Download Divided we stand: Tracking cell proliferation with carboxyfluorescein

Immunology and Cell Biology (1999) 77, 509–515
Special Feature
Divided we stand: Tracking cell proliferation with
carboxyfluorescein diacetate succinimidyl ester
A B RU C E LYO N S
Discipline of Pathology, The University of Tasmania, Hobart, Tasmania, Australia
Summary Most techniques for assessing cell division can either detect limited numbers of cell divisions (bromodeoxyuridine incorporation) or only quantify overall proliferation (tritiated thymidine incorporation). In the
majority of cases, viable cells of known division history cannot subsequently be obtained for functional studies.
The cells of the immune system undergo marked proliferation and differentiation during the course of an immune
response. The relative lack of an organized structure of the lymphohaemopoietic system, in contrast with other
organ systems, makes lineage interrelationships difficult to study. Coupled with the remarkable degree of mobility engendered by recirculation, the differentiation occurring along with cell division in the immune system has
not been readily accessible for investigation. The present article reviews the development of a cell division analysis procedure based on the quantitative serial halving of the membrane permeant, stably incorporating fluorescent
dye carboxyfluorescein diacetate succinimidyl ester (CFSE or CFDA, SE). The technique can be used both in vitro
and in vivo, allowing eight to 10 successive divisions to be resolved by flow cytometry. Furthermore, viable cells
from defined generation numbers can be sorted by flow cytometry for functional analysis.
Key words: carboxyfluorescein diacetate succinimidyl ester, dye dilution, lymphocyte division.
Introduction
The technique for analysing cell division using serial dilution
of the fluorescein-based dye carboxyfluorescein diacetate
succinimidyl ester (CFSE), described in 1994,1 developed
from our attempts to achieve long-term tracking of intravenously transferred lymphocytes. Previously, workers in the
Parish laboratory at the John Curtin School of Medical
Research had successfully tracked lymphocytes in vivo for 8
weeks using CFSE,2 a membrane-permeant dye that covalently attaches to free amines of cytoplasmic proteins.
However, one of our aims was to combine fluorescein-based
CFSE staining with an antibody against fluorescein, thereby
allowing the position of CFSE-labelled lymphocytes to be
determined many weeks after transfer. During the process of
investigating the longevity of CFSE-labelled lymphocytes
after transfer using flow cytometry, I noted a small but significant number of events that had one-half, one-quarter and
one-eighth of the fluorescence of the majority of events, as
shown in Fig. 1. Simultaneous staining with a B cell-specific
antibody revealed the events of decreased fluorescence intensity to be confined to the B cell compartment.1,3 Such a
pattern could most adequately be explained by cell division.
Correspondence: A Bruce Lyons, Discipline of Pathology,
The University of Tasmania, GPO Box 252-29, Hobart, Tas. 7001,
Australia. Email: <[email protected]>
Received 11 August 1999; accepted 11 August 1999.
Demonstration that cell division is responsible for
CFSE serial dilution
To formally demonstrate that the dilution of the fluorescence
was due to cell division, CFSE staining was used in conjunction with the well-established Hoechst quenching method.
Bromodeoxyuridine (BrdU) is incorporated into the DNA of
dividing cells, which allows the identification of up to two
rounds of cell division because BrdU-substituted DNA
quenches the fluorescence of the DNA-binding dye Hoechst
33342. Figure 2 shows that, as predicted, cells with half and
one-quarter the fluorescence intensity of CFSE fell within the
first and second cell divisions identified by the Hoechst
quenching technique.
Long-term survival and division tracking of
transferred lymphocytes using CSFE
Due to the exquisite sensitivity of modern flow cytometers,
CFSE-labelled murine lymphocytes transferred intravenously
to syngeneic recipients can be tracked in vivo for at least
6 months. After transfer, spleen cell suspensions were stained
with anti-CD45R-phycoerythrin (PE) to identify B cells.
Figure 3a–c shows contour plots of CD45R expression against
CFSE fluorescence of transferred cells, demonstrating division
within the B cell compartment. Note that a small amount of
T cell division is apparent from 12 weeks post transfer.
Figure 3d–f shows the fluorescence intensity of CFSE-labelled
lymphocytes compared with the autofluorescence levels of
the recipient’s own cells. Even at 24 weeks post injection,
510
AB Lyons
Figure 1 A proportion of carboxyfluorescein diacetate succidimidyl ester (CFSE)-labelled lymphocytes intravenously transferred to
4-week-old mice undergo serial halving of CFSE label. The original observation. Syngeneic splenic lymphocytes (3 × 107) labelled with
CFSE were injected intravenously into unmanipulated BALB/c recipients. Ten days after injection, spleen cell suspensions were prepared and
50 000 events collected by flow cytometry. Dot plot of forward scatter against CFSE fluorescence. (b) Histogram of CFSE fluorescence. Note
peaks evenly spaced on a logarithmic scale, suggestive of cell division. FL1, green fluorescence, logarithmic scale.
resolution of at least two division cycles is possible (Fig. 3f),
demonstrating the ability of this technique to allow the analysis of long-term survival and proliferative behaviour of transferred lymphocytes. Studies into the longevity and
proliferation behaviour of T lymphocytes are beginning to
provide insight into the still controversial areas of immunological memory and immune class regulation.4–7
Independent assessment of cell division within
subpopulations of lymphocytes is possible using CSFE
One of the major advantages of the CFSE division technique
is that dividing cells can be identified in complex mixtures by
immunophenotyping with appropriately conjugated monoclonal antibodies. Because CFSE acquires identical spectral
characteristics as fluorescein when carboxyl groups are
cleaved by cytoplasmic esterases, it is efficiently excited by
the 488 nm argon laser standard in most flow cytometers,
allowing concurrent use of fluorochromes including phycoerythrin (PE) and tandem dyes such as PE/Texas Red and
PE/Cy5 or peridinin chlorophyll-alpha protein (PerCP). Even
more complex studies can be performed using multilaser
instruments, allowing further fluoro-chromes such as
allophycocyanin (APC) to be used.
Figure 4 shows the response of murine splenic lymphocytes to different combinations of anti-CD3 and bacterial
lipopolysaccharide, demonstrating the power of this technique used in vitro. Table 1, derived from the data generated
from Fig. 4f, demonstrates the fidelity of partitioning of
CFSE fluorescence over a number of cell divisions.
This approach has also been used to probe for differentiation changes in homogenous populations of cells given
appropriate stimulus, such as B lymphocyte differentiation
and isotype switching,8–11 T lymphocyte differentiation,4,5,12
differentiation of haemopoietic precursors13 and NK cells14
and in evaluation of clinically relevant antigen-specific
responses of human peripheral T cells15 (see also further
reviews in this issue).
Carboxyfluorescein diacetate succinimidyl ester can be
used to label populations of lymphocytes at different
fluorescence intensities, allowing them to be
independently tracked in vivo
A major feature of CFSE staining is the ability to manipulate
the final fluorescence intensity of a stained population of
cells. As the fluorescence intensity obtained is directly proportional to the CFSE concentration during the staining
process, as well as the duration of staining, it is possible to
differentially label two or more cell populations. Figure 5
illustrates an experiment where murine lymphocytes were
‘fully’ labelled with CFSE, or labelled using one-quarter or
one-sixteenth of the CFSE concentration during staining.
Equal numbers of fully and one-quarter or one-sixteenth
labelled cells were mixed and cells injected into recipient
mice. After 4 days, spleens were removed and suspensions
made. After staining with a PE-conjugated B cell-specific
antibody (CD45R), events above a threshold of CFSE fluorescence were collected by flow cytometry. Figure 5 (lower
panels) shows that the two populations can be independently
Analysis of division by dilution of CFSE
511
Figure 2 Cell cycle analysis of a proliferating carboxyfluorescein diacetate succidimidyl ester (CFSE)-stained population of lymphocytes, coupled with bromodeoxyuridine (BrdU) quenching of Hoechst 33342 fluorescence, demonstrates that sequential halving of CFSE
fluorescence is due to cell division. (a) Fluorescence histograms of CFSE fluorescence. A, unstained control anti-CD3/LPS-stimulated
(proliferating) population; B, CFSE-stained anti-CD3/LPS-stimulated (proliferating) population; C, CFSE-stained unstimulated (nonproliferating) population. Note region markers over histogram B delineating undivided (M1), first division (M2) and second division (M3).
(b) Bivariate contour plot showing 7-AAD and Hoechst 33342 staining of cells in histogram B of (a). Vertical axis shows 7-AAD binding,
which distinguishes the G0/G1, S and G2/M phases of the cell cycle as shown. As a result of culture in the presence of bromodeoxyuridine (BrdU) these cells show typical quenching of Hoechst 33342 (horizontal axis), allowing two rounds of cell division to be visualized.
The regions shown delineate undivided (R1), first division (R2) and second division (R3). (c) Bivariate contour plot of cells under region
marker M1 of cells in histogram B (a). Note that the majority of these events fall within region 1, corresponding to undivided cells. (d)
Bivariate contour plot of cells under region marker M2 of cells in histogram B of (a). Note that the majority of these events fall within
region 2, corresponding to the first generation. (e) Bivariate contour plot of cells under region marker M3 of cells in histogram B of (a).
Note that the majority of these events fall within region 3, corresponding to the second generation. (Reprinted from Journal of Immunological Methods, volume 171, AB Lyons and CR Parish, Determination of lymphocyte division by flow cytometry, pages 131–7, Copyright 1994,
with permission from Elsevier Science.)
512
AB Lyons
Figure 3 Murine lymphocytes
labelled with carboxyfluorescein
diacetate succidimidyl ester
(CFSE) can be tracked in vivo for
at least 24 weeks. Murine splenocytes (5 × 107 per mL) were incubated at 37°C for 30 min at a
concentration of 10 µmol/L CFSE,
washed and 3 × 107 cells transferred to unmanipulated syngeneic BALB/c recipients via tail
vein injection. At various intervals
up to 24 weeks, recipients were
killed and splenocyte suspensions
prepared. Suspensions were stained
with CD45R-phycoerythrin (PE)
to allow B cells to be identified.
(a–c) Contour plots of CFSE fluorescence (gated above autofluorescence of unstained control
splenocytes) against CD45R-PE
fluorescence of samples taken 1,
12 and 24 weeks post-transfer,
respectively, demonstrating division within the B cell compartments. (d–f) Histograms derived
from the same data of CFSE fluorescence of transferred (j) versus
unstained
splenocytes
(—),
demonstrating that up to two division cycles can still be resolved
after 24 weeks. FL1, green fluorescence, logarithmic scale.
tracked in vivo. Antibody staining allows even more complex
analyses to be achieved, as shown in Fig. 5 (upper panels)
where four cell subpopulations are tracked. This approach has
been successfully used to determine the behaviour of transgenic B cells specific for hen egg lysozyme (HEL) after
transfer into different antigenic environments, using quarter
CFSE labelled non-transgenic B cells as an injection
control, allowing accurate measurement of cell survival and
migration.16 It is also possible to analyse cell division under
these conditions, provided antibodies are available to distinguish between the different transferred populations.16
some in vivo experiments, especially those using transgenic
T and B cells, the majority of transferred cells may undergo
division. Therefore, it is important to be able to determine
the fluorescence intensity of non-divided cells to enable
kinetic analyses to be performed. Figure 6 shows data from
an experiment where labelled splenic lymphocytes are both
transferred to syngeneic recipients, and set up in culture
under different conditions. Over a typical time course of 3
days, the peak of undivided cells has the same fluorescence
intensity in vitro or in vivo, allowing combined in vivo/
in vitro experiments to be performed.
Labelled lymphocytes maintain identical fluorescence
characteristics in vitro and in vivo
Discussion
As seen in Fig. 3a, CFSE fluorescence intensity in nondivided cells declines slowly over time due to catabolism. In
Since its introduction in 1994, the CFSE dye dilution technique has become the method of choice for investigating
cell division-linked differentiation of lymphocytes and
Analysis of division by dilution of CFSE
513
Figure 4 In vitro culture of
splenic lymphocytes stained with
carboxyfluorescein diacetate succidimidyl ester (CFSE) allows differential B and T cell division in
response to polyclonal stimuli to
be analysed. (a) Bivariate dot plot
of cells from a control unstimulated culture. Vertical axis shows
anti-CD45R-phycoerythrin (PE)
staining, horizontal axis shows
CFSE staining. (b) Bivariate dot
plot of cells from an LPS-stimulated culture shows division
occurring in a subpopulation of
anti-CD45R staining B lymphocytes. (c) Bivariate dot plot of
cells from an anti-CD3 stimulated
culture shows division occurring
in a subpopulation of non-antiCD45R staining T lymphocytes,
as well as in a minority of B lymphocytes. (d) Bivariate dot plot of
cells from a culture stimulated
with both LPS and anti-CD3
demonstrates pronounced cell
division of B and T lymphocytes.
(e) Fluorescence histogram of
gated B lymphocytes (R1) from
(d) demonstrates sequential half
loss of CFSE fluorescence. (f)
Fluorescence histogram of gated T
lymphocytes (R2) from (d)
demonstrates sequential half loss
of CFSE fluorescence. Analysis of
the fluorescence intensity of peaks
under markers was used to generate Table 1. (Reprinted from the
Journal of Immunological Methods,
volume 171, AB Lyons and CR
Parish, Determination of lymphocyte division by flow cytometry,
pages 131–7, Copyright 1994, with
permission from Elsevier Science.)
Table 1
Peak
M1
M2
M3
M4
M5
M6
The CFSE fluorescence intensity of proliferating T cells closely follows the predicted sequential halving due to cell division
Mean fluorescence intensity
1775
869
432
221
114
61
Fraction of undivided cell fluorescence
Actual
Predicted
1.0
0.49
0.243
0.125
0.064
0.034
1.0
0.5
0.25
0.125
0.0625
0.03125
Data is a comparison between mean fluorescence intensities of peaks in Fig. 4f. (Reprinted from the Journal of Immunological Methods,
volume 171, AB Lyons and CR Parish, Determination of lymphocyte division by flow cytometry, pages 131–7, Copyright 1994, with permission from Elsevier Science). CFSE, carboxyfluorescein diacetate succinimidyl ester.
514
AB Lyons
Figure 5 Differential labelling with carboxyfluorescein diacetate succidimidyl ester (CFSE) allows two populations of lymphocytes to
be tracked simultaneously in vivo. Murine splenocytes were stained with CFSE at a final concentration of 5 µmol/L (fully labelled),
1.25 µmol/L (one-quarter labelled) and 0.3125 µmol/L (one-sixteenth labelled) for 10 min at 37°C. After washing away excess stain, equal
quantities of fully labelled and one-quarter or one-sixteenth labelled cells were mixed and 3 × 107 cells injected intravenously into syngeneic BALB/c recipients. After 4 days, spleen suspensions were prepared and stained with a phycoerythrin (PE)-conjugated anti-CD45R
B cell specific antibody and CFSE-labelled events collected for analysis. The upper pair of dot plots represent the coinjected fully and
one-quarter labelled cells (a) and the fully and one-sixteenth labelled cells (b) stained with anti-CD45R-PE. Note that due to the intensity of CFSE fluorescence at 4 days, adequate compensation between the PE and CFSE collecting channels is not possible. However, by
day 6 the fluorescence has dropped sufficiently to allow complete compensation (not shown). The lower panels are histograms of CFSE
fluorescence derived from the upper panels, demonstrating the accuracy of resolution of the coinjected populations. FL1, green fluorescence, logarithmic scale.
haemopoietic cells and for the investigation of kinetics of
proliferation during immune responses (reviewed by Lyons
and Doherty17 and Lyons et al.18). Carboxyfluorescein diacetate succinimidyl ester is equally partitioned between
daughter cells with remarkable fidelity. With a uniformly
sized starting population, such as resting B or T lymphocytes,
more than eight sequential cell divisions can be visualized
without the need for computer-based deconvolution.8
In addition to being able to explore cell surface molecule
changes linked to cell division, the technique has also been
combined with techniques measuring production of bioactive
materials such as cytokines,4,5 as well as with uptake of BrdU
to allow rates of division to be determined.9 A major advantage of the technique is the ease in which viable cells of
defined generation number can be obtained by flow cytometric cell sorting for functional investigations. An everexpanding group of different cell types have proved amenable
to study by this technique, including T and B lymphocytes,
NK cells and haemopoietic precursors. The ability to use the
technique for both in vivo and in vitro studies has greatly
enhanced our understanding of the immune system at the
individual cell level.
Analysis of division by dilution of CFSE
515
Figure 6 Lymphocytes labelled with carboxyfluorescein diacetate succidimidyl ester (CFSE) can be simultaneously tracked in vivo and
in vitro, because undivided fluorescence intensity remains constant. Splenic lymphocytes were labelled with CFSE and either injected via
the lateral tail vein into syngeneic hosts or cultured in vitro with maximally stimulating amounts of anti-CD3 plus anti-CD28. At intervals of
24 (a), 48 (b) and 72 h (c), spleen cell suspensions from recipients were prepared and CFSE fluorescence of cultured (j) and in vivo samples
(—) analysed by flow cytometry. The markers identify the undivided cell peaks and the means differ by less than 5%. Thus, decay of undivided
cell fluorescence remains constant in vivo and in vitro, allowing direct comparison of cultured and transferred cell division behaviour. FITC LOG,
green fluorescence, logarithmic scale.
References
1 Lyons AB, Parish CR. Determination of lymphocyte division by
flow cytometry. J. Immunol. Meth. 1994; 171: 131–7.
2 Weston SA, Parish CR. New fluorescent dyes for lymphocyte
migration studies. Analysis by flow cytometry and fluorescence
microscopy. J. Immunol. Meth. 1990; 133: 87–97.
3 Lyons AB. Pertussis toxin pretreatment alters the in vivo division
behaviour and survival of B lymphocytes, after intravenous
transfer. Immunol. Cell Biol. 1997; 75: 7–12.
4 Gett AV, Hodgkin PD. Cell division regulates the T cell cytokine
repertoire, revealing a mechanism underlying immune class regulation. Proc. Natl Acad. Sci. USA 1998; 95: 9488–93.
5 Bird JJ, Brown DR, Mullen AC et al. Helper T cell differentiation is controlled by the cell cycle. Immunity 1998; 9: 229–37.
6 Kurts C, Carbone FR, Barnden M et al. CD4+ T cell help
impairs CD8+ T cell deletion induced by cross-presentation of
self-antigens and favors autoimmunity. J. Exp. Med. 1997; 186:
2057–62.
7 Lee WT, Pelletier WJ. Visualizing memory phenotype development after in vitro stimulation of CD4+ T cells. Cell. Immunol.
1998; 188: 1–11.
8 Hodgkin PD, Lee JH, Lyons AB. B cell differentiation and
isotype switching is related to division cycle number. J. Exp.
Med. 1996; 184: 277–81.
9 Hasbold J, Lyons AB, Kehry MR, Hodgkin PD. Cell division
number regulates IgG1 and IgE switching of B cells following
CD40L and IL-4 stimulation. Eur. J. Immunol. 1998; 28:
1040–51.
10 Kindler V, Zubler RH. Memory, but not naive, peripheral blood
B lymphocytes differentiate into Ig-secreting cells after CD40
ligation and costimulation with IL-4 and the differentiation
factors IL-2, IL-10, and IL-3. J. Immunol. 1997; 159: 2085–90.
11 Hodgkin PD, Bartell G, Mamchak A, Doherty KV, Lyons AB,
Hasbold J. The importance of efficacy and partial agonism in
evaluating models of B lymphocyte activation. Int. Rev.
Immunol. 1997; 15: 101–27.
12 Wells AD, Gudmundsdottir H, Turka LA. Following the fate of
individual T cells throughout activation and clonal expansion.
Signals from T cell receptor and CD28 differentially regulate the
induction and duration of a proliferative response. J. Clin.
Invest. 1997; 100: 3173–318.
13 Nordon RE, Ginsberg SS, Eaves CJ. High resolution cell
division tracking demonstrates the Flt3 ligand dependence of
human marrow CD34+CD38- cell production in vitro. Br. J.
Haematol. 1997; 98: 528–39.
14 Warren HS, Kinnear BF. Quantitative analysis of the effect of
CD16 ligation on human NK cell proliferation. J. Immunol.
1999; 162: 735–42.
15 Angulo R, Fulcher DA. Measurement of Candida-specific
blastogenesis: Comparison of carboxyfluorescein succinimidyl
ester labelling of T cells, thymidine incorporation, and CD69
expression. Cytometry 1998; 34: 143–51.
16 Fulcher DA, Lyons AB, Korn S et al. The fate of self-reactive Bcells depends primarily on the degree of antigen receptor
engagement and availability of T-cell help. J. Exp. Med. 1996;
183: 2313–28.
17 Lyons AB, Doherty KV. Flow cytometric analysis of cell division by dye dilution. In: Robinson JP, Darzynkiewicz Z, Dean P
et al. (eds). Current Protocols in Cytometry. New York: John
Wiley and Sons, 1998; Unit 9.11.
18 Lyons AB, Hasbold J, Hodgkin PD. Flow cytometric analysis of
cell division history using dilution of CFSE, a stably integrated
fluorescent probe. In: Darzynkiewicz Z, Crissman HA, Robinson JP (eds). Methods in Cell Biology; Cytometry. San Diego:
Academic Press, 1999; in press.