Measurement of Multiple Drug Resistance Transporter
Activity in Putative Cancer Stem/Progenitor Cells
Vera S. Donnenberg, E. Michael Meyer, and Albert D. Donnenberg
Multiple drug resistance, mediated by the expression and activity of ABC-transporters, is a major obstacle to
antineoplastic therapy. Normal tissue stem cells and their malignant counterparts share MDR transporter
activity as a major mechanism of self-protection. Although MDR activity is upregulated in response to
substrate chemotherapeutic agents, it is also constitutively expressed on both normal tissue stem cells and
a subset of tumor cells prior to the initiation of therapy, representing a built-in obstacle to therapeutic
ratio. Constitutive and induced MDR activity can be detected in cellular subsets of disaggregated tis-
sues, using the fluorescent substrates Rhodamine 123 and Hoechst 33342 for ABCB1 (also known as
P-gp and MDR1) and ABCG2 (BCRP1). In this chapter, we will describe the complete procedure for
the detection of MDR activity, including: (1) Preparing single-cell suspensions from tumor and normal
tissue specimens; (2) An efficient method to perform cell surface marker staining on large numbers of
cells; (3) Flow cytometer setup and controls; (4) Simultaneous measurement of Hoechst 33342 and
Rhodamine123 transport; and (5) Data acquisition and analysis.
Key words: Pre-existing multiple drug resistance, ABCB1 activity, ABCG2 activity, Hoechst
33342, Rhodamine 123, Epithelial tumors, Cancer stem cells, Flow cytometry
Multiple drug resistance (MDR) is mediated chiefly by a group of
transmembrane transporter glycoproteins known as ATP-binding
cassette (ABC) proteins. These energy-dependent cellular pumps
are highly conserved in eukaryotes. Because they preferentially
transport lipophilic cationic molecules, a single transporter can
act on structurally dissimilar substrates. Biologically, they have
been adapted to serve a wide variety of functions. In the central
1.1. MDR in Stem Cells
John S. Yu (ed.), Cancer Stem Cells, Methods in Molecular Biology, vol. 568
DOI 10.1007/978-1-59745-280-9_17, © Humana Press, a part of Springer Science + Business Media, LLC 2009
262 Donnenberg, Meyer, and Donnenberg
nervous system they are expressed by capillary endothelial cells,
forming the blood brain barrier ( 1 ) . In the gut they are expressed
by villus tip enterocytes of the small intestine ( 2 ) , limiting the
bioavailability of substrate drugs. In stem cells, MDR transport-
ers, principally ABCG2 (also known as breast cancer resistance
protein 1 (BCRP1)) and ABCB1 (also known as P-glycoprotein
(P-gp) and multiple drug resistance protein 1 (MDR1)), operate
at the single cell level, rather than at the tissue level, excluding
xenobiotics and other toxins and making stem cells among the
most resilient cells in the body. ABC transporters may be con-
stitutively expressed, as they are in some stem cell populations
( 3 ) , or may be induced by exposure to substrate, as they are in
memory T lymphocytes ( 4 ) .
MDR activity can be detected and quantified by exclusion
of fluorescent substrates. The association of constitutive MDR
expression with normal tissue stem cells (sometimes called adult
stem cells) was first demonstrated by Lansdorp, who showed
that the most primitive hematopoietic cells could be isolated
from the bone marrow (BM) by sorting for cells which excluded
the lipophilic, cationic dye rhodamine 123 ( 5 ) . Years later, this
phenomenon was rediscovered by Goodell while sorting BM
cells on the basis of DNA content (cell cycle). After staining
BM cells with the membrane permeant DNA stain Hoechst
33342 she noted, in addition to the usual cell cycle popula-
tions, a very small population of viable cells which excluded
the Hoechst dye. In very careful experiments, she determined
that this population was highly enriched for cells capable of
reconstituting hematopoiesis in radiation-ablated mice ( 3 ) . She
called these cells Side Population cells because of their location
on the far left side of blue fluorescence vs. red fluorescence his-
tograms. Today, Hoechst 33342 excluding cells are commonly
referred to as side population cells. Such cells can be found in a
wide variety of tissues, and the side population strategy has been
useful for isolating cells with high regenerative capacity from
hematopoietic ( 3, 5, 6 ) , airway ( 7 ) , pituitary ( 8 ) , small intestine
( 9 ) , and testicular ( 10 ) tissues. However, given the wide array
of biological functions that ABC transporters serve, and the fact
that they are sometimes induced in transporter-negative cells by
substrate exposure, the caveat must be given that MDR expres-
sion does not unequivocally identify stem cells and conversely,
its absence does not rule out the self-renewing capacity most
characteristic of stemness.
Multiple drug resistance transporters are so named because
they were first discovered in the context of antineoplastic ther-
apy ( 11 ) . Their discovery solved the conundrum posed by the
observation that cancer cells which developed resistance to a
1.2. MDR in Cancer
Measurement of Multiple Drug Resistance Transporter Activity 263
particular chemotherapeutic agent became simultaneously
resistant to a wide variety of unrelated agents, including drugs
with entirely different mechanisms of action. Today we know
that MDR is constitutively expressed by a subset (usually a small
subset) of neoplastic cells prior to treatment with substrate
drugs. Treatment results in selection for drug excluding MDR
active cells by a number of mechanisms, including regional gene
activation ( 12 ) , gene amplification ( 13 ) , and modification of
histone acetylation at the ABCG2 locus ( 14 ) . Recently it has
been suggested that MDR activity in some cancers is regulated
by the hedgehog signaling pathway ( 15, 16 ) , a key pathway in
embryonic morphogenesis ( 17 ) . Further, although the mecha-
nism remains unclear, there are data linking MDR expression to
radiation resistance ( 18 ) .
MDR activity has been investigated in multiple types of can-
cer as a possible means of identifying the cancer stem cell. The
vast majority of work has been done in cell lines, which have
undergone generations of selection for characteristics favorable
to in vitro growth in the absence of host- and therapy-mediated
selective pressures. Not all tumorigenic cancer cell lines exhibit a
side population. Investigators working with SP + cell lines derived
from ovarian cancer ( 19 ) , breast cancer ( 20 ) , glioma ( 21 ) , pros-
tate ( 22 ) , and thyroid cancer ( 23 ) all found enhanced tumori-
genicity or in vitro clonogenicity in sorted side population cells.
Harris et al. ( 21 ) and Mitsutake et al. ( 23 ) found that non-SP
cells could give rise to SP cells. In contrast, Lichtenauer et al.
found neither growth nor survival advantage in SP cells sorted
from an adrenocortical carcinoma cell line ( 24 ) . Our own data in
primary breast cancer isolates support plasticity in MDR expres-
sion. Sorted CD44+ CD90+ ABCG2− breast cancer cells gave rise
to heterogeneous tumors which included a subset of ABCG2+
cells when explanted to NOD/SCID mice ( 25 ) . Therefore, the
caveat given for normal stem cells, that MDR activity and stem-
ness are not one in the same, holds for neoplastic cells as well
( 26 ) . Although MDR activity is upregulated in response to sub-
strate chemotherapeutic agents, it is also constitutively expressed
on both normal tissue stem cells and a subset of tumor cells prior
to the initiation of therapy ( 19, 25– 27 ) , representing a built-in
obstacle to therapeutic ratio ( 28 ) . The take-home message is that
a cancer cell which is both self-renewing (i.e., tumorigenic) and
protected by MDR transporters constitutes a very difficult thera-
peutic target, having much in common with normal tissue stem
cells. Thus, detection and isolation of MDR active cells by simul-
taneous measurement of rhodamine 123 and Hoechst 33342
transport ( 29 ) represents an important tool for investigation of
those cancer cells capable of therapy resistance, dormancy, and
264 Donnenberg, Meyer, and Donnenberg
1. Tissue requirement 1 g minimum.
2. 60-mm Petri dishes Falcon (PL-056).
3. BD Bard-Parker protected disposal scalpel, blade#10 (Fisher
4. Collagenase/DNAase solution in PBS with calcium and
(a) Collagenase type I (Sigma C0130), 0.4% final concen-
(b) DNAse (Sigma D-5025-750KU), 350 kU/mL final
5. 50-mL Polypropylene conical tube, Falcon (Fisher 14-959-
6. DMEM 1×, high glucose, SuperCase (Invitrogen 11965-118).
7. Cell strainer, 70- m m nylon, 26-mm diameter (Becton Dick-
8. Cellector (TM) tissue sieve kit, 130 mL with pan, glass
pestle, and screens (Bellco 205020).
9. Newborn calf serum (Atlanta Biologicals S11210).
10. Phosphate-buffered saline without calcium or magnesium
(PBS-A, Sigma D5652).
11. PBS with CaCl 2 and MgCl 2 (PBS, Sigma D8662).
12. Ammonium chloride lysing solution 10× (500 mL).
(a) 41.5 g Ammonium chloride.
(b) 5.0 g Potassium bicarbonate.
(c) 400 mL Glass distilled water.
(d) pH to 7.2–7.4 with 1 M HCl or NaOH as needed.
(e) QS to 500 mL with distilled water.
(f) Dilute with distilled water to 1× prior to use.
13. Hemacytometer (VWR Scientific 48312-002).
14. Trypan Blue (Sigma T8154).
15. Histopaque (Sigma 10771).
16. 10-mL Pipettes, Falcon (Fisher 13-675-20).
17. 15-mL polypropylene conical, Falcon (Fisher 14-959-70C).
18. Mouse serum (Sigma M5905).
19. Microcentrifuge tube, 1.5 mL (Eppendorf Fisher 05-402-25).
20. L -Glutamine Gibco 25030-081.
21. 2-ME (Fisher BP176-100).
Measurement of Multiple Drug Resistance Transporter Activity 265
22. HEPES buffer (Sigma H-3375).
23. Rhodamine 123(Calbiochem 001004).
24. Hoechst 33342 (Molecular Probes H-3570).
25. Verapamil (Sigma V-4629).
26. Cyclosporin A (Sigma C1832).
27. Fumitremorgin C (Alexis Biochemicals 350-127-C250).
28. Vincristine (Sigma V8879).
29. Propidium Iodide (Sigma P-4170).
30. 70- m m Filter cap tubes, 12 × 75 (Fisher 08-771-23).
31. CompBead Antimouse Ig, k (BD 552843).
32. SpectrAlign beads (DAKO, Cat. No. KO111).
33. 8-Peak Rainbow Calibration Particles (Spherotech, Liberty-
ville, IL, Cat. No. RCP-30-5A).
34. Calibrite beads (PE) (BD 349502).
35. Calibrite (APC) (BD 340487).
1. Place tissue into a tared 60-mm Petri dish.
2. If desired remove a small piece for histology (paraffin and
3. Record the weight of remaining tissue.
4. Finely mince using sterile paired scalpels. Add a few drops of
collagenase/DNase solution while mincing to prevent tissue
from drying (Fig. 1 ) .
5. Place a 70- m m cell strainer in the mouth of a 50-mL polypro-
pylene conical tube labeled “Filtrate.”
6. Using a scalpel, a 10-mL pipette and PBS-A plus 2% calf
serum (PBS-A-CS), transfer all tissue into the cell strainer.
Repeat until all tissue fragments have been removed from
the Petri dish, pipetting PBS-A-CS vigorously against the
cell strainer to release individual cells ( see Notes 1 and 2 ).
7. Bring volume of Filtrate tube to 50 mL with PBS-A-CS,
centrifuge at 400 × g for 10 min, and discard supernatant.
Resuspend in 2 mL PBS-A-CS and hold on ice.
8. Using a scalpel, transfer minced tissue in the strainer to an
additional 50-mL polypropylene conical tube containing 10
mL collagenase/DNase solution. Label “Digest.”
3.1. Tissue Collection
266 Donnenberg, Meyer, and Donnenberg
9. Place Digest tube in shaking waterbath (e.g., Bellydancer,
Stovall Life Science) at 37°C for 30 min, maximum agitation
10. Place a new 70- m m cell strainer in the mouth of the conical
tube labeled Filtrate.
11. Using a 10-mL pipette, transfer material from the Digest
tube into the cell strainer on the Filtrate tube.
12. Add 10 mL PBS-A-CS to the Digest tube and transfer any
remaining tissue to the cell strainer. Repeat if necessary.
13. Using a scalpel, transfer the undigested material from the cell
strainer back into the Digest tube.
14. Add 10 mL collagenase/DNase solution and return Digest
tube to the shaking 37°C waterbath for 30 min.
15. Bring volume of Filtrate tube to 50 mL with PBS-A-CS,
centrifuge at 400 × g for 10 min, and discard supernatant.
Resuspend in 2 mL PBS-A-CS and hold on ice.
16. Repeat steps 10 and 11 .
Fig. 1. Preparation of normal breast tissue and breast tumor for digestion. Freshly excised normal breast tissue ( a , left )
and breast tumor ( a , right ) was transported in ice-cold tissue culture medium and transferred to 60-mm Petri dishes.
The tissue was minced into fine fragments using paired scalpels ( b ). A few drops of collagenase/DNase were added to
the minced tissue to prevent drying and clumping. PBS-A is added to the Petri dish ( c ) and carefully pipetted through a
cell strainer into a 50-mL conical tube ( d , Filtrate). Tissue remaining in the strainer is digested with collagenase/DNase
and the process is repeated as necessary .
Measurement of Multiple Drug Resistance Transporter Activity 267
17. Discard strainer from Filtrate tube.
18. Bring volume of Filtrate tube to 50 mL with PBS-A-CS, cen-
trifuge at 400 × g for 10 min, and discard supernatant.
19. Add 45 mL of 1× NH 4 Cl lysing solution and mix ( see Note 3 ).
20. Centrifuge at 400 × g for 10 min at room temperature.
21. Pour off the supernatant, and loosen cell pellet.
22. Resuspend in 40 mL PBS-A-CS.
23. Centrifuge at 400 × g for 10 min at room temperature.
24. Resuspend in 2 mL PBS with calcium and magnesium plus
2% calf serum and hold on ice ( see Note 4 ).
25. Count cells on a hemacytometer (Tuerk’s solution to elimi-
nate RBC, Trypan blue for viability) ( 30 ) .
1. Pellet cells at 400 × g for 10 min at 4°C. Discard superna-
2. Resuspend cell pellet in 5 m L neat decomplemented (56°C,
30 min) mouse serum ( see Note 5 ).
3. Pellet cells (400 × g , 10 min) and aspirate residual superna-
tant ( see Note 6 ).
4. Stain the dry pellet for surface markers by the addition of 2
m L of each monoclonal antibody.
5. Add 2μL antibodies in the following order ( see Note 7 ):
(a) EpCAM-APC (Miltenyi Biotech, Bergisch Gladbach,
Germany, Cat. No. 12000420)
(b) CD90-PE (Beckman Coulter, Cat. No. IM1840)
(c) CD45-APCCy7 (BD, Cat. No. 557833)
(d) ( see Note 8 ).
6. Incubate for 30 min on ice in the dark.
7. Dilute surface stained cell pellets in DMEM medium con-
taining 10% calf serum, L -glutamine (2 mM), and 2-beta
mercaptoethanol (50 m M) to a concentration of 2 × 10 6
cells/mL ( see Note 9 ).
1. Load surface stained cells for 90 min at 37°C with 0.13 m M
R123 and 8.12 m M Ho33342 in the absence and presence of
inhibitors ( Table 1 , see Notes 10 – 12 ).
2. When the incubation is complete, wash the cells once (400 ×
g , 10 min) with ice-cold PBS-A-CS and resuspend to 5–10 ×
10 6 cells/mL in ice cold PBS-A 10% CS ( see Note 13 ).
3. Add Propidium Iodide, 10 m g/mL final concentration, as a
viability dye ( see Note 14 ).
4. Strain cells immediately prior to flow cytometric acquisition
through a 70- m m cap filter.
Hoechst 33342 and
Rhodamine 123 Dye
Transport (Fig. 2 )
3.3.1. Dye Loading in the
Presence and Absence of
268 Donnenberg, Meyer, and Donnenberg
1. Vortex CompBeads thoroughly before use ( see Note 15 ).
2. Label a separate 1.5-mL Eppendorf tube for each mouse
monoclonal antibody conjugated to a tandem dye (e.g., ECD,
PE-Cy5, PE-Cy7, and APC-Cy7).
3. Add one full drop (approximately 60 m L) of antimouse Ig
CompBeads to each Eppendorf tube.
4. Centrifuge for 10 min at 400 × g . Carefully aspirate superna-
tant to ensure a “dry pellet” ( see Note 16 ).
5. Sonicate each tube for 10 s in a water bath sonicator ( see
Note 17 ).
3.3.2. Instrument Setup
and MDR Standards
IgG Capture Bead Staining
Fig. 2. Simultaneous detection of Ho33342 and R123 transport. Flow cytometry was performed on the parental cell line
K562 ( bottom panels , human erythroleukemia) and the MDR1 transfectant K562-G185 ( top panels ). A gating strategy
was used to analyze only singlet viable cells. Monoclonal antibody staining on separate aliquots revealed that parental
K562 cells were ABCB1 (MDR1) negative and expressed a very small subpopulation of ABCG2+, low side scatter cells
(0.08% of viable cells, top right ). In contrast, the transfectant line K562-G158, which is maintained in the presence of
100 ng/mL vincristine, is uniformly positive for ABCB1, and expresses a small subpopulation of ABCG2+ cells identical
to that of the parent cell line. Both cell lines were coincubated with Ho33342 and R123 in the absence and presence
of MDR inhibitors. In the absence of inhibitors, only a minor subpopulation within the parental K562 exhibited the SP
phenotype (3.36%), whereas almost all K562-G185 (ABCB1 transfected) cells transported Ho33342 (71% SP). Addition of
5 m M cyclosporin A (CsA) abrogated transport in both cell lines. Vincristine had no effect on either cell line, whereas the
ABCG2 specific inhibitor, fumitremorgin (1 m M), only inhibited Ho33342 transport in parental K562 (native ABCG2 medi-
ated transport) and had no effect on Ho33342 transport through ABCB1. Verapamil (50 m M, a MDR nonspecific inhibitor)
blocked SP phenotype in both cell lines. R123 transport was limited to cells with ABCB1 expression (K562-G185 only),
the parental K562 had no constitutive ABCB1 activity or expression. The R123 dull phenotype was reversed by the addi-
tion of either CsA or verapamil but was not blocked by the addition of ABCG2-specific inhibitor fumitremorgin.
FSC Peak Width
FSC Peak Height
FSC Peak Width
FSC Peak Height
0%INH 0%INH 66%INH
74%INH 0%INH 72%INH 80%INH
Measurement of Multiple Drug Resistance Transporter Activity 269
6. Add 2 m L of each antibody directly to beads (one antibody
per tube) and gently reflux.
7. Incubate for 15 min at room temperature in the dark ( see
Note 18 ).
8. Add 1 m L of mouse serum; incubate 5 min at room tempera-
ture ( see Note 19 ).
9. Add 100 m L of staining buffer and reflux.
10. Sonicate each tube for 10 s.
11. Add 1 mL of staining buffer. For manual compensation, add
one drop of negative CompBeads to each test tube that con-
tains antibody stained beads ( see Note 20 ).
12. Centrifuge beads for 10 min at 400 × g ; decant and carefully
blot to remove residual supernatant ( see Note 21 ).
13. Resuspend washed beads in 0.5 mL of staining buffer.
14. Transfer to 12 × 75 mm snap cap tubes for flow cytometry.
15. Sonicate for 10 s prior to acquisition on the flow cytometer.
1. We use an eight-color MoFlo cell sorter equipped with ultra-
violet, 488 and 633 nm lasers and an automated sample sta-
tion capable of holding the sample at 4°C (Beckman Coulter,
Fort Collins, CO). Other similarly equipped cytometers may
be used. The filters used for each fluorescence channel are
shown in Fig. 3 ( see Note 22 ).
2. The cytometer is calibrated to predetermined photomul-
tiplier target channels prior to each use using SpectrAlign
beads (DAKO, Cat. No. KO111) and 8-peak Rainbow
Calibration Particles (Spherotech, Libertyville, IL, Cat. No.
RCP-30-5A) ( see Note 23 ).
Instrument Setup and
Characteristics of common MDR inhibitors
Inhibitor Vehicle [Final concentration] Notes
Vehicle controls (EtOH
Not applicable 1/1,000
50 m M
5 m M
Potent dead end MDR
10 m M
Highly specific ABCG2
1 m g/mL
Used to maintain
ABCB1 expression in
270 Donnenberg, Meyer, and Donnenberg
3. All fluorescence parameters are collected in the logarithmic
mode, with the exception Hoechst 33342 emissions (blue and
red) which are collected in the linear mode ( see Note 24 ).
4. PMT settings for the two Hoechst channels must be fine
tuned empirically for each sample because data are acquired
in a linear mode and because fluorescence intensity varies
with the total DNA per sample. For initial settings it is a great
help to calibrate with cells having a known side population
( see Notes 10 and 22 ), adjusting PMT gain until the G1/
G0 peak is comfortably placed in the upper right third of the
Hoechst Red by Hoechst Blue histogram (Fig. 3 ) . For the
side population itself, blue fluorescence intensity should be
slightly greater than red fluorescence intensity, displacing the
side population slightly upward on the histogram ( 31 ) .
5. Acquire unstained cells or beads first, and then each single
stained sample (bead or cells) from the shortest emission
wavelength (R123 stained cells) to the longest (e.g., CD45
Fig. 3. Three-laser filter configuration to measure side population plus six additional fluorochromes. Routine analysis
and sorting of SP cells is performed on the MoFlo cell sorter by using the 488-nm laser line as a primary exciter of FITC,
PE, and PE tandem dyes. The 488 line is also used to provide forward (FSC) and side scatter (SSC) signals. Multiline UV
is used to excite Hoechst 33342 for SP detection. An additional 633-nm laser is used to excite APC and APC tandem
dyes. The MoFlo allows the operator to direct individual lines toward specific detector arrays on the Detection Table, as
illustrated in the figure ( upper left ). The convergence of the three lasers on the sample stream ( top view ) and the paths
of forward and side scattered light are shown in the lower right .
Measurement of Multiple Drug Resistance Transporter Activity 271
6. Calculate the compensation matrix from single stained tubes
using automated software (supplied with most new cytometer
software and all offline analysis packages).
7. Prior to running Ho33342/R123 loaded cells, minimal logical
gating is performed in order to eliminate cellular clumps
(doublets) ( 32 ) , subcellular debris (forward scatter thresh-
old), and PI+ dead and dying cells (Figs. 4 and 5 ) MDR
transfected reference cells ( see Note 12 ) are particularly useful
for this purpose.
8. Acquire Ho33342/R123 loaded cells, modifying “Hoechst
red 670/20” and “Hoechst blue 450/65” PMT voltage so
that nontransporting/non-SP cells with 2N DNA content
are visualized in the last third of the Hoechst red 670/20 vs.
Hoechst blue 450/65 histogram.
9. Acquire tumor/tissue cell suspensions loaded in the absence
and presence of inhibitors ( Table 1 , see Note 25 ).
We perform data analysis offline, using VenturiOne software
(Applied Cytometry) which has been designed to accommodate
very large datafiles. We create an analysis template into which
we load our spectrally compensated data. Our strategy usually
proceeds in three steps:
1. First we eliminate sources of interference with logical gates.
These include cell-cell doublets and clusters, subcellular
debris, and dead cells (Fig. 5 ) .
2. Next we decide on our classifier parameters. In this example CD45
expression (rt. 5 ) serves as a primary classifier, and Hoechst dye
exclusion (Fig. 6 ) is a secondary (branching) classifier.
3.4. Data Analysis
Fig. 4. Instrument set up for Hoechst 33342 detection using cell lines as process controls. Parental K562 and ABCB1-
transfected K562-G185 cells were used as low-negative and positive controls, respectively. A gating strategy was used
to restrict analysis to singlet viable (PI negative) cells. Approximately 650,000 viable singlet events were acquired for
both samples. The majority of parental K562 had no Ho33342 transport. The majority of events had uniform Ho33342
fluorescence in both blue and red detection PMTs, representing cells with 2N DNA. Almost all ABCB1 transfected cells
transported Ho433342 and therefore exhibited the side population (SP) phenotype, with greater blue than red fluores-
cence. Only a small minority of K562-G185 cells failed to transport Ho33342 and therefore had fluorescence intensities
characteristic of 2N DNA cells .
Ho BLUE (450/65)
SingletsViableK562 ParentalK562 G185
Ho RED (670/20)
272 Donnenberg, Meyer, and Donnenberg
3. Next outcomes , parameters to be measured on each population
defined by the classifiers, are determined. In this case (Fig. 6 )
the outcome variables are R123 dye efflux, CD90 and EpCAM
expression, and light scatter ( see Note 26 ).
4. Lymphocytes are often present in tumor specimens and are
identifiable by bright CD45 expression and low light scatter
(Fig. 5 ) . Since virtually none of these cells transports Hoechst
33342, the dye reaches the nucleus, where it binds to DNA
Fig. 5. Gating strategy to eliminate doublets, nonviable cells, and hematopoietic lineage positive events. Enzymatically
and mechanically digested nonsmall-cell lung carcinoma was stained for the expression of surface CD45 and EpCAM
(CD326); propidium iodide was added and cells were passed through a 70- m M filter immediately prior to acquisition.
Singlet cells were identified on a 2-parameter histogram of forward scatter (FSC) pulse height vs. FSC pulse width (A).
Viable cells which are impermeable to propidium iodide (PI) were identified (center histogram, B). Viable singlet cells
were analyzed for the expression of CD45 (common leukocyte antigen, D) and the epithelial surface antigen EpCAM.
CD45 bright/EpCAM-negative events were identified as lymphocytes (D) for use as a 2N DNA standard .
Fig. 6. Functional and phenotypic characterization of “side population” cells. Freshly isolated lung cancer cells (CD45−
viable singlets) were divided into Ho33342 excluding (F) and retaining (G) populations. Rhodamine 123 (R123) exclusion
and the surface expression of EpCAM, CD90 (Thy1) and light scatter properties (FSC, SSC) were evaluated. Side population
cells ( top panels ) cotransported R123 (15.4%). Cells which cotransported Ho33342 and R123 were mostly EpCAM nega-
tive/CD90− and had low light scatter comparable to resting lymphocytes. Nonside population cells, which did not transport
R123, were of mixed surface phenotype (EpCAM/CD90) and were predominantly (97.2%) large cells with high light scatter .
A and B
A and B not D
A and B and F not D
F and H
F and H
A and B and G not D
G and I
G and I
Measurement of Multiple Drug Resistance Transporter Activity 273
stoichiometrically. The great majority of lymphocytes have 2N
DNA content, but a distinct population of cycling cells can also
be seen (Fig. 7 ) . A population with very low blue and red fluo-
rescence is often present as a source of artifact in digested tissue
(Fig. 8 ) . These acellular events have sufficient light scatter to trig-
ger the cytometer to acquire a pulse, but have no DNA content
(which also prevents them from being eliminated by PI staining).
Another potential source of artifact is cells with laddered DNA
(Fig. 8 ) , the result of apoptosis-induced endonuclease release.
The feature which distinguished them from bona fide side popu-
lation cells is that cells with degraded DNA lie on the blue/red
diagonal, whereas side population cells have greater blue than red
fluorescence and are therefore slightly above the diagonal.
5. There is some confusion in the literature concerning the spe-
cificity of the side population, with early reports attributing
the side population exclusively to cells expressing the ABCG2
transporter. Figure 2 shows results obtained with K562 cells
(parent) and cells transfected with an ABCB1 vector (K562
G185). Surface staining with anti-ABCB1 (UIC2) and anti-
ABCG2 antibodies ( see Note 27 ), and real-time PCR per-
formed on these cells (not shown) confirmed that both cell
lines had low ABCG2 expression but that only the transfect-
ant had detectable ABCB1. The Instrument was set up using
parental and ABCB1 transfected cell lines and presence of
a battery of inhibitors. Two important points emerge from
the analysis: (1) Both ABCG2 (parental and transfected)
and ABCB1 (transfected only) mediate Hoechst 33342 trans-
port and therefore confer the side population phenotype; (2)
The transporters responsible for the side population and
Fig. 7. Identification of “side population” cells by Ho33342 transport in freshly digested tumor tissue. Events were
gated as shown in Fig. 5 . Lymphocytes identified as CD45bright/EpCAM negative (Fig. 5 ) were largely homogeneous for
Ho33342 uptake and DNA staining (2N, left histogram); no “side population” was detected. Nonhematopoietic (CD45−)
cells evidenced a significant “side population” accounting for 2.3% of events (center histogram). Cells with 2N and >2N
DNA content (non-Ho33342 transporting cells) comprised the majority of cells. Abrogation of the “side population” (right
histogram) by the addition of MDR inhibitors cyclosporine and fumitremorgin confirms that Ho33342 elimination is mediated
by MDR transporters.
A and B and C
A and B not D
A and B not D
No Inhibitors CsA+Fumi
274 Donnenberg, Meyer, and Donnenberg
R123 dull phenotypes may be inferred from inhibitor studies.
Fumitremorgin, an exquisitely specific inhibitor of ABCG2,
abrogated the side population phenotype in parental cells, but
not in ABCB1 transfectants.
1. If material appears slimy this indicates DNA release. Add
a few drops of DNase/collagenase solution and continue.
DNAse activity is measured in Kunitz units. DNase is
Fig. 8. Interpretation of Ho33342 exclusion histograms. Cells are permeant to the DNA-
binding dye Ho33342. In the absence of constitutively active MDR transport, Ho33342
concentrates in the nucleus and binds to the DNA, resolving 2N and 4N populations (top
histogram). DNA-bound Ho33342 autoquenches, shifting emission toward the red. Side
population cells, which do not permit accumulation of Ho33342 in the nucleus, exhibit
greater blue than red fluorescence, placing them above the diagonal (center histogram).
Events in the lower left-hand corner are not Ho33342 transporting cells, but rather
events that triggered on FSC but lack DNA. Another commonly seen artifact is illustrated
in the bottom histogram. The diagonal streak exhibits equal red and blue fluorescence
and represents cells with degraded DNA, as is seen in early apoptotic cells. In this
example, the diagonal streak was not abrogated by the addition of MDR inhibitors .
Measurement of Multiple Drug Resistance Transporter Activity 275
necessary because cells that are killed during the disaggregation
procedure release viscous DNA strands which entrap live cells.
2. Occasionally, tumors are too sclerotic for the minced pieces to
be disaggregated through a nylon cell strainer. In these cases
we substitute a stainless tissue sieve with a 10-mesh screen
overlayed with a 100-mesh screen. Tissue fragments are forced
through the screen with a glass pestle ( see Subheading 2 ).
3. Treatment with ammonium chloride lyses red blood cells.
This step may be omitted if sample is not visually contami-
nated with red blood cells.
4. DNase requires Ca 2+ or Mg 2+ for its activity.
5. Addition of mouse serum, which contains mouse immu-
noglobulins, blocks nonspecific binding of murine mono-
6. We refer to this as a dry the pellet elsewhere in the protocol.
Cells pelleted in a 15-mL conical tube and aspirated dry
actually contain anywhere from 10 to 50 m L of residual
7. The choice of antibodies and fluorochromes is specific to the
question being addressed and the available instrumentation,
and can be modified at will. The order of antibody addition
may influence staining as binding occurs very rapidly, and
sequential addition of many antibodies progressively reduces
the concentration of individual antibodies in the mixture
( 26, 33 ) .
8. We have found the following antibodies useful for subsetting
MDR+ tumor cells: CD44-PECy7 (Abcam, Cat. No. AB46793),
CD14-PECy5 (Beckman Coulter, Cat. No. IM2640U), CD33-
PECy5 (Beckman Coulter, Cat. No. IM2647U), glycophorin
A-PECy5 (Becton Dickinson, Cat. No. 559944). CD44 has been
used by many investigators to identify tumorigenic cancer cells
( 34 ) . Anti-CD14, anti-CD33, and antiglycophorin A (hema-
topoietic markers) are all labeled with the same fluorochrome
and collectively will serve as a dump gate during data analysis.
This means that all positive events within the dump gate will be
eliminated from the analysis.
9. Prewarm medium to 37°C. The medium formulation is not
important and others may be substituted. Medium must
contain serum, calcium, magnesium, and glucose and be
buffered to an appropriate pH. Antimicrobials, especially
and antifungal agents, should be avoided as some are MDR
10. Make all inhibitors at 1,000× final concentration to avoid
adding an excessive amount of the vehicle (EtOH for
276 Donnenberg, Meyer, and Donnenberg
11. Prepare tubes with inhibitors ahead of time and transfer cells
immediately after the addition of R123 and Ho33342.
12. It is useful, especially when first establishing this assay, to
use reference cell lines known to express MDR activity. We
use the chronic myelogenous leukemia cell line K562 (low
MDR activity) and K562 cells transfected with the ABCB1
vector G185 (very high MDR activity) ( 35 ) as process con-
trols (Fig. 4 ) , incubated in the absence and presence of
inhibitors (Fig. 2 ) . The G185 transfectants were kindly pro-
vided by Dr. Suresh Ambudkar.
13. Keep on ice under aluminum foil until acquisition; maximum
wait time before acquisition is 1 h. Holding cells longer will
result in loss of side population cells.
14. Viable cells exclude PI. PI stains dead and dying cells very
brightly and facilitates their removal from the analysis. Since
it is broadly fluorescent, PI-positive cells must be removed
from the analysis with a logical gate (Fig. 4 ) .
15. Single stained compensation standards are essential for
correct spectral compensation. We find it best to use periph-
eral blood mononuclear cells incubated with rhodamine 123
(per earlier protocol) for the FL1 standard, integrally stained
Calibrite beads for PE and APC, and Ig capture beads (Comp-
Beads) for antibodies conjugated to tandem dyes. Hoechst
33342 staining does not require spectral compensation. It
should be noted that R123, which is acquired with the same
filter as is used for FITC (530/40 BP), has more spillover
into the PE channel than FITC and therefore requires more
16. Beads do not pack as tightly as cells, so care must be taken
not to aspirate the beads.
17. Sonication is not essential but is very useful because it disag-
gregates bead clumps better than vortexing or refluxing.
18. Tandem dyes are easily degraded by exposure to light. Reagents,
stained cells, and stained beads must be carefully protected
from ambient light. We perform staining in an unilluminated
biological safety cabinet, and cover stained cells and beads
with aluminum foil to minimize light exposure.
19. Addition of mouse serum prevents clumping of beads due to
20. Negative beads are not required for some automated com-
pensation algorithms, such as those implemented on Beck-
man-Coulter instruments or in VenturiOne software.
21. Decanting and blotting must be done in one smooth motion
to prevent loss of beads.
Measurement of Multiple Drug Resistance Transporter Activity 277
22. The sample station is prechilled and kept at 4°C during the
23. Determining balanced PMT settings is an art in itself and
has been addressed by us elsewhere ( 32 ) . Once these settings
have been established, and bead target channels have been
determined, they can be reproduced from experiment to
experiment by adjusting PMT voltage to place the brightest
peak of the Rainbow particles in the predetermined tar-
get channel. The intensity of Hoechst 33342 staining will
vary from specimen to specimen, because it is measured on
a linear scale and is highly dependent on the total amount
of DNA in the sample. PMT voltages for the Hoechst Red
and Blue channels are adjusted for each sample to place the
median fluorescence of the 2N population at about channel
192 (of 255) (Fig. 4 ) .
24. Our MoFlo is an analog model in which log or linear
modes of amplification are user selected. Newer digital
flow cytometers use high-resolution linear amplification
exclusively and logarithmic transformations are performed
25. CD45− Lin− CD44+ side population cells may be very rare
events. We routinely acquire 10 million events per sample
at rates not exceeding 10,000 events/second, to assure suf-
ficient cells for subset analysis.
26. Which parameters are used to define classifiers, and which
define outcomes is somewhat plastic and depends on the
goal of the analysis. For example, one could classify cells
into several populations on the basis of CD90 and EpCAM
expression and then look at dye efflux as an outcome. The
classifier/outcome strategy has the advantages of focusing the
analysis and avoiding the all possible permutations problem
inherent in multiparameter data files.
27. These antibodies interfere with transporter activity and
cannot be used with dye transport assays.
The authors would like to thank Melanie Pfeifer and Amber
McCauslin for technical assistance. This work was supported
by grants BC032981 and BC044784 from the Department of
Defense, the Hillman Foundation, and the Glimmer of Hope Foun-
dation. Vera Donnenberg is a CDMRP Era of Hope Scholar.
278 Donnenberg, Meyer, and Donnenberg
1 . Schinkel AH, Mol CA, Wagenaar E, van
Deemter L, Smit JJ, Borst P. Multidrug resist-
ance and the role of P-glycoprotein knockout
mice. Eur J Cancer 1995;31A:1295–1298.
2 . Benet LZ, Izumi T, Zhang Y, Silverman JA,
Wacher VJ. Intestinal MDR transport proteins
and P-450 enzymes as barriers to oral drug
delivery. J Control Release 1999;62:25–31.
3 . Goodell MA, Brose K, Paradis G, Conner
AS, Mulligan RC. Isolation and functional
properties of murine hematopoietic stem
cells that are replicating in vivo. J Exp Med
4 . Donnenberg VS, Burckart GJ, Zeevi A, et al.
P-glycoprotein activity is decreased in CD4+
but not CD8+ lung allograft-infiltrating T
cells during acute cellular rejection. Trans-
5 . Udomsakdi C, Eaves CJ, Sutherland HJ,
Lansdorp PM. Separation of functionally
distinct subpopulations of primitive human
hematopoietic cells using rhodamine-123.
Exp Hematol 1991;19:338–342.
6 . Chaudhary PM, Roninson IB. Expression and
activity of P-glycoprotein, a multidrug efflux
pump, in human hematopoietic stem cells.
7 . Giangreco A, Shen H, Reynolds SD, Stripp
BR. Molecular phenotype of airway side
population cells. Am J Physiol Lung Cell Mol
8 . Chen J, Hersmus N, Van Duppen V, Caesens
P, Denef C, Vankelecom H. The adult pituitary
contains a cell population displaying stem/
progenitor cell and early embryonic character-
istics. Endocrinology 2005;146:3985–3998.
9 . He DN, Qin H, Liao L, et al. Small intestinal
organoid-derived SP cells contribute to repair
of irradiation-induced skin injury. Stem Cells
10 . Riou L, Bastos H, Lassalle B, et al. The telom-
erase activity of adult mouse testis resides in
the spermatogonial alpha6-integrin-positive
side population enriched in germinal stem
cells. Endocrinology 2005;146:3926–3932.
11 . Ling V, Thompson LH. Reduced permeability
in CHO cells as a mechanism of resistance to
colchicine. J Cell Physiol 1974;83:103–116.
12 . Wang YC, Juric D, Francisco B, et al. Regional
activation of chromosomal arm 7q with and
without gene amplification in taxane-selected
human ovarian cancer cell lines. Genes Chro-
mosomes Cancer 2006;45:365–374.
13 . Chen GK, Lacayo NJ, Duran GE, et al. Prefer-
ential expression of a mutant allele of the ampli-
fied MDR1 (ABCB1) gene in drug-resistant
variants of a human sarcoma. Genes Chromo-
somes Cancer 2002;34:372–383.
14 . Calcagno AM, Fostel JM, To KK, et al.
Single-step doxorubicin-selected cancer cells
overexpress the ABCG2 drug transporter
through epigenetic changes. Br J Cancer
15 . Sims-Mourtada J, Izzo JG, Ajani J, Chao KS.
Sonic Hedgehog promotes multiple drug
resistance by regulation of drug transport.
16 . Peacock CD, Wang Q, Gesell GS, et al.
Hedgehog signaling maintains a tumor stem
cell compartment in multiple myeloma. Proc
Natl Acad Sci U S A 2007;104:4048–4053.
17 . Lum L, Beachy PA. The Hedgehog response
network: sensors, switches, and routers. Science
18 . Johnstone RW, Ruefli AA, Tainton KM, Smyth
MJ. A role for P-glycoprotein in regulating
cell death. Leuk Lymphoma 2000;38:1–11.
19 . Szotek PP, Pieretti-Vanmarcke R, Masia-
kos PT, et al. Ovarian cancer side population
defines cells with stem cell-like characteris-
tics and Mullerian Inhibiting Substance
responsiveness. Proc Natl Acad Sci U S A
20 . Patrawala L, Calhoun T, Schneider-Broussard
R, Zhou J, Claypool K, Tang DG. Side popu-
lation is enriched in tumorigenic, stem-like
cancer cells, whereas ABCG2+ and ABCG2−
cancer cells are similarly tumorigenic. Cancer
21 . Harris MA, Yang H, Low BE, et al. Cancer
stem cells are enriched in the side population
cells in a mouse model of glioma. Cancer Res
22 . Patrawala L, Calhoun-Davis T, Schneider-
Broussard R, Tang DG. Hierarchical organi-
zation of prostate cancer cells in xenograft
tumors: the CD44+ alpha2beta1+ cell popu-
lation is enriched in tumor-initiating cells.
Cancer Res 2007;67:6796–6805.
23 . Mitsutake N, Iwao A, Nagai K, et al. Charac-
terization of side population in thyroid cancer
cell lines: cancer stem-like cells are enriched
partly but not exclusively. Endocrinology
24 . Lichtenauer UD, Shapiro I, Geiger K, et al.
Side population does not define stem cell-
like cancer cells in the adrenocortical carci-
noma cell line NCI h295R. Endocrinology
25 . Donnenberg VS, Luketich JD, Landreneau
RJ, DeLoia JA, Basse P, Donnenberg AD.
Tumorigenic epithelial stem cells and their
Measurement of Multiple Drug Resistance Transporter Activity 279 Download full-text
normal counterparts. Ernst Schering Found
Symp Proc 2006;5:245–263.
26 . Donnenberg VS, Landreneau RJ, Donnen-
berg AD. Tumorigenic stem and progenitor
cells: implications for the therapeutic index
of anti-cancer agents. J Control Release
27 . Lou H, Dean M. Targeted therapy for can-
cer stem cells: the patched pathway and ABC
transporters. Oncogene 2007;26:1357–1360.
28 . Donnenberg VS, Donnenberg AD. Thera-
peutic index and the cancer stem cell para-
digm. In: Bagley R, Teicher B, eds. Stem Cells
and Cancer Series: Cancer Drug Discovery and
Development . New York: Springer, Humana
29 . Bertoncello I, Williams B. Hematopoietic
stem cell characterization by Hoechst 33342
and rhodamine 123 staining. Methods Mol
30 . Turgeon ML. Clinical hematology: theory
and procedures (4th ed). Philadelphia, PA:
Lippincott Williams & Wilkins; 2005.
31 . Ibrahim SF, Diercks AH, Petersen TW, van
den Engh G. Kinetic analyses as a critical
parameter in defining the side population
(SP) phenotype. Exp Cell Res 2007;313:
32 . Donnenberg AD, Donnenberg VS. Under-
O’Gorman MR, Donnenberg AD, eds.
Handbook of Human Immunology (2nd ed).
Boca Raton: CRC Press Taylor and Francis;
33 . Donnenberg AD, Donnenberg VS. Rare-
event analysis in flow cytometry. Clin Lab
34 . Al-Hajj M, Wicha MS, Benito-Hernandez A,
Morrison SJ, Clarke MF. Prospective identifi-
cation of tumorigenic breast cancer cells. Proc
Natl Acad Sci U S A 2003;100:3983–3988.
35 . Kane S, Reinhard D, Fordis M, Pastan I,
Gottesman M. A new vector using the multi-
ple drug resistance gene as a selectable marker
enables overexpression of foreign genes in
eukaryotic cells. Gene 1989;84:439–446.
flow cytometry. In: