlorotetracycline as fluorescent probe. J Cell Biol76:371, cytophotometry: application in tumor cell biology and in clinical
1978 oncology. In: Flow Cytometry and Sorting, Melamed MR, Mul-
35. Cohen RL, Muirhead KA, Gill J E , Waggoner AS, Horan PK: A laney PF, Mendelsohn ML (eds.) John Wiley and Sons, New
cyanine dye distinguishes between cycling and non-cycling fibro- York, 1979, p. 599-620
blasts. Nature 290:593, 1981 57. Goldstein JL, Brown MS, Kreiger M, Anderson RGW, M i n k B:
36. Cowden RR, Curtis SK: Demonstration of protein-bound sulfhy- Demonstration of low density lipoprotein receptors in mouse
dry1 and disulfide groups with fluorescent mercurials. Histo- teratocarcinoma stem cells and description of a method for
chemie 22247, 1970 producing receptor-deficient mutant mice. Proc Natl Acad Sci
37. Crissman HA, Steinkamp JA: Rapid, simultaneous measurement USA 76:2843, 1979
of DNA, protein and cell volume in single cells from large 58. Goldstein S, Korczack LB: Status of mitochondria in living
mammalian cell populations. d Cell Biol 59:766, 1973 human fibroblasts during growth and senescence in uitro: use of
38. Crissman HA, Tobey RA: Cell cycle analysis in 20 minutes. the laser dye rhodamine 123. J Cell Biol 91:392, 1981
Science 184:1297, 1974 59. Gratzner HG, Leif RC: An immunofluorescence method for
39. Crissman HA, Stevenson AP, Orlicky DJ, Kissane RJ: Detailed monitoring DNA synthesis by flow cytometry. Cytometry 1:385,
studies on the application of three fluorescent antibiotics for 1981
DNA st,aining in flow cytometry. Stain Technol 53:321, 1978 60. Gray J W , Dean PN, Mendelsohn M L Quantitative cell-cycle
40. Crissman HA, Stevenson AP, Kissane R J , Tobey RA: Tech- analysis. In: Flow Cytometry and Sorting, Melamed MR, Mul-
niques for quantitative staining of cellular DNA for flow cyto- laney PF, Mendelsohn M L (eds.) John Wiley & Sons, New York,
metric analysis. In: Flow Cytometry and Scoring, Melamed MR, 1979, p. 383-407
Mullaney PF, Mendelsohn ML (eds.) John Wiley and Sons, New 61. Hafeman DG, McConnell HM, Gray JW, Dean PN: Neutrophil
York, 1979, p. 243-261 activation monitored by flow cytometry: stimulation by phorbol
 MULTISTATION MULTIPARAMETER FLOW CYTOMETRY 24 1

diester is an all-or-none event. Science 215673, 1982 with Hoechst 33342. Cell Tissue Kinet 15:105, 1982
62. Hamori E, Arndt-Jovin DJ, Grimwade BG, Jovin TM: Selection 83. Kobayashi S, Kaede K, Nishiki K, Ogata E: Microfluorometry
of viable cells with known DNA content. Cytometry 1:132, 1980 of oxidation-reduction states of the rat kidney in ~itu.J Appl
62a. Hardy RR, Hayakawa K, Haaijman J, Herzenberg LA: B-cell Physiol 31:693, 1971
subpopulations identified by two-colour fluorescence analysis. 84. Koenig SH, Brown HD, Kamentsky L, Sedlis A, Melamed MR:
Nature 297:589, 1982 A report of the efficacy of a rapid cell spectrophotometer in
63. Hartmann W, Beisker W, Eisert R, Eisert WG: Fluorescence screening for cervical cancer. Cancer 21:1091, 1968
polarization measurements on human lymphocytes from patients 85. Kohen E,Kohen C, Thorell B: A comparative study of pyridine-
with and without neoplasia. Acta Pathol Microbiol Scand, Sect nucleotide metabolism in yeast and mammalian cells by micro-
A, Suppl 274: 183, 1980 fluorometry-microelectrophoresis. Histochemie 12:95, 1968
64. Heiple J M , Taylor DL: Intracellular pH in single motile cells. J 86. Krishan A, Ganapathi R Laser flow cytometric studies on the
Cell Biol 86:885, 1980 intracellular fluorescence of anthracyclines. Cancer Res 40:3895,
65. Hercher M. Mueller W, Shapiro HM: Detection and discrimi- 1980
nation of individual viruses by flow cytometry. J Histochem 87. Lalande ME, Miller RG: Fluorescence flow analysis of lympho-
Cytochem 27:350, 1979 cyte activation using Hoechst 33342 dye. J Histochem Cytochem
66. Hiebert RI), J e t t J H , Salzman GC: Modular electronics for flow 27:394, 1979
cytometry and sorting: the LACEL system. Cytometry 1:337, 88. Lalande ME, McCutcheon MJ, Miller RG: Quantitative studies
1981 on the precursors of cytotoxic lymphocytes. VI. Second signal
67. Hoffman PM, Davidson WF, Ruscetti SK, Chused TM, Morse requirements of specifically activated precursors isolated 12 h
HC 111: Wild mouse ecotropic murine leukemia virus infection of after stimulation. J Exp Med 151:12, 1980
inbred mice: Dual-tropic virus expression precedes the onset of 89. Latt SA: Fluorescent probes of DNA microstructure and synthe-
paralysis and lymphoma. J Virol39:597, 1981 sis. In: Flow Cytometry and Sorting, Melamed MR, Mullaney
68. Hoffman RA, Kung PC, Hansen P, Goldstein G: Simple and PF, Mendelsohn ML (eds.) John Wiley & Sons, New York, 1979,
rapid measurement of human T lymphocytes and their sub- p. 263-284
classes in peripheral blood. Proc Natl Acad Sci USA 77:4914, 90. Leary JF, Todd P, Wood JCS, Jett JH: Laser flow cytometric
1980 light scatter and fluorescence pulse width and pulse rise-time
68a. Hoffman RA, Hansen WP: Immunofluorescent analysis of sizing of mammalian cells. J Histochem Cytochem 27:315, 1979
blood cells by flow cytometry. Int J Immunopharmacol 3:249, 91. Ledbetter JA, Rouse RV, Micklem HS, Herzenberg LA: T cell
1981 subsets defined by expression of Lyt-1,2,3 and Thy-1 antigens.
69. Hulett HR, Bonner WA, Barrett J , Herzenberg LA: Cell sorting: Two-parameter immunofluorescence and cytotoxicity analysis
automated separation of mammalian cells as a function of intra- with monoclonal antibodies modifies current views. J Exp Med
cellular fluorescence. Science 166747, 1969 152:280, 1980
70. Hulett HR, Bonner WA, Sweet HG, Herzenberg LA: Develop- 92. Leeman U, Ruch F: Cytofluorometric determination of basic and
ment and application of a rapid cell sorter. Clin Chem 192313, total proteins with sulfaflavine. J Histochem Cytochem 20:659,
1973 1972
71. James TW, Bohman H: Proliferation of mitochondria during the 93. Levenson R, Macara IG, Smith RL, Cantley L, Housman D:
cell cycle of the human cell line (HL-60). J Cell Biol89:256, 1981 Role of mitochondrial membrane potential in the regulation of
72. Jensen KH, Langlois RG, Mayall BH: Strategies for choosing a murine erythroleukemia cell differentiation. Cell 28:855, 1982
deoxyribonucleic acid stain for flow cytometry of metaphase 94. Loew LM, Bonneville GW, Surow J : Charge shift optical probes
chromosomes. J Histochem Cytochem 25:953, 1977 of membrane potential. Theory. Biochemistry 17:4065, 1978
73. Johnson LV, Walsh ML, Bockus BJ, Chen LB: Monitoring of 95. Loken MR, Herzenberg LA: Analysis of cell populations with a
relative mitochondrial membrane potential in living cells by fluorescence-activated cell sorter. Ann NY Acad Sci 254:163,
fluorescence microscopy. J Cell Biol88:526, 1981 1975
74. Johnson TS,Swartzendruber DE, Martin JC: Nuclear size of 96. Loken MR, Sweet RG, Herzenberg LA: Cell discrimination by
G I / S transition cells measured by flow cytometry. Exp Cell Hes multiangle light scattering. J Histochem Cytochem 24:284, 1976
134:201, 1981 97. Loken MH, Parks DR, Herzenberg LA: Two-color immunofluo-
74a. Jongkind JF, Verkerk A, Visser WJ, Van Dongen JM: Isolation rescence using a fluorescence-activated cell sorter. J Histochem
of autofluorescent ‘aged’ human fibroblasts by flow sorting. Exp Cytochem 25:899, 1977
Cell Hes 138:409, 1982 98. Loken MR, Stout RD, Herzenberg LA: Lymphoid cell analysis
75. Kamentsky LA, Melamed MR, Derman H: Spectrophotometer: and sorting. In: Flow Cytometry and Sorting, Melamed MH,
new instrument for ultrarapid cell analysis. Science 150630,1965 Mullaney PF, Mendelsohn ML (eds.) John Wiley & Sons, New
76. Kamentsky LA, Melamed M R Instrumentation for automated York, 1979, p. 505-528
examination of cellular specimens. Proc Inst Electric Electron 99. Loken MR. Simultaneous quantitation of Hoechst 33342 and
Hng 57:2007, 1969 immunofluorescence on viable cells using a fluorescence acti-
77. Kamentsky LA: Cytology automation. Adv Biol Med Phys 14:93, vated cell sorter. Cytometry 1:136, 1980
1973 100. Loken MR, Houck DW: Light scattered a t two wavelengths can
78. Kanaoka Y, Machida M, Sekine T: Fluorescent thiol reagents. discriminate viable lymphoid cell populations on a fluorescence-
VI. N-(l-anilinonaphthyl-4)-maleimide; a fluorescent hydropho- activated cell sorter. J Histochem Cytochem 29609, 1981
bic probe directed to thiol groups in protein. Biochim Biophys 101. Mansberg HP, Saunders AM, Groner W: The Hemalog D white
Acta 317:563, 1973 cell differential system. J Histochem Cytochem 22:711, 1974
79. Kaplow LS, Lerner E: Computer-assisted monocyte esterase 102. Mariani BD, Slate DL, Schimke R T S phase-specific synthesis
assay by flow-cytophotometry. J Histochem Cytochem 25:590, of dihydrofolate reductase in Chinese hamster ovary cells. Proc
1977 Natl Acad Sci USA 78:4985, 1981
80. Kaplow LS: The application of cytochemistry to automation. J 103. McCutcheon MJ, Miller RG: Fluorescence intensity resolution
Histochern Cytochem 25:990, 1977 in flow systems. J Histochem Cytochem 27:246, 1979
81. Kerker M, Chew H, McNulty PJ, Kratohvil JP, Cooke DD, 104. Metezeau P, Djavadi-Ohaniance L, Goldberg ME: The kinetics
Sculley M, Lee M-P: Light scattering and fluorescence by small and homogeneity of endocytosis of a receptor-bound ligand in a
particles having internal structure. J Histochem Cytochem heterogeneous cell population studied by flow cytofluorometry.
27:250, 1979 J Histochem Cytochem 30:359, 1982
82. Kissane HJ, Tobey RA, Crissman HA, McLaughlin SR, Kraemer 105. Mikkelsen RB, Schmidt-Ullrich R. Concanavalin A induces the
PM: Detailed FCM and cell sorting studies of dye-binding ki- release of intracellular Ca” in intact rabbit thymocytes. J Biol
netics, viability and cell growth of cells following DNA staining Chem 255:5177, 1980
242 SHAPIRO

106. Mikkelsen RB, Schmidt-Ullrich R, Wallach DFH: Concanavalin receptors on rat basophilic leukemia cells. J Immunol 126:138,
A induces an intraluminal alkalinization of thymocyte membrane 1981
vesicles. J Cell Physiol 102113, 1980 131. Shapiro HM, Bier M, Zukoski CF: Continuous redox state mon-
107. Miller MH, Powell JI, Sharrow SO, Schultz AR: Rapid data itoring of preserved organs. Proc Annu Conf Eng Med Biol
collection, analysis and graphics for flow microfluorometry in- 15:274, 1973
strumentation. Rev Sci Instrum 49:1137, 1978 132. Shapiro HM, Schildkraut ER, Curbelo R, Laird CW, Turner
108. Murphy RF, Powers RS, Verderame M, Cantor CR, Pollack R RB, Hirschfeld T: Combined blood cell counting and classifica-
Flow cytofluorometric analysis of insulin binding and internali- tion with fluorochrome stains and flow instrumentation. J His-
zation by Swiss 3T3 cells. Cytometry 2402, 1982 tochem Cytochem 24:396,1976
109. Naccache PH, Volpi M, Showell H J , Becker EL, Sha-afi RI: 133. Shapiro HM, Schildkraut ER, Curbelo R, Turner RB, Webb
Chemotactic factor-induced release of membrane calcium in RH, Brown DC, Block MJ: Cytomat-R: a computer-controlled
rabbit neutrophils. Science 203:461, 1979 multiple laser source multiparameter flow cytophotometer sys-
110. Nicola NA, Metcalf D, von Melchner H, Burgess AW: Isolation tem. J Histochem Cytochem 25:836, 1977
of murine fetal hemopoietic progenitor cells and selective frac- 134. Shapiro HM: Fluorescent dyes for differential counts by flow
tionation of various erythroid precursors. Blood 58:376, 1981 cytometry: Does histochemistry tell us much more than cell
111. Nicolini C, Belmont A, Parodi S, Lessin S, Abraham S: Mass geometry? J Histochem Cytochem 25:976, 1977
action and acridine orange staining. Static and flow cytofluorom- 135. Shapiro HM, Natale PJ, Kamentsky LA: Estimation of mem-
etry. J Histochem Cytochem 27102, 1979 brane potentials of individual lymphocytes by flow cytometry.
112. Niedel J, Kahane I, Lachman L, Cuatrecasas P: A subpopulation Proc Natl Acad Sci USA 76:5728, 1979
of cultured human promyelocytic leukemia cells (HL-60) dis- 136. Shapiro HM: Flow cytometric probes of early events in cell
plays the formyl peptide chemotactic receptor. Proc Natl Acad activation. Cytometry 1:301, 1981
Sci USA 77:1000, 1980 137. Shapiro HM: Flow cytometric estimation of DNA and RNA
113. Noguchi PD, Johnson JB, Browne W: Measurement of DNA content in intact cells stained with Hoechst 33342 and pyronin
synthesis by flow cytometry. Cytometry 1:390, 1981 Y. Cytometry 2:143, 1981
114. Ohkuma S, Poole B: Fluorescence probe measurement of the 138. Shapiro HM: New methods in cytochemistry for flow cytometry.
intralysosomal pH in living cells and the perturbation of pH by In: Transplantation and Clinical Immunology, Vol XIII, Tour-
various agents. Proc Natl Acad Sci USA 75:3327, 1978 aine JL, Traeger J , Betuel H, Brochier J , Dubernard JM, Kevil-
114a. Oi VT, Glazer AN, Stryer L: Fluorescent phycobiliprotein con- lard J P , Triau R (eds.) Excerpta Medica, Amsterdam, 1981, p.
jugates for analyses of cells and molecules. J Cell Biol 93:981, 207-215
1982 139. Shapiro HM, Williams JM, Christenson L, Cotner T, Strominger
115. Ornstein L, Ansley HR: Spectral matching of classical cytochem- J , Strom TB: Rapid tissue typing and immunological monitoring
istry to automated cytology. J Histochem Cytochem 22:453,1974 using multiparameter flow cytometry to detect lymphocyte ac-
116. Osband ME, Cohen EB, McCaffrey RP, Shapiro HM: A tech- tivation. Abst Conf Cytom Clin Lab 57, 1982
nique for the flow cytometric analysis of lymphocytes bearing 140. Shapiro HM, Feinstein DM, Kirsch AS, Christenson L: Multi-
histamine receptors. Blood 56923, 1980 station multiparameter flow cytometry: Some influences of in-
117. Parks DR, Stove1 RT, Herzenberg LA: A tunable dye laser in a strumental factors on system performance.
two laser FACS system. Cytometry 2:119, 1981 (abstr) 141. Shapiro HM, Feinstein DM, Kirsch AS, Christenson L, Connor
118. Pollack A, Prudhomme D, Irvin GL 111, Claflin AJ, Block NL: JF Jr: Multistation multiparameter flow cytometry: Effective
Flow cytometric analysis of RNA content per cell using pyronin- minimal implementations.
Y and methyl green. Cytometry 2:122, 1981 (abstr) 142. Sharpless T, Traganos F, Darzynkiewicz Z, Melamed MH: Flow
119. Porro T J , Dadik SP, Green M, Morse HT: Fluorescence and cytofluorimetry: Discrimination between single cells and cell
absorption spectra of biological dyes. Stain Technol38:37, 1963 aggregates by direct size measurements. Acta Cytol 19:577, 1975
120. Porro T, Morse HT: Fluorescence and absorption spectra of 143. Sharpless T, Melamed MR: Estimation of cell size from pulse
biological dyes. Stain Technol 40:173, 1965 shape in flow cytofluorometry. J Histochem Cytochem 24:257,
121. Prewitt JMS: Parametric and nonparametric recognition by 1976
computer: an application to leukocyte image processing. Ad- 144. Sharpless T, Bartholdi M, Melamed MR: Size and refractive
vances in Computers 12:285, 1972 index dependence of simple forward angle scattering measure-
122. Roper PR, lkewinko B: Comparison of in uitro methods to ments in a flow system using sharply focused illumination. J
determine drug-induced cell lethality. Cancer Res 362182, 1976 Histochem Cytochem 25845, 1977
123. Rotman B, Papermaster BW: Membrane properties of living 145. Sharpless T: Cytometric data processing. In: Flow Cytometry
mammalian cells as studied by enzymatic hydrolysis of fluoro- and Sorting, Melamed MR, Mullaney PF, Mendelsohn ML (eds.)
genic esters. Proc Natl Acad Sci USA 55:134, 1966 John Wiley & Sons, New York, 1979, p. 359-379
124. Salzman GC, Crowell JM, Goad CA, Hansen KM, Hiebert RD, 146. Shechter Y, Schlessinger J, Jacobs S , Chang K-J, Cuatrecasas
LaBauve PM, Martin JC, Ingram ML, Mullaney PF: A flow- P: Fluorescent labeling of hormone receptors in viable cells:
system multiangle light-scattering instrument for cell character- Preparation and properties of highly fluorescent derivatives of
ization. Clin Chem 21:1297, 1975 epidermal growth factor and insulin. Proc Natl Acad Sci USA
125. Salzman GC, Crowell J M , Martin JC, Trujillo TT, Romero A, 75:2135, 1978
Mullaney PF, LaBauve PM: Cell identification by laser light 147. Shinitzky M, Inbar M: Microviscosity parameters and protein
scattering: identification and separation of unstained leukocytes. mobility in biological membranes. Biochim Biophys Acta
Acta Cytol 19374, 1975 433:133, 1976
126. Salzman GC, Hiebert RD, Crowell JM: Data acquisition and 148. Sims PJ, Waggoner AS, Wang C-H. Hoffman JF: Studies on the
display for a highspeed cell sorter. Comput Biomed Kes 11:77, mechanism by which cyanine dyes measure membrane potential
1978 in red blood cells and phosphatidylcholine vesicles. Biochemistry
127. Salzman GC, Mullaney PF, Price BJ: Light-scattering ap- 13:3315, 1974
proaches to cell characterization. In: Flow Cytometry and Sort- 149. Steen HB, Boye E, Skarstad K, Bloom B, Godal T, Mustafa S:
ing, Melamed MR, Mullaney PF, Mendelsohn ML (eds.) John Applications of flow cytometry on bacteria: Cell cycle kinetics,
Wiley & Sons, New York, 1979, p. 105-124 drug effects, and quantitation of antibody binding. Cytometry
128. Salzman GC, Wilder ME, Jett JH: Light scattering with stream- 2:249, 1982
in-air flow systems. J Histochem Cytochem 27:264, 1979 150. Steinkamp JA, Fulwyler MJ, Coulter JR, Hiebert RD, Horney
129. Scott RB, Grogan WM, Collins JM: Separation of rabbit marrow JL, Mullaney PF: A new multiparameter separator for micro-
precursor cells by combined isopycnic sedimentation and elec- scopic particles and biological cells. Rev Sci Instrum 44:1301,
tronic cell sorting. Blood 51:1137, 1978 1973
130. Segal lIM, Sharrow SO, Jones JF, Siraganian RP: F,(IgG) 151. Steinkamp JA, Kraemer PM: Quantitation of lectin binding by
 MULTISTATION MULTIPARAMETER FLOW CYTOMETRY 243
cells. In: Flow Cytometry and Sorting, Melamed MR, Mullaney changes in cytoplasmic free Ca2+ and membrane potential in
PF, Mendelsohn ML (eds.) John Wiley & Sons, New York, 1979, lymphocytes. Nature 29558, 1982
p. 497-504 173. Tycko B, Maxfield F R Rapid acidification of endocytic vesic!es
152. Steinkamp JA, Orlicky DA, Crissman HA: Dual-laser flow cy- containing an-macroglobulin. Cell 28643, 1982
tometry of single mammalian cells. J Histochem Cytochem 174. Udkoff R, Chan S, Norman A: Identification of mitogen respond-
27:273, 1979 ing lymphocytes by fluorescence polarization. Cytometry 1:265,
153. Steinkamp JA, Stewart CC, Crissman HA: Three-color fluores- 1981
cence measurements on single cells excited a t three laser wave- 175. Valet G, Bamberger S, Hofmann H, Schindler R, Ruhenstroth-
lengths. Cytornetry 2:226, 1982 Bauer G: Flow cytometry as a new method for the measurement
154. Steinkamp JA, Hiebert RD: Signal processing electronics for of electrophoretic mobility of erythrocytes using membrane
multiple electronic and optical measurements on cells. Cytome- charge staining by fluoresceinated polycations. J Histochem
try 2:232, 1982 Cytochem 27:342, 1979
155. Stewart S, Pritchard KI, Meakin JW, Price GB: A flow system 176. Valet G, Jenssen HL, Krefft M, Kuhenstroth-Bauer G: Flow-
adaptation of the SCM test for detection of lymphocyte response cytometric measurements of the transmembrane potential, the
in patients with recurrent breast cancer. Clin Immunol Immu- surface charge density and the phagocytic activity of guinea pig
nopathol 13:171, 1979 macrophages after incubation with lymphokines. Blut 42379,
156. Stohr M, Gebhardt U, Goerttler K: Computer assistance in 1981
multiparameter flow microphohmetry of mammalian cells. Bio- 177. Valet G, Haffael A, Moroder L, Wunsch E, Ruhenstroth-Bauer
technol Bioeng 181057, 1976 G: Fast, intracellular pH determinat,ion in single cells by flow-
157. Stohr M, Eipel H , Goerttler K, Vogt-Schaden M: Extended cytometry. Naturwissenschaften 68:265, 1981
application of flow microfluorometry by means of dual laser 178. van den Engh GJ, Visser JWM: T h e morphology of the hae-
excitation. Histochemistry 51305, 1977 mopoietic stem cell derived from scatter measurements on a
158. Stiihr M. Vogt-Schaden M, Knobloch M, Vogel R, Futterman light activated cell sorter. Annual Report, Radiobiological Lab-
G: Evaluation of eight fluorochrome combinations for simulta- oratories, Amsterdam, 1977, p 47
neous DNA-protein flow an . Stain Technol53:205, 1978 179. Van Dilla MA, Trujillo TT, Mullaney PF, Coulter JR: Cell
159. Stahr M, Futterman G: Vis tion of multidimensional spec- microfluorometry: a method for rapid fluorescence measure-
tra in flow cytometry. J Histochem Cytochem 27:560, 1979 ment. Science 1631213, 1969
160. Stiihr M, Vogt-Schaden M: A new dual staining technique for 180. Van Dilla MA, Langlois R, Pinkel I), Yajko D, Hadley WK:
simultaneous flow cytometric DNA analysis of living and dead Flow cytometric characterization of bacteria in culture and clin-
cells. Acta Pathol Microbiol Scand, Sect A, Suppl 274:96, 1980 ical specimens. Abstracts of the Conference on Cytometry in the
161. Sutherland R, Delia D, Schneider C, Newman R, Kemshead J , Clinical Laboratory 7:7, 1982
Greaves M: Ubiquitous cell-surface glycoprotein on tumor cells 180a. Van Wauwe J P , D e Mey J R , Goossens JG: OKT3: A monoclo-
is proliferation-associated receptor for transferrin. Proc Natl nal anti-human T lymphocyte antibody with potent mitogenic
Acad Sci USA 784515, 1981 properties. J Immunol 124:2708, 1980
161a. Swartzendruber DE: A bromodeoxyuridine (BUdRbmithra- 181. Visser JWM, Jongeling AAM, Tanke HJ: Intracellular pH-de-
mycin technique for detecting cycling and non-cycling cells by termination by fluorescence measurements. J Histochem Cyto-
flow microfluorometry. Exp Cell Res 109439, 1977 chem 27:32, 1979
162. Tanke H J , Niewenhuis IAB, Koper GJM, Slats JCM, Ploem JS: 182. Vyth-Dreese FA, Kipp JBA, D e Jong TAM: Simultaneous meas-
Flow cytometry of human reticulocytes based on RNA fluores- urement of surface immunoglobulins and cell cycle phase of
cence. Cytometry 1:313, 1981 human lymphocytes. Acta Pathol Microbiol Scand, Sect A,
163. Taylor IW: A rapid single step staining technique for DNA Suppl 274:207, 1980
analysis by flow microfluorimetry. J Histochem Cytochem 183. Wachtler F, Musil R: Nucleoli visualized by silver staining com-
28:1021, 1980 bined with a new RNA-specific fluorochrome. Stain Technol
164. Thomas JA, Buchsbaum RN, Zimniak A, Racker E: Intracellular 54:265, 1979
p H measurements in Ehrlich ascites tumor cells utilizing spec- 184. Wade CG, Rhyne R H Jr, Woodruff WH, Bloch DP, Bartholo-
troscopic probes generated in situ. Biochemistry 18:2210, 1979 mew JC: Spectra of cells in flow cytometry using a vidicon
165. Thorell B: Intracellular red-ox steady states as basis for cell detector. J Histochem Cytochem 27:1049, 1979
characterization by flow cytofluorometry. Blood Cells 6745,1980 185. Watson JV: Enzyme kinetic studies in cell populations using
166. Thorell B: Flow cytometric analysis of cellular endogenous flu- fluorogenic substrates and flow cytometric techniques. Cytome-
orescence simultaneously with emission from exogenous fluoro- try I:143, 1980
chromes, light scatter, and absorption. Cytometry 239, 1981 186. Watt SM, Burgess AW, Metcalf D, Battye FL: Isolation of mouse
166a. Titus JA, Haugland R, Sharrow SO, Segal DM: Texas red, a bone marrow neutrophils by light scatter and autofluorescence.
hydrophilic red-emitting fluorophore for use with fluorescein in J Histochem Cytochem 28:934, 1980
dual parameter flow microfluorometric and fluorescence micro- 187. Weil GJ, Chused TM: Eosinophil autofluorescence and its use in
scopic studies. J Immunol Meth 50:193, 1982 isolation and analysis of human eosinophils using flow micro-
167. Traganos F, Darzynkiewicz Z, Sharpless T , Melamed MR: Si- fluorometry. Blood 57:1099, 1981
multaneous staining of ribonucleic and deoxyribonucleic acids in 188. Wheeless LL Jr: Slit-scanning and pulse width analysis. In: Flow
unfixed cells using acridine orange in a flow cytofluorimetric Cytometry and Sorting, Melamed MR, Mullaney PF, Mendel-
system. J Histochem Cytochem 25:46, 1977 sohn ML (eds.) John Wiley & Sons, New York, 1979, p.125-135
168. Trask B J , van den Engh GJ, Elgershuizen JHBW: Analysis of 189. Williams J M , Shapiro HM, Milford EL, Strom TB: Multipa-
phytoplankton by flow cytometry. Cytometry 2:258, 1982 rameter flow cytometric analysis of lymphocyte subpopulation
169. Trujillo TT, Van Dilla MA: Adaptation of the fluorescent Feul- activation in lectin-stimulated cultures. J Irnmunol 128:2676,
gen reaction to cells in suspension for flow microfluorometry. 1982
Acta Cytol 16:26, 1972 190. Wofsy L, Henry C, Cammisuli S: Hapten-sandwich labeling of
170. Tsien RY: New calcium indicators and buffers with high sensi- cell-surface antigens. Contemp T o p Mol Immunol 7:215, I978
tivity against magnesium and protons: Design, synthesis and 191. Yataganas X, Mitomo Y, Traganos F, Strife A, Clarkson B:
properties of prototype structures. Biochemistry 19:2396, 1980 Evaluation of a Feulgen-type reaction in suspension using flow
171. Tsien RY: A non-disruptive technique for loading calcium microfluorimetry and a cell separation technique. Acta Cytol
buffers and indicators into cells. Nature 290:527, 1981 19:71. 1975
172. Tsien RH, Pozzan T, Rink T J : T-cell mitogens cause early