ARTICLE IN PRESS

Immunobiology 213 (2008) 629–640

www.elsevier.de/imbio

Comparative study of four fluorescent probes for evaluation of natural

killer cell cytotoxicity assays
Dana Cholujováa,1, Jana Jakubı́kováb,1, Miroslav Kubešc, Barbora Arendackád,
Michal Sapáke, Robert Ihnatkoc, Ján Sedlákb,
a
Laboratory of Molecular Oncology, Cancer Research Institute, Slovak Academy of Sciences, Vlárska 7, Bratislava, Slovakia
b
Laboratory of Tumor Immunology, Cancer Research Institute, Slovak Academy of Sciences, Vlárska 7, Bratislava, Slovakia
c
Institute of Virology, Slovak Academy of Sciences, Dubravska cesta 9, Bratislava, Slovakia
d
Institute of Measurement Science, Slovak Academy of Sciences, Dubravska cesta 9, Bratislava, Slovakia
e
Institute of Immunology, Medical Faculty of Comenius University, Sasinkova 4, Bratislava, Slovakia

Received 21 June 2007; received in revised form 20 January 2008; accepted 28 February 2008

Abstract
Cytotoxicity is one of the major defence mechanisms against both virus-infected and tumor cells. Radioactive
51
chromium (51Cr) release assay is a ‘‘gold standard’’ for assessment of natural killer (NK) cytolytic activity in vitro.
Several disadvantages of this assay led us to design alternative tools based on flow cytometry analysis. Four different
fluorescent dyes, calcein acetoxymethyl ester (CAM), carboxyfluorescein succinimidyl ester (CFSE), Vybrant DiO
(DiO) and MitoTracker Green (MTG) were tested for labeling of NK target K-562 cells. Target staining stability,
spontaneous release of fluorochromes and subsequent accumulation in bystander unstained cells were measured using
fluorimetry and flow cytometry. Healthy donor peripheral blood mononuclear cells and affinity column purified NK
cells were used as effectors coincubated with target K-562 cells at different E:T ratios for 3 h and 90 min, respectively.
Fluorescent probe 7-amino-actinomycin D was used for live and dead cell discrimination. Bland–Altman statistical
method was applied to measure true agreement for all CAM–51Cr, CFSE–51Cr, DiO–51Cr and MTG–51Cr pairs
analyzed.
Based on the data, none of the four proposed methods can be stated equivalent to the standard 51Cr release assay.
Considering linear relationships between data obtained with four fluorochromes and 51Cr release assay as well as linear
regression analysis with R2 ¼ 0.9393 value for CAM–51Cr pair, we found the CAM assay to be the most closely related
to the 51Cr assay.
r 2008 Elsevier GmbH. All rights reserved.

Keywords: Calcein-AM; CFSE; Cr release; Cytotoxicity; Mitotracker Green; Natural killer; Vybrant DiO

Abbreviations: 51Cr assay, 51Cr release assay; CAM, acetoxymethyl

ester of calcein; CFDA-SE, carboxyfluorescein diacetate succinimidyl
ester; CFSE, carboxyfluorescein succinimidyl ester; DiO, Vybrant
DiO; FC assay, flow cytometric assay; MTG, MitoTracker Green FM;
NK, natural killer.
Corresponding author.
E-mail address: exonsedl@savba.sk (J. Sedlák).
1
These two authors contributed equally to this work.
Introduction
Natural killer (NK) cells are the primary effector cells of
the innate immune system through their ability to
eliminate pathogen-infected and tumor cells (Whiteside
and Herberman 1994). They sense pathological changes in

0171-2985/$ - see front matter r 2008 Elsevier GmbH. All rights reserved.
doi:10.1016/j.imbio.2008.02.006
 ARTICLE IN PRESS
630 D. Cholujová et al. / Immunobiology 213 (2008) 629–640

tissues via balanced cognate activities of the NK inhibitory themselves. For this reason, we correlated the most
receptors and NK activating receptors, which recognize suitable fluorescent probes acetoxymethyl ester of calcein
decreased expression of major histocompatibility complex (CAM) (Neri et al. 2001; Roden et al. 1999), carboxy-
(MHC) class I molecules and increased expression of non- fluorescein succinimidyl ester (CFSE) (Jedema et al. 2004;
classical MHC molecules (MICA and MICB) on affected Westerhuis et al. 2005), Vybrant DiO (DiO) (Hoppner
cells, respectively. The receptor stimulation induces et al. 2002; Piriou et al. 2000), and MitoTracker Green
activation of NK cells and leads to direct elimination of FM (MTG) (Vizler et al. 2002) among other fluorescent
target cells through NK cell-mediated cytotoxicity (Lanier fluorochromes, which could be used for labeling target
2005). Therefore, monitoring of NK functional alteration cells, with the still widely used 51Cr release assay
emerges as important and may be useful in prognosis, or considered as the reference assay (Brunner et al. 1968).
follow-up during treatment of patients with a variety of The aim was to find a stable alternative flow cytometric
diagnoses (Mazodier et al. 2005; Santin et al. 2000; von method for monitoring NK activity that could be
Zons et al. 1997; Wiltschke et al. 1994). incorporated into the evaluation of immune parameters
Cytotoxicity follows a number of steps such as the in patients and possible correlate NK cell activity with
recognition of the foreign antigen or other molecules disease outcome and progression.
expressed on the surface of target cells, the creation
of conjugates between effector and target cells and the
Material and methods
activation of effector killer cells. NK cells mediate
non-major-histocompatibility-complex-restricted lysis and
antibody-dependent cytotoxicity, which can be mediated
Effector and target cells preparation
by two major classes of contact-dependent mechanisms
Effector PBMNCs were isolated from buffy coats of
such as killing by a secretory/necrotic cytotoxic mechanism,
healthy donors (National Transfusion Service, Bratislava)
associated with a perforin/granzyme-mediated pathway,
by Pancol density gradient centrifugation (1.077 g/ml,
or by an apoptotic mechanism based on receptor–ligand
PAN-Biotech, Germany). The mononuclear cells from
interactions (Lanier 2005).
the interface were collected, washed twice with PBS and
The standard method for determination of NK
once in RPMI, and then resuspended in 10 ml of complete
cytotoxic activity in vitro is the 51chromium (51Cr) release
culture medium (CM), consisting of RPMI 1640 medium,
assay (Brunner et al. 1968). Though this method has the
benefits of being reproducible and relatively easy to 2 mM L-glutamine, 100 IU/ml penicillin, 100 mg/ml strepto-
mycin and supplemented with 10% fetal calf serum (FCS).
perform, it has several limitations including short half-life
Purified NK cells were obtained from PBMNCs by positive
of chromium, poor loading and high spontaneous release
selection using CD56 mAb-conjugated magnetic beads
by some cell types and the measurement of cytolysis
according to the manufacturer’s instructions (Miltenyi,
at the population versus single-cell level. The most
Biotec). Target K-562 cells were maintained in RPMI
serious disadvantage is potential environmental and
1640 cell CM supplemented with 10% FCS and 2 mM
health hazard associated with the use of radioactive
glutamine. The cells were split every 3–4 days. For assay
isotope (Jakubek et al. 1983; Slezak and Horan 1989).
Besides assays analyzing the release of endogenous purposes, they were centrifuged once, washed with PBS
and the pellet resuspended in FCS free RPMI medium.
enzymes (e.g., lactate dehydrogenase, alkaline phospha-
To determine the absolute number of viable and dead
tase), other methods using non-radioactive compounds
cells equal volumes (100 ml) of both Flow-Count fluoro-
(e.g., dimethyl-thiazol-diphenyl bromide tetrazolium
spheres (Coulter) and cells were mixed. Cell concentr-
bromide, Alamar blue) measure directly the proportion
ations of the samples were calculated according to
of viable cells by evaluating variations in metabolic state
cells/ml ¼ (countscells  volumebeads  concentrationbeads)/
(Hussain et al. 1993; Korzeniewski and Callewaert 1983;
(countsbeads  volumecells) where ‘‘concentrationbeads’’
Nociari et al. 1998; Szekeres et al. 1981). In addition,
several groups have developed multicolor flow cytome- indicates the concentration of the beads suspension as
given by the manufacturer and ‘‘volume’’ indicates
try-based assays to study cell-mediated cytotoxicity
volumes pipetted per sample. Target and effector
(Fischer and Mackensen 2003; Kantakamalakul et al.
cells viability was determined by 7-amino-actinomycin
2003; Kasatori et al. 2005; Langhans et al. 2005;
D (7-AAD; Molecular Probes, USA) viability assay;
Lecoeur et al. 2001; Lee-MacAry et al. 2001; Sheehy
viability of 495% was required to proceed.
et al. 2001). However, the leakage of the dyes used for
cell discrimination causes labeling of nearby cells during
the assay and thus prevents proper discrimination of Labeling of target cells
target and effector cell populations. Thus, the choice of
a stable primary fluorochrome is highly important. Four different fluorescent dyes were used for labeling
Although many flow cytometry-based alternatives of K-562 cells before effector and target coincubation.
have been designed, they were not correlated among All fluorochromes were purchased from Molecular
 ARTICLE IN PRESS
D. Cholujová et al. / Immunobiology 213 (2008) 629–640
Probes (Eugene, USA). Cells were loaded with 0.1 mM
CAM (stock solution 1 mM in DMSO), 1 mM CFSE
(5 mM stock solution in DMSO), 2 ml of DiO (1 mM
stock solution in DMSO), or 200 nM MTG (stock
solution 1 mM in DMSO), respectively, and further
incubated for 15 min at 37 1C in the dark. The CFSE
staining was stopped by addition of an equal volume of
2% FCS RPMI medium.
Fluorescent dyes efflux
Analysis by fluorimetry and flow cytometry techni-
ques allowed us to determine the spontaneous release
of accumulated dyes into supernatant and related
decrease of target cells fluorescence intensities, as well
as subsequent accumulation of dyes in bystander
unstained PBMNCs. Unless otherwise stated, all in-
cubations were done in thermostat at 37 1C in humidi-
fied air atmosphere with 5% CO2. Fluorescent labeled
K562 cells (2  105 per well in 200 ml of CM) were seeded
in 96-well V-bottom microplates and CM was collected
after 1, 2, 3 and 4 h of incubation. The fresh CM was
added into each well and cells were further incubated to
reach the total incubation time of 4 h. The new-added
medium was also harvested (1+3 h, 2+2 h and 3+1 h)
and fluorescence of collected supernatants was measured
thereafter in black 96-well plates at 485/520 nm (excita-
tion/emission) using POLARstar fluorimeter (BMG
Labtech, Offenburg, Germany). For assessment of
bystander staining of effector cells, the fluorochrome
stained K562 cells (1  105 cells/100 ml) and unstained
PBMNCs (1  105 cells/100 ml) were mixed at the ratio
1:1 in a total volume of 200 ml CM in 96-well
V-bottomed microplates. Plates were centrifuged at
200  g for 3 min and incubated for 1, 2, 3 and 4 h. At
the end of each incubation period, the samples were
transferred to cytometric tubes (Sarstedt, Germany),
cells were resuspended in cold PBS and analyzed using a
Coulter Epics Altra flow cytometer. Appropriate con-
trols containing only stained or unstained cells were
carried out simultaneously.
Flow cytometric cytotoxicity assay
Three replicates of effector and fluorochrome-stained
target cells were coincubated at different effector:target
ratios. Target K-562 cells (2  104 per well) were seeded
in 96-well V-bottomed microplates in 50 ml of CM.
Appropriate six serial two-fold dilutions of PBMNCs
were made from which 100 ml aliquots were added to
K562 target cells to obtain the six effector:target (E:T)
ratios started at 50:1. Additionally, two sets of controls
containing only effector cells or only target cells were
prepared. Plates were centrifuged at 200  g for 3 min
and incubated for 3 h. Then the samples were trans-
631
ferred into cytometric tubes and 2 mg/ml of 7-AAD
(stock solution 1 mg/ml in PBS) was added to mark dead
cells. Samples were incubated on ice for 10 min,
protected from light and analyzed using a Coulter Epics
Altra flow cytometer. For assessment of purified NK cell
cytotoxicity, three 8:1, 4:1 and 2:1 effector:target ratios,
the incubation time of 90 min and target K562 cells
number of 104 cells per well were used.
Annexin V-PE/7-AAD staining
The cells were resuspended in 100 ml of 1  binding
buffer (10 mM Hepes/NaOH, 140 mM NaCl, 2.5 mM
CaCl2, pH 7.4), and mixed with 5 ml of Annexin V-PE
and 2 ml of 7-AAD (1 mg/ml). After 15 min incubation at
room temperature in the dark, the cells were analyzed
using a Coulter Epics Altra flow cytometer.
Flow cytometric data acquisition
Samples were collected using an Altra Epics four-
color flow cytometer equipped with an argon laser
operating at 488 nm. The emission of fluorochromes was
recorded through specific band-pass fluorescence filters:
C-AM, CFSE, Vibrant DiO or MTG; 525 nm (FL1),
Annexin V-PE; 575 nm (FL2), 7-AAD; 675 nm (FL4).
Data were analyzed using WinMDI version 2.7 software
(J. Trotter, Scripps Research Institute, La Jolla, CA).
The gating was done on side scatter (SSC; ordinate)
versus log-scale green fluorescence of one out of four
fluorescent dyes (abscissa) to separate target cells from
effector cells. To measure target-cell death, green
fluorescence-positive events were gated on, and analysis
of 7-AAD was performed. An average of 3000 target
cells was collected per sample.
Chromium (51Cr) release cytotoxic assay
The 51Cr release assays were performed as described
previously (Brunner et al. 1968). Briefly, K562 cells were
labeled with Na2(51Cr)O4 (Hartmann Analytic,
Braunschweig, Germany) for 2 h. After co-cultivation
with effector cells, the radioactivity was measured
(Clinigamma, LKB) per 100 ml aliquot. To determine
the spontaneous release of 51Cr from target cells
(cpmspontaneous), the aliquots from wells with only target
cells were used. To determine the maximal release of
51
Cr from target cells (cpmmaximal), 5% of Triton X-100
was added to the wells.
Calculations and statistical analysis
The cytotoxic activity toward the target cells was
determined on the basis of the enumeration of viable
cells using flow cytometric data. The percent specific
 ARTICLE IN PRESS
632 D. Cholujová et al. / Immunobiology 213 (2008) 629–640

lysis (PSL) was calculated at each E:T ratio as follows:

% specific lysis ¼ (CTTE/CT)  100 (where CT is
the percentage of viable fluorescent target cells in
control tubes and TE is the percentage of viable
fluorescent target cells in test(target+effector) tubes). Lytic
unit (LU) is defined as the number of effector cells
required to lyse 20% of predetermined standard number
(TSTD ¼ 2  104) of target cells. The calculation of
LU was performed by fitting of curve on semilog2
plot of logarithmically transformed E:T values versus
the specific lysis according to Bryant et al. (1992)
and Pross et al. (1981). For Bland–Altman statistical
method, the formula: 107/(E:T)20  TSTD for calculation
of LU number contained in 107 effector cells was
used.
For 51Cr release assay, cytotoxic activity was calcu-
lated as the percentage of specific 51Cr release using the
following equation: % specific lysis ¼ (cpmexperimental
cpmspontaneous)/(cpmmaximalcpmspontaneous)  100.

Results

A clear distinction between effector and target cells is

crucial in cytotoxicity assays. Therefore, we decided to
examine four fluorescent probes CAM, CFSE, DiO and
MTG for target cell labeling applicable in NK cytotoxic
activity assay to see whether any of these stainings are
equivalent to the standard 51Cr release assay. None of
the four fluorochromes was toxic at the concentration
used here and did not affect the viability of labeled cells,
as confirmed by flow cytometry 7-AAD viability assay
(data not shown).
Analysis of spontaneous dye leakage and effector cell
cross-staining
In order to study dye leakage, the spontaneous release
of fluorochromes into supernatants of CAM, CFSE,
DiO and MTG labeled K-562 was measured by
fluorimetric technique (Fig. 1). Supernatants of fluor-
escent labeled K-562 cells were collected at indicated
time points after seeding and fluorescent intensity of
CM was analyzed using POLARstar fluorimeter
(Fig. 1(A)). After each particular medium collection at
the times indicated, fresh CM was added into wells
immediately and cells were further incubated to reach
the total incubation time of 4 h. After 4 h, the newly
added medium was also harvested and the fluorescence
intensity was measured (Fig. 1(B)). The high fluorescent
signal recorded after 1 h incubation in supernatants of
CFSE and MTG-labeled K-562, with a significant
increase in time, indicated that in the case of these two
fluorochromes the highest spontaneous release into CM
occurred (Fig. 1(A)). In samples where CM was changed
after 1 h of incubation and fluorochrome-labeled K-562
Fig. 1. Fluorometric measurement of spontaneous release of
fluorescent probes into supernatants of CAM, CFSE, DiO and
MTG-labeled K-562 cells. (A) The fluorescent labeled K562
cells (2  105/200 ml CM per well) were incubated for 1, 2, 3
and 4 h when the medium was collected and fluorescence
intensity (mean fluorescence units) was measured at 485/
520 nm using POLARstar fluorimeter. (B) Fresh CM was
added into wells, cells were further incubated to reach the total
incubation time of 4 h and fluorescence intensity in super-
natants was measured. (C) The time-dependent proportion of
fluorescent probes release (last 1 (1), 2 (2) and 3 h (3) of
incubation in fresh CM) in comparison to 4 h (4) release (set
to 1.00). The data presented are representative of results
obtained from three independent experiments each in quad-
ruplicate; the results are means7S.E.
was further incubated in fresh CM for 3 h, the
fluorescence intensity of CFSE-labeled K-562 media
decreased to 16,400 a.u., nearly to the fluorescence levels
of CM of CAM or DiO labeled cells. Although the
fluorescent signal of MTG-labeled samples decreased
with CM exchange, the mean fluorescent units remained
relatively high, indicating that MTG leakage continued
even after 3 h incubation. When the fluorescence of
newly added media was compared to the level of
respective 4 h CM fluorescence (Fig. 1(C)), fluoro-
chromes can be arranged according the leakage rate in
the order: CFSE4MTG4CAMDiO. It is evident
that a substantial part of CFSE release into media
occurred within the first hour of incubation (the extent
of difference between 4 and 3 values). The lowest
 ARTICLE IN PRESS
D. Cholujová et al. / Immunobiology 213 (2008) 629–640 633

release of fluorochromes during the last hour of

incubation (1 value) was observed in CM of CFSE-
stained K-562 cells, followed by CAM, DiO and MTG-
labeled cells.
Targets and effectors were co-cultivated in order to
investigate whether the fluorescent probes which leaked
out from the target cells into CM can stain the effector
cells non-specifically. Fluorescence intensity of CAM,
CFSE, DiO and MTG-labeled K-562 in co-cultures with
PBMNCs and potential cross-staining of primary
unlabeled PBMNCs was analyzed by flow cytometry.
Cell debris was gated out and histogram overlays of
PBMNCs and labeled K-562 co-cultivated at 1:1 ratio
for 1 h (thin line, gray fill) or 4 h (thick line, no fill) with
mean fluorescence values are shown (Fig. 2). Time-
dependent decrease of cell fluorescence (26%) was
observed in CAM stained cells (Fig. 2(A)), followed by
48% decrease of the mean CFSE fluorescence intensity
(Fig. 2(B)), but in both cases the effector (M1 region)
and target (M2 region) cell populations were still clearly
distinguishable. In the case of DiO-labeled cells
(Fig. 2(C)), small but significant increase (1.5-fold) of
target cell fluorescence was observed after 4 h co-
cultivation in comparison to 1 h data. However, the
clear separation between stained and unstained cells was
still maintained regardless of relatively wide width of
fluorescence histograms of DiO-labeled K-562 cells. In
contrast, in MTG-labeled K-562 cells (Fig. 2(D)),
evident cross-staining of primarily unstained PBMNCs
appeared already 1 h after co-cultivation and became
more considerable after 4 h as confirmed by shift of
mean fluorescence values. Both populations fused and
the distinction between effector and target cells was
almost impossible. The best discrimination between
labeled target cells and unstained effector cells was
obtained with CAM followed by CFSE and DiO, while
MTG staining was unusable.

Flow cytometric assays for NK cytotoxicity

Cell-mediated cytotoxicity is a dynamic process. A

centrifugation step was carried out to enforce immediate
effector–target cell contact. Due to the high values of
cytotoxicity and the plateau observed in the preliminary
kinetic experiments (data not shown), the duration of
cytotoxicity assays was shortened from conventional 4
to 3 h for PBMNCs:K-562 co-cultures and to 90 min for
isolated NK:K-562 ones. The same gating procedure on
SSC/probe green fluorescence dot plots for all tested Fig. 2. Flow cytometric analysis of fluorescence intensity in
fluorescence dyes was used. The wider setting of gate at CAM, CFSE, DiO, MTG-labeled K-562 cells co-cultured with
PBMNCs. Histograms overlay of fluorescence intensity of
least 1 log of fluorescence intensity below the fluores-
PBMNCs and CAM, CFSE, DiO, MTG-labeled K-562 co-
cence of population in the control stained target cells cultured at 1:1 ratio for 1 h (thin line, gray fill) or 4 h (thick line,
(Fig. 3(A)) guarantees the incorporation of both lysed no fill) with the mean fluorescence values written in the rectangles
and viable target cells (Fig. 3(B), blue and green (1 h: gray; 4 h: transparent). The data presented are representative
population) in test(target+effector) tubes into the gate. of results obtained from the three independent experiments.
 ARTICLE IN PRESS
634 D. Cholujová et al. / Immunobiology 213 (2008) 629–640

Fig. 3. Gating strategy on SSC/probe fluorescence and 7-AAD/probe fluorescence dot plots. The CAM-labeled target cells only in
control(target) tube (A) and test(target+effector) tube (B). CAM (C), CFSE (D), DiO (E), MTG (F)-labeled K-562 target cells were
incubated with PBMNCs for 3 h at 12.5:1 ratio. Abscissa: respective probe fluorescence, ordinate: 7-AAD fluorescence. The results
are representative of nine independent experiments.

Subsequently, it was used for generation of 7-AAD/
green fluorescence dot plots for quantification of viable
and dead target cells (Fig. 3(C–F)). Similar percentages
of dead cells were obtained in CAM or CFSE-stained
K-562 cells (Fig. 3(C and D)), while lesser extent of
dead cells was observed in DiO or MTG-stained cells
(Fig. 3(E and F)) incubated with PBMNCs for 3 h at
12.5:1 ratio. This is due to non-specific increase of
PBMNCs fluorescence in DiO and MTG-stained target
cells as was shown in Fig. 2(C and D). CAM-labeled
and CFSE-labeled K-562 cells are clearly distinguishable
from unstained effector cells that allows the exact
quantification of the E:T ratio in each sample. Conse-
quently, we were able to check correctly adjusted cell
concentrations.
Analysis of apoptotic cell population in CAM assay
A subpopulation of CAM-labeled K-562 with de-
creased green fluorescence intensity, but without 7-AAD
positivity, was identified after coincubation with
PBMNCs when analyzing 7-AAD versus CAM dot
plots (Figs. 3(C) and 4(D). In order to ascertain if the
decrease of fluorescence intensity in this K-562 cell
subpopulation is related to the apoptotic process,
 ARTICLE IN PRESS
D. Cholujová et al. / Immunobiology 213 (2008) 629–640 635

Fig. 4. Analysis of apoptotic cell population in CAM assay. Control CAM-labeled K-562 cells (A and B) and co-cultured ones with
PBMNCs at E:T ratio 12.5:1 (C and D) were stained by Annexin V-PE/7-AAD. (A and C) Regions of viable (blue), apoptotic
(green) and necrotic (red) cells were set on Annexin V-PE/7-AAD dot plot, respectively. (B and D) Dot plots (7-AAD/fluorochrome)
used in cytotoxicity assays revealed the apoptotic cell population with decreased CAM staining (green). Oval depicts the viable cells
with CAM decreased/Annexin V-PE negative fluorescence. The results shown are representative of three independent experiments.

Annexin V-PE/7-AAD staining was performed at the
end of the coincubation time (Fig. 4). CAM-labeled
K-562 subpopulations (gated as in Fig. 3(B)) were
analyzed on 7-AAD versus Annexin V-PE dot plots.
Three different subpopulations were identified as
demonstrated in Fig. 4(A–C): Annexin V-negative/7-
AAD negative viable cells, Annexin V-positive/7-AAD
negative apoptotic cells and Annexin V-positive/7-AAD
positive late apoptotic/necrotic cells. The increase in
percentage of the apoptotic and necrotic cells was
accompanied by decrease in the number of viable cells
with increased E:T ratio. We found that there are viable
cells with decreased CAM fluorescence, but still Annexin
V-PE negative (Fig. 4(D), oval), thus not showing marks
of apoptosis yet.
Sensitivity and kinetics of CAM and CFSE based
FC assays
Because of clear distinction of target and effector cells
in CAM and CFSE assays, sensitivity and kinetics were
tested by 2 or 3 h co-cultivation of CAM or CFSE-
labeled K-562 with PBMNCs at ratio indicated
(Fig. 5(A)). CAM assay seems to be more sensitive than
CFSE assay as higher values of PSL were observed. To
obtain the relevant results, we used the same conditions
for flow cytometric assays and 51Cr release assay. In this
gold standard method, target cells are loaded with
Na51CrO, which passively enters cells and binds to
intracellular proteins. Upon target lysis, 51Cr is released
into the supernatant and the amount of 51Cr is
quantified using a beta or gamma counter. For
comparison of flow cytometric methods with 51Cr
release assay, the determination of PSL is important
for assessing the validity of each cytotoxic assay, since a
valid cytotoxicity assay will result in an increasing
dose–response relationship between PSL and increasing
E:T ratios. Fig. 5(B) shows representative results of
specific lysis obtained at three E:T ratios using
PBMNCs (Fig. 5(B): 1, 2, 3) and corresponding isolates
of NK cells (Fig. 5(B): 4, 5, 6) isolated from blood
samples of three donors. It is evident that at high
E:T ratio, there are higher PSL values in CAM and
 ARTICLE IN PRESS
636 D. Cholujová et al. / Immunobiology 213 (2008) 629–640

Fig. 5. (A) Time-dependent (2 and 3 h) lysis obtained by co-

cultivation of CAM or CFSE-labeled K-562 with PBMNCs.
The data presented are from four independent experiments;
and the results are means7S.E. (B) PBMNC (1, 2, 3) and
affinity column isolated NK (4, 5, 6) specific lysis of respective
fluorochrome labeled K-562 cells at ratio indicated. Repre-
sentative data of three blood donor samples (1+4, 2+5, 3+6)
are shown.

CFSE-stained K-562 cells in comparison to 51Cr release

assay. The use of isolated NK cells at adequate lower
E:T ratios resulted in higher convergence of PSL data
than in PBMNC assays. It is clear that shortened assay
time using isolated NK cells reduced the cross-staining
due to probe leakage as the lines of cytotoxicity for all
probes are closer together.
Statistical comparison of fluorescent methods with
51
Cr release assay
First, we examined whether any of the four fluor-
escent methods is equivalent to the standard 51Cr release
assay. The numbers of LU in 107 effector cells were
used for Bland–Altman analysis and representative plot
comparing CAM method with 51Cr method (Fig. 6(A)).
The test that the slope of the regression line of the
differences versus the averages is zero yields a p-value of
0.1168, however, the mean difference of 10.98 is not
ideally close to zero and the limits of agreement (12.13,
Fig. 6. (A) Bland–Altman plot for comparison of CAM and
51
Cr methods and the regression line in regression of
differences versus averages. (B) LU measured by CAM method
versus LU measured by 51Cr release assay with a regression
line and its Scheffé 95% confidence band. The estimated
regression line is CAM ¼ 0.396+1.192*Cr. (C) Linear
correlation between assessment of LU in both CAM and
51
Cr release assays.
34.09) are rather large (meaning that CAM method can
be 12.13 LU below or 34.09 LU over the 51Cr method).
Therefore, we cannot conclude that CAM method and
51
Cr method are equivalent. As we could not claim any
of the four tested assays equivalent to the standard, we
 ARTICLE IN PRESS
D. Cholujová et al. / Immunobiology 213 (2008) 629–640
tried to find a method which is the most closely related
to 51Cr assay in sense of linearity. To this purpose, we
examined the linear relationship between each of the
four methods and the standard. The quality of this
relationship is assessed by the width of Scheffé 95%
confidence band for the regression lines in regressions
of each of the four methods versus 51Cr assay. The
narrowest band is obtained for CAM method (the
regression line and the confidence band are plotted in
Fig. 6(B)). The estimated regression line is CAM ¼
0.396+1.192*51Cr. R2 value 0.9393 of linear fitting in
Excel data analysis indicates an acceptable correlation
between number of LU in 107 effector cells obtained in
CAM and 51Cr assays (Fig. 6(C)).
Discussion
In our study, we have used four fluorescent probes for
labeling of NK sensitive target K-562 cells in order to
assess their potential to replace the gold standard 51Cr
release assay. It is an essential need for multiparameter
flow cytometry that the ideal fluorochrome should be
stably integrated into the cell without altering the cell
characteristics and should stain the intended cell
population homogenously. However, the spontaneous
release of fluorescent dyes can be quite high giving rise
to misleading results and limiting their use in short-term
assays. In our assays, using four probes for target cell
staining, only significant MTG leakage was observed,
similar to coincubation studies that caused serious
difficulties making it impossible to evaluate cytotoxicity
assays properly (Goldberg et al. 1999; Johann et al.
1995). Mitochondrion-selective probe MTG was origin-
ally used for staining mitochondria via its thiol-reactive
chloromethyl moiety and to our knowledge MTG was
once previously used for measuring cytotoxicity by flow
cytometric assay (Vizler et al. 2002). In our hands, MTG
was the least usable probe for target cell discrimination
in co-cultivation. Bimodal fluorescence distribution
could be a consequence of either probe leakage or
possible mitochondrial membrane potential changes
(Keij et al. 2000), although the independence of MTG
fluorescence intensity from mitochondrial membrane
potential was published recently (Pendergrass et al.
2004).
CAM is a lipid-soluble fluorogenic esterase substrate
that passively crosses the cell membrane and is widely
used in functional assays of multidrug resistance
transporter activity (Dhar et al. 1998; Jakubikova
et al. 2002). Inside the cells, it is converted by
intracellular esterases into a polar lipid green fluorescent
product calcein that is retained by cells with intact
membranes. This property allows its use in fluorimetric
and flow cytometric assays for cell viability, including
cytotoxicity assays (Lichtenfels et al. 1994; Neri et al.
637
2001; Roden et al. 1999). In our study, 26% decrease
of the mean CAM fluorescence did not preclude a clear
separation of both populations, and because of low
spontaneous release CAM probe seems a suitable
candidate for discrimination of both cell populations.
Similarly, a non-fluorescent CFDA-SE is also con-
verted by intracellular esterases into a fluorescent
product CFSE, which forms covalent derivatives with
lysine residues of intracellular proteins and other
available amino groups within the cells stable for several
days. CFSE is a standard dye for cell division analysis
by flow cytometry and is also currently used for tracking
the migration of hemopoietic cells (Lyons 2000; Matera
et al. 2004; Parish 1999). According to previous studies,
most of the strong fluorescence signal detected just after
labeling was lost within the first hours because of efflux
of non-protein-bound dye and catabolism of CFSE-
bound proteins in the cells (Lyons et al. 2001; Nordon
et al. 1999). We confirmed that CFSE did not cause any
significant cross-contamination of effector cells (Johann
et al. 1995).
Numerous structurally related PKH dyes have
been developed that either resulted in labeling and
leakage problems, such as PKH-2, PKH26 (Fischer and
Mackensen 2003; Johann et al. 1995; Lee-MacAry et al.
2001) or by contrast were reported as stable and
long-durable cell membrane markers, e.g., DiOC18
(Hoppner et al. 2002; Mattis et al. 1997; Piriou et al.
2000). Vybrant DiO is a dye delivery solution that
can be added directly to culture media to uniformly
label suspended or attached culture cells but its cell
discrimination in co-cultivation was lower in compar-
ison to CAM or CFSE-labeling. It is known that
carbocyanine fluorochromes with shorter alkyl chains
are sensitive to membrane potential and that their
dimers formed at increased concentration exhibit a
lower fluorescence emission (Badley et al. 1973).
Quantitative analysis would be required to elucidate
this phenomenon more thoroughly, but that is beyond
the scope of this study.
Using the viable population for specific lysis determi-
nation helped us to account for all forms of cell death
that occurred and was appropriate for comparison of
targets staining with four different fluorescent probes.
Unlike the 51Cr release assay, flow cytometry-based
CAM/7-AAD assay has an advantage of detection of
cell death at the single cell level and also allows the
simultaneous quantification of apoptotic cell death by
Annexin V-PE. Although the level of 7-AAD and PI
uptake by dead target cells is almost identical (Derby
et al. 2001; Lecoeur et al. 2002), the use of 7-AAD is
more advantageous allowing free channels for immuno-
phenotyping. Observed decrease of CAM fluorescence
of Annexin V negative cells is neither due to early
apoptotic intracellular pH changes as fluorescence
of CAM is pH-insensitive nor probe leakage as cell
 ARTICLE IN PRESS
638 D. Cholujová et al. / Immunobiology 213 (2008) 629–640
membrane was PI impermeable (Lagadic-Gossmann
et al. 2004).
Analysis of different E:T ratios revealed the ability to
determine dose–response and also time–response effects.
Because labeled K-562 target cells are the minor
population in co-cultivation, the cross-staining of
bystander cells due to probe leakage is less effective
in comparison to the 1:1 ratio. In addition, the reduction
of time of assay caused the lowering of probe release.
Previous kinetic studies indicated that the flow cyto-
metric assay could identify lytic events before they
were detectable by the 51Cr release assay; the PSL
obtained from the 4 h 51Cr assay was comparable to
values obtained from a 2 h cytometric assay (Slezak and
Horan 1989). This method is still widely used and
considered as the reference assay (Brunner et al. 1968).
However, several disadvantages are associated with this
assay, for example the use of radioactive compounds,
poor loading of some target cell lines and high
spontaneous release by some cell types, but the major
limitation is that it only measures the final outcome
of the cytolytic process at the population level (Jakubek
et al. 1983).
The data consisting of numbers of LU per 107 effector
cells obtained from measurement of cytotoxicity in blood
samples by four proposed flow cytometric assays were
compared with data from conventional 51Cr release assay
in order to test reliability of flow cytometric assays and to
select the most suitable method to replace the standard.
The flow cytometric assay and 51Cr assay were conducted
at the same time with the same human effector cells and
the same target K-562 cells. Based on the data, none of the
four proposed methods can be stated equivalent to the
standard method 51Cr. But considering linear relationships
between number of LU in 107 effector cells obtained by the
four proposed methods and those obtained by 51Cr assay,
we can conclude that the best linear fit was achieved for
CAM method.
Conclusions
CAM assay can be adapted into a rapid, quantitative
and sensitive method to measure cellular cytotoxicity on
a single cell basis. We show that the use of this assay is
useful for calculation of LU numbers that correlate
linearly to values obtained from the 51Cr release assay.
With the advantage of multiparameter flow cytometry,
several different variables could be integrated into the
setting, which would enable study of the mechanisms of
death/apoptosis in in vitro cell-mediated cytotoxicity.
Moreover, any further phenotypic surface marker
staining on target cells can be carried out simulta-
neously. One advantage of flow cytometry is its ready
availability in many clinical immunology laboratories,
and this assay would be a suitable alternative for
measuring NK activity in the clinical laboratory.
Acknowledgments
The authors gratefully acknowledge expert technical
assistance by Mrs. Margita Sulikova and Mrs. Jana
Chovancova. This work was supported by a grant from
Daiwa Pharmaceutical Co. Ltd., Slovak Grant Agency
APVV 51-017505, VEGA Grants 2/5042 and 2/3437,
and European Social Fund (Project code 13120200038).
We are grateful to all the blood donors for donating
blood for this study.
References
Badley, R.A., Martin, W.G., Schneider, H., 1973. Dynamic
behavior of fluorescent probes in lipid bilayer model
membranes. Biochemistry 12, 268–275.
Brunner, K.T., Mauel, J., Cerottini, J.C., Chapuis, B., 1968.
Quantitative assay of the lytic action of immune lymphoid
cells on 51-Cr-labelled allogeneic target cells in vitro;
inhibition by isoantibody and by drugs. Immunology 14,
181–196.
Bryant, J., Day, R., Whiteside, T.L., Herberman, R.B., 1992.
Calculation of lytic units for the expression of cell-mediated
cytotoxicity. J. Immunol. Methods 146, 91–103.
Derby, E., Reddy, V., Kopp, W., Nelson, E., Baseler, M.,
Sayers, T., Malyguine, A., 2001. Three-color flow cyto-
metric assay for the study of the mechanisms of cell-
mediated cytotoxicity. Immunol. Lett. 78, 35–39.
Dhar, S., Nygren, P., Liminga, G., Sundstrom, C., de la Torre,
M., Nilsson, K., Larsson, R., 1998. Relationship between
cytotoxic drug response patterns and activity of drug efflux
transporters mediating multidrug resistance. Eur. J. Phar-
macol. 346, 315–322.
Fischer, K., Mackensen, A., 2003. The flow cytometric PKH-
26 assay for the determination of T-cell mediated cytotoxic
activity. Methods 31, 135–142.
Goldberg, J.E., Sherwood, S.W., Clayberger, C., 1999. A novel
method for measuring CTL and NK cell-mediated cyto-
toxicity using annexin V and two-color flow cytometry.
J. Immunol. Methods 224, 1–9.
Hoppner, M., Luhm, J., Schlenke, P., Koritke, P., Frohn, C.,
2002. A flow-cytometry based cytotoxicity assay using
stained effector cells in combination with native target cells.
J. Immunol. Methods 267, 157–163.
Hussain, R.F., Nouri, A.M., Oliver, R.T., 1993. A new
approach for measurement of cytotoxicity using colori-
metric assay. J. Immunol. Methods 160, 89–96.
Jakubek, P., Thorsby, E., Hirschberg, H., 1983. Cell mediated
lympholysis; a modified technique using 111indium-oxine-
labelled targets. J. Immunol. Methods 60, 379–390.
Jakubikova, J., Duraj, J., Hunakova, L., Chorvath, B., Sedlak,
J., 2002. PK11195, an isoquinoline carboxamide ligand of
the mitochondrial benzodiazepine receptor, increased drug
uptake and facilitated drug-induced apoptosis in human
 ARTICLE IN PRESS
D. Cholujová et al. / Immunobiology 213 (2008) 629–640
multidrug-resistant leukemia cells in vitro. Neoplasma 49,
231–236.
Jedema, I., van der Werff, N.M., Barge, R.M., Willemze, R.,
Falkenburg, J.H., 2004. New CFSE-based assay to
determine susceptibility to lysis by cytotoxic T cells of
leukemic precursor cells within a heterogeneous target cell
population. Blood 103, 2677–2682.
Johann, S., Blumel, G., Lipp, M., Forster, R., 1995. A versatile
flow cytometry-based assay for the determination of
short- and long-term natural killer cell activity. J. Immunol.
Methods 185, 209–216.
Kantakamalakul, W., Jaroenpool, J., Pattanapanyasat, K.,
2003. A novel enhanced green fluorescent protein (EGFP)-
K562 flow cytometric method for measuring natural killer
(NK) cell cytotoxic activity. J. Immunol. Methods 272,
189–197.
Kasatori, N., Ishikawa, F., Ueyama, M., Urayama, T., 2005.
A differential assay of NK-cell-mediated cytotoxicity in
K562 cells revealing three sequential membrane impairment
steps using three-color flow-cytometry. J. Immunol. Meth-
ods 307, 41–53.
Keij, J.F., Bell-Prince, C., Steinkamp, J.A., 2000. Staining of
mitochondrial membranes with 10-nonyl acridine orange,
MitoFluor Green, and MitoTracker Green is affected by
mitochondrial membrane potential altering drugs. Cyto-
metry 39, 203–210.
Korzeniewski, C., Callewaert, D.M., 1983. An enzyme-release
assay for natural cytotoxicity. J. Immunol. Methods 64,
313–320.
Lagadic-Gossmann, D., Huc, L., Lecureur, V., 2004. Altera-
tions of intracellular pH homeostasis in apoptosis: origins
and roles. Cell Death Differ. 11, 953–961.
Langhans, B., Ahrendt, M., Nattermann, J., Sauerbruch, T.,
Spengler, U., 2005. Comparative study of NK cell-mediated
cytotoxicity using radioactive and flow cytometric cyto-
toxicity assays. J. Immunol. Methods 306, 161–168.
Lanier, L.L., 2005. NK cell recognition. Annu. Rev. Immunol.
23, 225–274.
Lecoeur, H., Fevrier, M., Garcia, S., Riviere, Y., Gougeon,
M.L., 2001. A novel flow cytometric assay for quantitation
and multiparametric characterization of cell-mediated
cytotoxicity. J. Immunol. Methods 253, 177–187.
Lecoeur, H., Oliveira-Pinto, L.M., Gougeon, M.L., 2002.
Multiparametric flow cytometric analysis of biochemical
and functional events associated with apoptosis and oncosis
using the 7-aminoactinomycin D assay. J. Immunol.
Methods 265, 81–96.
Lee-MacAry, A.E., Ross, E.L., Davies, D., Laylor, R.,
Honeychurch, J., Glennie, M.J., Snary, D., Wilkinson,
R.W., 2001. Development of a novel flow cytometric cell-
mediated cytotoxicity assay using the fluorophores PKH-26
and TO-PRO-3 iodide. J. Immunol. Methods 252, 83–92.
Lichtenfels, R., Biddison, W.E., Schulz, H., Vogt, A.B.,
Martin, R., 1994. CARE-LASS (calcein-release-assay), an
improved fluorescence-based test system to measure cyto-
toxic T lymphocyte activity. J. Immunol. Methods 172,
227–239.
Lyons, A.B., 2000. Analysing cell division in vivo and in vitro
using flow cytometric measurement of CFSE dye dilution.
J. Immunol. Methods 243, 147–154.
639
Lyons, A.B., Hasbold, J., Hodgkin, P.D., 2001. Flow cytometric
analysis of cell division history using dilution of carboxy-
fluorescein diacetate succinimidyl ester, a stably integrated
fluorescent probe. Methods Cell Biol. 63, 375–398.
Matera, G., Lupi, M., Ubezio, P., 2004. Heterogeneous cell
response to topotecan in a CFSE-based proliferation test.
Cytometry A 62, 118–128.
Mattis, A.E., Bernhardt, G., Lipp, M., Forster, R., 1997.
Analyzing cytotoxic T lymphocyte activity: a simple and
reliable flow cytometry-based assay. J. Immunol. Methods
204, 135–142.
Mazodier, K., Marin, V., Novick, D., Farnarier, C., Robitail,
S., Schleinitz, N., Veit, V., Paul, P., Rubinstein, M.,
Dinarello, C.A., Harle, J.R., Kaplanski, G., 2005. Severe
imbalance of IL-18/IL-18BP in patients with secondary
hemophagocytic syndrome. Blood 106, 3483–3489.
Neri, S., Mariani, E., Meneghetti, A., Cattini, L., Facchini, A.,
2001. Calcein-acetyoxymethyl cytotoxicity assay: standar-
dization of a method allowing additional analyses on
recovered effector cells and supernatants. Clin. Diagn. Lab.
Immunol. 8, 1131–1135.
Nociari, M.M., Shalev, A., Benias, P., Russo, C., 1998. A
novel one-step, highly sensitive fluorometric assay to
evaluate cell-mediated cytotoxicity. J. Immunol. Methods
213, 157–167.
Nordon, R.E., Nakamura, M., Ramirez, C., Odell, R., 1999.
Analysis of growth kinetics by division tracking. Immunol.
Cell Biol. 77, 523–529.
Parish, C.R., 1999. Fluorescent dyes for lymphocyte migration
and proliferation studies. Immunol. Cell Biol. 77, 499–508.
Pendergrass, W., Wolf, N., Poot, M., 2004. Efficacy of
MitoTracker Green and CMXrosamine to measure changes
in mitochondrial membrane potentials in living cells and
tissues. Cytometry A 61, 162–169.
Piriou, L., Chilmonczyk, S., Genetet, N., Albina, E., 2000.
Design of a flow cytometric assay for the determination of
natural killer and cytotoxic T-lymphocyte activity in
human and in different animal species. Cytometry 41,
289–297.
Pross, H.F., Baines, M.G., Rubin, P., Shragge, P., Patterson,
M.S., 1981. Spontaneous human lymphocyte-mediated
cytotoxicity against tumor target cells. IX. The quantitation
of natural killer cell activity. J. Clin. Immunol. 1, 51–63.
Roden, M.M., Lee, K.H., Panelli, M.C., Marincola, F.M.,
1999. A novel cytolysis assay using fluorescent labeling and
quantitative fluorescent scanning technology. J. Immunol.
Methods 226, 29–41.
Santin, A.D., Hermonat, P.L., Ravaggi, A., Bellone, S.,
Roman, J., Pecorelli, S., Cannon, M., Parham, G.P.,
2000. Effects of concurrent cisplatinum administration
during radiotherapy vs. radiotherapy alone on the immune
function of patients with cancer of the uterine cervix. Int. J.
Radiat. Oncol. Biol. Phys. 48, 997–1006.
Sheehy, M.E., McDermott, A.B., Furlan, S.N., Klenerman, P.,
Nixon, D.F., 2001. A novel technique for the fluorometric
assessment of T lymphocyte antigen specific lysis.
J. Immunol. Methods 249, 99–110.
Slezak, S.E., Horan, P.K., 1989. Cell-mediated cytotoxicity. A
highly sensitive and informative flow cytometric assay.
J. Immunol. Methods 117, 205–214.
 ARTICLE IN PRESS
640 D. Cholujová et al. / Immunobiology 213 (2008) 629–640
Szekeres, J., Pacsa, A.s., Pejtsik, B., 1981. Measurement of
lymphocyte cytotoxicity by assessing endogenous alkaline
phosphatase activity of the target cells. J. Immunol.
Methods 40, 151–154.
Vizler, C., Nagy, T., Kusz, E., Glavinas, H., Duda, E.,
2002. Flow cytometric cytotoxicity assay for measuring
mammalian and avian NK cell activity. Cytometry 47,
158–162.
von Zons, P., Crowley-Nowick, P., Friberg, D., Bell, M.,
Koldovsky, U., Whiteside, T.L., 1997. Comparison of
europium and chromium release assays: cytotoxicity in
healthy individuals and patients with cervical carcinoma.
Clin. Diagn. Lab. Immunol. 4, 202–207.
Westerhuis, G., Maas, W.G., Willemze, R., Toes, R.E., Fibbe,
W.E., 2005. Long-term mixed chimerism after immunologic
conditioning and MHC-mismatched stem–cell transplanta-
tion is dependent on NK-cell tolerance. Blood 106,
2215–2220.
Whiteside, T.L., Herberman, R.B., 1994. Role of human
natural killer cells in health and disease. Clin. Diagn. Lab.
Immunol. 1, 125–133.
Wiltschke, C., Tyl, E., Speiser, P., Steininger, A., Zeillinger, R.,
Kury, F., Czerwenka, K., Kubista, E., Preis, P., Krainer, M.,
1994. Increased natural killer cell activity correlates with low
or negative expression of the HER-2/neu oncogene in
patients with breast cancer. Cancer 73, 135–139.