Experimental Cell Research 266, 126 –134 (2001)
doi:10.1006/excr.2001.5200, available online at http://www.idealibrary.com on
Cell Proliferation and CD11b Expression Are Controlled Independently
during HL60 Cell Differentiation Initiated by 1,25␣-Dihydroxyvitamin D 3
or All-trans-Retinoic Acid
Mark T. Drayson, Robert H. Michell,* Jennifer Durham, and Geoffrey Brown 1
LRF Differentiation Programme, Division of Immunity & Infection, and *School of Biosciences, University of Birmingham,
Birmingham B15 2TT, United Kingdom
When 1␣,25-dihydroxyvitamin D 3 (D 3) induces HL60
cells to differentiate to monocytes, a burst of approximately three shortened cell cycles (“maturation divisions”) precedes exit from cell cycle and completion of
maturation. Here we show that similar maturation divisions occur during neutrophil differentiation induced by all-trans-retinoic acid (ATRA), but without
shortening of the cell cycle. Both ATRA and D 3 initiate
these maturation divisions as cells pass through a
“window of sensitivity” during early G1. We also investigated whether the initiation of maturation divisions
and of the expression of CD11b, an early-expressed
maturation marker, are linked. Cells treated with D 3
or ATRA start to express CD11b after 9 –14 h, before
completing the first maturation division. Elutriation
was used to isolate small HL60 cells (almost all in G1)
and larger cells (in G1 and S phases) from unsynchronized populations. When these were cultured with D 3
or ATRA, most reentered cycle synchronously, multiplied, and differentiated. Following D 3 treatment, the
G1-enriched small cells expressed CD11b slightly
faster than unsynchronized cultures or fractions dominated by late G1 cells and/or S phase cells. D 3-induced
CD11b expression occurred at a similar rate even in
G1 cells that were held at the G1/S boundary by thymidine. In conclusion, changes in the control of the
cell cycle that characterize the onset of monocytic and
neutrophil differentiation are only triggered in early
G1, but CD11b expression can be initiated from most
points in the cell cycle. Differentiating agents must
therefore regulate the proliferation and the maturation of differentiating myeloid cells by mechanisms
that are at least partly independent. © 2001 Academic Press
Key Words: HL60 cells; proliferation; differentiation;
retinoic acid; vitamin D 3; gene expression; cell cycle;
CD11b.
1
To whom correspondence and reprint requests should be addressed at Division of Immunity and Infection, Medical School, University of Birmingham, Birmingham B15 2TT, UK. Fax: 0121-4143599. E-mail: G.Brown@bham.ac.uk.
0014-4827/01 $35.00
Copyright © 2001 by Academic Press
All rights of reproduction in any form reserved.
INTRODUCTION
In culture, cells of the promyeloid line HL60 are
oligopotent: they can differentiate to neutrophils,
monocytes, eosinophils, and basophils [1– 4]. Contrary
to earlier views, recent studies have revealed that
HL60 and U937 cells undergoing D 3-induced monocytic
differentiation do not undergo growth arrest immediately and then develop differentiated character. Rather
there is a burst of rapid proliferation, typically of approximately three shortened cell cycles, before growth
arrest occurs [5, 6]. D 3 initiates this period of accelerated cell division just after a cell enters G1. As a result,
one D 3-treated HL60 cell typically gives rise to 6 –12
monocytes in 3– 4 days [6].
It has not been known whether such continued proliferation is also characteristic of HL60 cells differentiating to neutrophils. We therefore examined the extent and kinetics of proliferation during neutrophilic
differentiation induced by all-trans-retinoic acid
(ATRA). Moreover, it is not clear whether a single
temporal program in differentiating cells controls both
the induction of “maturation divisions” and the expression of “markers of differentiation” such as CD11b.
Recent work has suggested that CD11b expression is
not always preceded by cessation of division. For example, differentiating HL60 cells can simultaneously
incorporate bromodeoxyuridine and express proliferating cell nuclear antigen (PCNA) and CD11b before they
undergo growth arrest [7]. Moreover, some variant
HL60 cell sublines continue to proliferate even when
1␣,25-dihydroxyvitamin D 3 (D 3)-treated but also remain capable of the D 3-induced expression of CD14 and
other monocyte markers [8]. Results such as these
imply that monocytic maturation starts while HL60
cells are still proliferating.
In the present study, we show that CD11b expression starts even before cells differentiating in response
to D 3 or ATRA complete the first of the approximately
three cell cycles that they undergo during the terminal
“divide and differentiate” program. We also show that
126
CELL PROLIFERATION, CD11B EXPRESSION, AND HL60 CELL DIFFERENTIATION
127
FIG. 1. The proliferation kinetics during 4 days of exponentially growing and differentiating HL60 single-cell cultures. Cells were
counted daily in wells containing the exponentially growing progeny of a single cell (lines without symbols) and in similar single-cell cultures
to which either 100 nM D 3 (Œ, A) or 100 nM ATRA, in the presence of 0.036 pM D 3 (F, B), was added at time zero. The results are plotted
cumulatively as the percentages of the cultures in which a parental cell had achieved or surpassed a particular number of divisions. For
example, a well recorded as having undergone 1.5 divisions contained 3 cells or as having achieved 3.375 divisions if it contained 11 cells (also
see Fig. 3). The labels for the individual curves indicate for how many days the cells yielding that curve had been cultured. Each curve
combines information from between 191 and 543 cultures.
control of the initiation of maturation divisions and
control of the expression of CD11b are, at least in part,
independent processes.
EXPERIMENTAL METHODS
Cell culture. The methods used for analyzing the differentiation
of single-cell and bulk HL60 cultures were largely as described
before [6, 9; and see also Results and figure legends]. Briefly, the
technique for the single-cell cultures was as follows [for detail, see 6].
Cells from growing serum-free HL60 cultures were diluted to 10
cells/ml in HL60-conditioned medium, and 100-␮l cultures (⬃1 cell
per well) were set up. Wells containing one cell were identified after
3 h and were reexamined 2 h later to identify the 10% or so that had
divided to yield two (smaller and adjacent) daughter cells: these were
in the first 2 h of G1. One-hundred nanomolar D 3 (for monocytes) or
100 nM ATRA plus 0.036 pM D 3 (for neutrophils) was added, either
immediately or after various intervals, to induce differentiation.
Viable cells and differentiated cells (that reduce nitroblue tetrazolium, NBT) were enumerated in each well 4 days later (and sometimes earlier).
Immunodetection of CD11b expression. CD11b on unfixed cells
was labeled with direct fluorochrome conjugates of a mouse monoclonal anti-CD11b antibody (Immunotech phycoerythrin and FITC),
with an isotype-matched conjugate used as a negative control. After
fixation, the cells were analyzed on a Becton Dickinson FACScalibur
flow cytometer and the characteristics of the viable cells (identified
by their light scatter characteristics) were analyzed using
CELLQuest software. To detect both surface and intracellular
CD11b, cells were formalin-fixed, permeabilized (0.5% saponin in
PBS containing 4% fetal bovine serum, FBS), and stained with the
FITC conjugate.
Size fractionation of HL60 cell populations by countercurrent elutriation. Exponentially growing HL60 cells were harvested by centrifugation at room temperature, resuspended in phosphate-buffered
saline containing 5% FBS, and loaded into a Beckman JE-6B centrifugal elutriation rotor fitted with a standard chamber, spinning at
1950 rpm, in a Beckman J-6M centrifuge. The same FBS-containing
buffer at room temperature was pumped through the rotor at progressively increasing flow rates, expelling the smallest cells (Fraction 1) first, followed by progressively larger cells (Fractions 2 and 3).
The elutriated cells were spun down and quickly returned to culture,
using medium conditioned by exponentially growing HL60 cells.
RESULTS
Population Expansion Occurs during Neutrophilic
Differentiation of Single-Cell Cultures
Serum-free HL60 cultures were initiated with a single cell, and growth and differentiation were monitored
for up to 4 days. Exponentially growing cultures were
compared with cultures differentiating in 100 nM D 3
(to monocytes) or 100 nM ATRA (in the presence of
0.036 pM D 3, to neutrophils) [6, 9]. Figure 1 summarizes the proliferative kinetics of many such cultures,
expressed in terms of the cumulative number of divisions undergone by one parental cell. Division continued at an unchanged rate throughout the 4 days in
control wells, with the resulting curve (Fig. 1, lines
without symbols) moving rightward by slightly less
than one division per day (see below for further discussion).
Cultures that were treated with D 3 (Fig. 1, left)
showed a transient acceleration of their proliferation.
Few of the control cultures moved beyond one division
during the first day, but the initial progeny of ⬃30% of
128
DRAYSON ET AL.
the D 3-treated cells embarked on their second (or occasionally third) division within 1 day. Cells in more than
one-third of the D 3-treated wells embarked on their
third round of division within 2 days, whereas only half
as many control wells progressed this far by then. The
proliferation of D 3-treated cells slowed dramatically
during the third day, and few divided during day 4.
These results substantially expand the evidence that a
burst of accelerated proliferation occurs during D 3induced monocytic differentiation of HL60 cells [5, 6].
ATRA-treated cultures continued to divide at much
the same pace as control cultures for 3 days, and proliferation slowed markedly during the fourth day (Fig.
1, right): there is no further proliferation in such cultures after day 4 [9]. Although the ATRA-treated cells
did not show accelerated proliferation like that induced
by D 3, it is clear that ATRA-induced neutrophilic differentiation does involve a period of continued population expansion before the cells drop out of cycle and
terminally mature.
These observations establish that the maturation
divisions of cells en route to neutrophils and of cells en
route to monocytes are different in character.
The Median Cell Cycle Period of HL60 Cells
Is ⬃27.5 h
Knowledge of the median cell cycle period and of the
periods spent in each phase of the cell cycle was essential for the planning and interpretation of the experiments that follow. Previous analyses of population expansion in exponentially dividing bulk HL60 cultures
have generally yielded doubling periods of around 24 h
[e.g., 6]. When the daily cell totals from many singlecell cultures were pooled, so as to simulate a small
“bulk” culture, they also yielded an apparent doubling
time of 24 h (not shown).
Such results have been taken to mean that the typical cell cycle period of a multiplying HL60 cell is ⬃24
h. However, this ignores two characteristics of exponentially multiplying populations: the cycle periods of
individual cells vary widely, and the progeny of any
single founder cell accumulate exponentially. In the
one-cell cultures, the cell cycle periods in individual
wells were distributed symmetrically about a median
value of about 3.5 cycles in 4 days, rather than four
(Fig. 2): the slowest and fastest deciles completed ⱕ2.4
and ⱖ4.6 cycles, respectively. Only about one-third of
the cultures completed four or more cycles in 4 days.
Cells at the median traverse 3.5 cycles in 4 days (i.e.,
0.88 divisions per day), which translates to a cell cycle
period of ⬃27.5 h rather than ⬃24 h.
The apparent disparity between the median cycle
period of ⬃27.5 h and the shorter bulk doubling time of
23–24 arises from the exponential nature of proliferation. A cell showing the median proliferation rate of
FIG. 2. The distribution of cell division outcomes among many
single-cell cultures. The figure presents a fitted Poisson trend curve
derived from the first derivative of the cumulative curve representing the proliferative outcomes over 4 days of the 466 single-cell
control cultures (as plotted in Fig. 1). It shows that the most common
behavior of these cells was to achieve ⬃3.5 divisions during 4 days of
exponential proliferation in such cultures.
0.88 divisions per day yields 12 progeny in 4 days.
However, cells that divide fast (e.g., 4.5–5 divisions,
yielding ⬃28 progeny—16 in excess of the median)
contribute disproportionately more to the final population than the similar number of cells that divide slowly
(e.g., 1–2.5 divisions, yielding ⬃4 progeny— or 8 below
the median).
We know the cell cycle distribution of our exponentially growing HL60 cultures (38% are in G1, 43% in S,
and 19% in G2/M [6]), so we can calculate that the
median behavior of these HL60 cells was to spend ⬃10
h in G1, ⬃12 h in S, and ⬃5 h in G2/M.
Commitment to Maturation Divisions Occurs Early in
G1 for D 3 and ATRA
Proliferating HL60 cells that encounter D 3 during
early G1 immediately switch on a terminal divide “two
to five times and differentiate to monocytes” program,
and cells that encounter D 3 later initiate this program
only when they start the next cycle [6]. As a result,
cells that pass through early G1 before D 3 is added
undertake one division more than cells that encounter
D 3 at the start of a cycle [6].
Figure 3 shows that this is also the case for ATRA. It
depicts, cumulatively, how many divisions individual
single-cell cultures underwent when D 3 or ATRA
was added during the first 2 h of G1 (curves labeled
⬍2 h) or between 6 and 24 h into cycle (curves labeled
⬎6⬍24 h): for relevant methods, see [6] and the legend
to Fig. 3. The proliferative profile of cells that received
D 3 or ATRA early in G1 (for which the median outcome
is ⬃2.5 divisions, yielding ⬃6 cells) is centered at one
CELL PROLIFERATION, CD11B EXPRESSION, AND HL60 CELL DIFFERENTIATION
129
The Switch in Cell Cycle Behavior Also Occurs
during Early G1 in Bulk Cultures
FIG. 3. Growing HL60 cells are switched into their terminal
maturation divisions by D 3 and ATRA only during early G1. Singlecell cultures were established and reexamined 2 h later to identify
cells that immediately underwent mitosis (see Experimental Methods and [6]). D 3 or ATRA was added to these cells either immediately
[curves labeled ⬍2 h, representing the interval between mitosis and
the addition of D 3 (Œ) or ATRA (F)] or at later times: 1–3 h after
mitosis (D 3, 䊐); 2– 4 h after mitosis (D 3, 〫), or more than 6 but less
than 20 h after mitosis (D 3, ‚; or ATRA, E). Cells were enumerated
4 days later. The results are plotted, as in Fig. 1, in terms of the
number of divisions undergone by a single cell that was present at
the time of ATRA or D 3 addition. Wells in which D 3 or ATRA was
added at some time between 6 and 20 h after the initial cell divided
all gave similar results, so the ⬎6 ⬍ 20 h curves depict these pooled
data. Each curve summarizes information from between 45 and 221
single-cell cultures.
division fewer than for cells that encountered D 3 or
ATRA later and responded only after completing their
ongoing cycle (median ⬃3.5 divisions, ⬃12 cells).
Additional studies with D 3 showed that the proportion of single-cell cultures that undergoes an “extra”
division increases progressively as D 3 treatment is delayed beyond 2 h into G1. All cells were D 3 sensitive in
the first 2 h after G1 (Fig. 3, ⬍2 h), but about onequarter of the cultures that encountered D 3 in the 1- to
3-h window were not immediately D 3 sensitive (Fig. 3).
The 1- to 3-h window includes the fully sensitive second hour of the 0- to 2-h window, so about half of the
cells must lose D 3 sensitivity during the third hour of
G1. There was a further loss of sensitivity during the 2to 4-h window (Fig. 3), and a small number of cultures
that encountered D 3 4 – 6 h into G1 behaved like the
⬎6⬍24 h population (not shown).
These results establish, at least in relation to initiating maturation divisions, that HL60 cells are
ATRA/D 3 sensitive at the start of G1 but undergo a
rapid transition to an ATRA/D 3-insensitive state between ⬃2 and ⬃4 h later. The time resolution of our
methods did not permit us to determine whether the
transitions for ATRA and D 3 occur at precisely the
same time.
Single-cell cultures have provided conclusive evidence that an encounter with D 3 or ATRA early in G1
sets in train a restricted number of maturation divisions. However, most studies of differentiating myeloid
cells examine bulk cultures, so we used cell populations
enriched in early G1 cells (obtained by countercurrent
centrifugal elutriation) to gain confirmatory information from such cultures.
The elutriation technique fractionates HL60 cells on
the basis of cell size: the smallest cells, which are
mainly those that divided recently, emerge from the
rotor first. When such an HL60 subpopulation [Fraction 1, comprising the smallest 4.99 ⫾ 1.84% (n ⫽ 12)
of the cells] was subjected to cell cycle analysis, all cells
were in G1/G0 (Fig. 4A). Two populations of larger cells
were also harvested. Fraction 2 comprised a mixture of
cells in late G1 and in S phase (in an ⬃3:1 ratio, Fig.
4B). Fraction 3 was dominated by cells in S phase, but
included some G1 and a few G2/M cells (Fig. 4C).
When Fraction 1 cells were returned to culture at
250,000 cells/ml, ⬃70% started to move into S phase
after 5–7 h (Fig. 5). For the following ⬃10 h these cells
progressed through S phase or were in G2/M, and they
then began to enter G1 of the next cycle. The remaining
20 –30% of the elutriated cells failed to leave G0/G1
when returned to culture.
To estimate whether the point in the cell cycle at
which D 3 is encountered influences how much Fraction
1 cells proliferate before maturing, Fraction 1 cultures
were either immediately exposed to D 3 (when all were
in early G1 or in G0) or given D 3 after 12 h in culture
(by which time most had entered S phase but none had
divided; Fig. 5). Four days later, we counted their progeny and assessed what proportion had differentiated.
Cells that were immediately treated with D 3 multiplied
approximately eightfold, but those that were D 3
FIG. 4. The cell cycle distributions of the size-fractionated HL60
subpopulations obtained by elutriation. Fractions 1– 3, representing
HL60 subpopulations of progressively increasing average cell size,
were obtained by countercurrent elutriation (see Experimental
Methods). These were stained with propidium iodide and their cell
cycle profiles were determined by FACS analysis. The resulting
profiles are from: (A) Fraction 1; (B) Fraction 2; and (C) Fraction 3.
130
DRAYSON ET AL.
FIG. 5. The small G1 cells in elutriated Fraction 1 reenter cell
cycle when cultured. Freshly isolated Fraction 1 cells were immediately returned to culture. Their cell cycle status was analyzed immediately and at intervals thereafter. The solid line with plotted
points records the percentage of cells that were in S, G2, or M phases,
and the dotted line those in G1. Data were pooled from six independent experiments, and each point is from one to three determinations.
treated only after 12 h in culture proliferated substantially more (Table 1).
If we assume that the cultures in Table 1 behaved in
the same way as in the single-cell cultures [see 6], we
can estimate the likely spectrum of proliferative behaviors of the cells therein. In the “delayed D 3” cultures, a
cell has two options: (a) it can fail to reenter cycle and
so contribute one cell to the final tally; or (b) it can
complete its ongoing cycle, encounter D 3 at the start of
the next G1 and thereafter yield ⬃15 progeny (see Fig.
3 of Ref. [6]). The observed yield, of ⬃10 progeny per
starter cell from the delayed D 3 cultures, is about twothirds of the maximum potential yield. This tallies
with the cell cycle analyses, which suggested that
around two-thirds of the Fraction 1 cells that were
returned to culture reentered cycle and proliferated.
Assuming that two-thirds of the cells in the immediate
D 3 cultures also reentered cycle, then the immediate
D 3 cells averaged ⬃11–12 progeny (7.8 ⫻ 3/2), a value
midway between typical yields from cells that encounter D 3 in early G1 (⬃8 progeny; see [6]) or encounter D 3
later (⬃15 progeny; [6]). This suggests that around half
of the small G1 cells in Fraction 1 were near enough to
the start of G1 to embark on their D 3-triggered maturation divisions immediately upon their return to culture.
When Do Cells Start to Express CD11b, a Marker
Characteristic of Monocytes and Neutrophils?
Monocytes and neutrophils express surface CD11b,
the ␤-subunit of the integrin-␣ M␤ 2 (also known as
CD11b/CD18, MAC-1, or CR3), but exponentially proliferating HL60 cells do not. Most HL60 cells become
CD11b positive following treatment with D 3 (⬃90% at
3– 4 days) or ATRA (⬃80% at 4 –5 days) [e.g., 6, 9]. It is
common practice to assay the expression of such differentiation markers when cells have fully matured
[typically at 3– 4 days (monocytes) or 4 –5 days (neutrophils)], but it is known that CD11b expression starts
relatively quickly [e.g., 7]. The speed of this expression
offered a way to determine to what extent there is
functional linkage between the onset of CD11b expression and the recruitment of cells into the proliferate
and mature program.
When unsynchronized cell populations were treated
with ATRA or D 3, their recruitment into the CD11bpositive compartment was essentially complete in ⬃36
h (Fig. 6A). During ATRA treatment, CD11b expression started after 12–14 h, and thereafter cells were
recruited approximately linearly to the CD11b-positive
population for about 1 day (Fig. 6A). CD11b expression
started a few hours earlier in D 3-treated cells, and
most cells became CD11b positive during only about
15 h (Fig. 6A).
Intracellular CD11b was not immunologically detectable any earlier than cell surface CD11b (not
shown). The hiatus between ATRA or D 3 addition and
surface CD11b expression therefore reflects a genuine
delay in CD11b expression rather than a postponement
of the transit of newly synthesized CD11b to the cell
surface.
The ⬃27-h cell cycle period of the proliferating cells
to which differentiating agents were added (see above)
is much longer than the ⬃15-h period during which
CD11b-positive cells were recruited after treating unsynchronized cells with D 3. Initiation of CD11b expression by D 3 cannot therefore be restricted to the brief
time window in early G1 when maturation divisions
are initiated— or to any other relatively brief segment
of the cell cycle. This immediately raises the question
TABLE 1
Elutriated HL60 Cells That Were Exposed to D 3 from
Early G1 Yielded Fewer Differentiated Progeny than Cells
Treated from Midcycle
Treatment
D 3 added immediately
D 3 added after 12 h
Increase in cell number (fold)
(mean ⫾ SD (n))
7.8 ⫾ 0.23 (5)
9.9 ⫾ 0.9 (6)
P ⬍ 0.002
(immediate D 3 vs D 3 after 12 h)
Note. Elutriated HL60 cells in G1 were returned to culture for 4
days. D 3 was added to some cultures immediately and to others after
12 h (by which time the cell number remained unchanged but the
most cells had reentered cycle and progressed into S phase and/or
G2/M; see Fig. 5). After 4 days in culture, cells were counted and
their differentiation was assessed: ⬎80% of the cells had differentiated under both conditions (not shown).
CELL PROLIFERATION, CD11B EXPRESSION, AND HL60 CELL DIFFERENTIATION
FIG. 6. Kinetics of the expression of CD11b in HL60 cells exposed to D 3 or ATRA. HL60 cells (exponentially growing bulk cultures or elutriated fractions) were returned to culture and treated for
the indicated periods with D 3 or ATRA (concentrations as in Fig. 1).
(A) The time courses of CD11b expression in bulk cultures that were
exponentially growing when D 3 (F) or ATRA (}) was added (data
from 10 independent experiments, with each point the average of one
to six determinations). (B) The time courses of CD11b expression in
elutriated and cultured HL60 cell subpopulations. Data are presented for Fraction 1 (E), for Fraction 2 (■), and for Fraction 3 (Œ)
and also for Fraction 1 cells that were cultured for 6 h before D 3
addition (F). To facilitate comparison between A and B, the trend
curve for CD11b expression in D 3-treated bulk cultures from A is
reproduced as the gray line in B. The Fraction 1 data are from eight
independent experiments (with one to four analyses contributing to
each point), and other data are derived from three experiments.
131
of whether initiation of CD11b expression may be permitted through most or all of the cycle? We used two
sets of experiments to address this question. First, we
compared the kinetics of CD11b expression in unsynchronized cells and in the three semisynchronized cell
subpopulations obtained by elutriation: the small cells
of Fraction 1; Fraction 2 (cells in G1 and S phases, in a
3:1 ratio); and Fraction 3 (mainly cells in S phase) (see
above). D 3 was also added to Fraction 1 cells 6 h after
they were returned to culture, by which time most were
in late G1 and about to enter S phase (see above and
Fig. 5).
Remarkably, the differences between the time
courses of CD11b expression in the four cell populations to which D 3 was immediately added were slight
(Fig. 6). CD11b expression started at about the same
time, after 8 –11 h, in unsynchronized cells and in each
elutriated fraction, and the rate of recruitment of
CD11b-positive cells thereafter was similar in unsynchronized populations and in Fractions 2 and 3 (Fig. 6).
CD11b expression was slightly quicker in the small G1
cells of Fraction 1 (Fig. 6B). They progressed from
being CD11b negative to almost maximal CD11b expression during little more than 10 h. Most Fraction 1
cells that were exposed to D 3 for 15–18 h expressed
CD11b (Fig. 6B), even though they had only progressed
from G1 into S phase (see Fig. 5). When small G1 cells
were cultured to late G1 or early S phase before D 3
addition, CD11b expression was slightly slower: the
cells behaved more like Fraction 2 or Fraction 3 cells
(Fig. 6B).
Finally, when the small G1 cells of Fraction 1 were
cultured for 6 h with thymidine, the cell cycle halted
before cells entered S phase (Fig. 7C) (serum-grown
cells were used for these experiments, as they tolerated
the thymidine treatment better). Despite the continued
presence of thymidine, these cells responded to D 3 addition with prompt CD11b expression. After 6 h of
thymidine treatment followed by an additional 21.5 h
FIG. 7. CD11b is expressed normally even by cells in which cell cycle progression has been blocked at S phase entry by thymidine.
Fraction 1 cells (in G1) were returned to culture in the presence or absence of 1 mM thymidine. D 3 was added to thymidine-free cells at the
start of the culture and to the thymidine-containing cultures after 6 h, by which time cells were accumulating at the G1/S boundary. Each
panel depicts the cell cycle status of the cells and (inset) a FACS analysis of CD11b expression. (A) Newly isolated Fraction 1 cells (no
thymidine, no D 3). (B) Fraction 1 cells exposed to D 3 for 21.5 h. (C) Thymidine-treated Fraction 1 cells (6 h) to which D 3 was then added for
an additional 21.5 h.
132
DRAYSON ET AL.
with D 3 and thymidine, 72–75% of the thymidineblocked cells expressed CD11b (Fig. 7C). Similar observations have been made previously with HL60 cells
that were thymidine blocked from exponential culture
and then exposed to D 3 [10]. Thymidine-blocked cells
that were treated with ATRA also expressed CD11b to
the same extent and with similar kinetics as cells that
were treated only with ATRA (not shown).
DISCUSSION
Maturation Divisions during Myeloid Differentiation
We recently established, by directly counting the
progeny of D 3-treated HL60 cells, that a 2-day period of
rapid population expansion is an intrinsic element of
their D 3-initiated monocytic differentiation program
[6]. Independently, Rots et al. used a formazan-based
dye reduction assay to obtain evidence for a transient
period of rapid proliferation when U937 cells, another
myelomonocytic cell line, embark on D 3-induced monocytic differentiation [5].
Here we show that continued proliferation, in this
case for about 3 days, is also a part of the ATRAinitiated neutrophilic differentiation program of HL60
cells. In this case, there is no shortening of the cell
cycle (Fig. 1B). A similar pattern may be evident in a
recent study of ATRA-induced neutrophilic differentiation of NB 4 promyeloid cells [12]. The authors of that
study interpreted their data (in Fig. 3 of Ref. [12]) as
evidence that NB 4 cells promptly drop out of cycle
following ATRA addition, but the published FACS profiles do not seem to support this conclusion. These show
substantial numbers of cells in S phase and in G2/M
long after ATRA addition—for example, 2 days after
ATRA addition there were two-thirds as many cells in
S phase as at the start. Although it is possible that
some of these S and G2/M cells may have halted division in midcycle, the more obvious interpretation of
these data is that during the onset of ATRA-induced
neutrophilic differentiation of NB 4 cell proliferation
continues for 2–3 days, as it does in HL60 cells.
Our previous evidence that the only phase of the cell
cycle during which cells can be launched into maturation divisions is during the first 2–3 h of G1 has also
been strengthened [6]— cells that receive a differentiating agent later do not embark on maturation divisions until G1 of the next cycle. This is true both for D 3
(monocytic differentiation) and for ATRA (neutrophil
differentiation), and we detected no difference between
the durations of the sensitive “windows” for the two
pathways (Fig. 3). The particular sensitivity of cells in
early G1 to the instruction to initiate maturation divisions was further confirmed when we showed that
early G1-enriched HL60 bulk cultures yielded fewer
differentiated progeny if immediately cultured with D 3
than if they encountered D 3 12 h later (about halfway
through the same cycle).
Control of Proliferation and of the Expression of
Mature Characters Are Not Coupled
Until recently, a frequent assumption about differentiating promyeloid cells was that differentiation
markers such as CD11b are only expressed as and/or
after cells drop out of cycle. Recent work, which showed
that CD11b expression starts in D 3-treated HL60 cells
before division ceases, has shown this to be incorrect.
For example, HL60 cells can simultaneously incorporate bromodeoxyuridine into DNA and express PCNA
and CD11b [7]; and some variant HL60 cell sublines
continue to proliferate when D 3 treated but still show
an induced expression of CD14 and other markers of
differentiation [8].
Our results take this conclusion much further. They
demonstrate that whether or when CD11b expression
is triggered is influenced little by when in the cycle D 3
is added— or even whether cells are in cycle at all.
HL60 cells that are isolated by elutriation in early G1
and returned to culture with D 3 typically undergo
three maturation divisions, but they start to express
CD11b long before they complete the first cell cycle.
Moreover, the cells are sensitive to initiation of CD11b
expression through much of the cell cycle, but D 3 and
ATRA can only instruct cells to embark on their maturation divisions during a brief “window of sensitivity”
early in G1. This means that cells which encounter a
differentiating agent after closure of this early G1 window of sensitivity initiate the expression of genes characteristic of their differentiated fate even before they
embark on the maturation divisions that precede cell
cycle exit. At least in part, therefore, CD11b expression
and the initiation of maturation divisions must be controlled by independent processes downstream of the
receptors for the differentiating agents—and the signals that control the maturation divisions can only
function productively during early G1. These regulatory processes are yet to be identified.
However, there was a modest difference between the
speed of CD11b expression in early G1 cells (Fraction
1) and the rates of CD1b appearance in the other
populations (unsynchronized and Fractions 2 and 3).
The former became CD11b positive more quickly— during 10 –12 h rather than ⬃15 h (Fig. 6B). The most
obvious interpretation is that there is a brief phase
during the cell cycle (most likely G2/M) during which
cells are transiently refractory to the induction of
CD11b expression, and Fraction 1 contains few cells in
this phase.
Recent gene array analyses dramatically demonstrate how quickly differentiating agents start to reprogram the gene expression patterns of myeloid cells.
CELL PROLIFERATION, CD11B EXPRESSION, AND HL60 CELL DIFFERENTIATION
In two studies, of phorbol ester-induced monocyte differentiation of HL60 cells [12] and of ATRA-induced
neutrophilic differentiation of NB 4 cells [11], substantial numbers of genes related to mature myeloid cell
function were activated within 1 day—well before the
cells drop out of cycle. These include the genes for
interleukin-8, ICAM1, the G-CSF receptor and CD11a,
as well as some less predictable proteins (e.g., VEGF in
both conditions). A few genes—such as that for macrophage inhibitory protein-␣ (MIP1␣)—are switched on
in 4 – 8 h. In those studies, CD11b mRNA was detected
at times consonant with our immunological detection
of newly synthesized CD11b after ⬃8 –12 h: at 24 (but
not at 4) h in the phorbol ester study and at 12 (but not
8) h in the ATRA microarrays.
What Is the Relationship between Proliferation and
Differentiation?
In recent years, several studies of differentiating
myeloid cells have noted the early activation of processes that are more usually associated with cell proliferation. These include increased Cdk5 activity [13];
up-regulation of the JNK and ERK2 MAP kinases (and
down-regulation of p38) [14]; up-regulation of cyclins
A, D1, and E [5]; phosphorylation of histone H3 [15];
and activation of protein kinase C␨ [16]. There is also a
rapid activation of phosphoinositide 3-kinase—which
generates either a survival or a proliferative signal in
most cells, but which seems in this context to be needed
for successful monocytic maturation [16]. Unexpectedly, this phosphoinositide 3-kinase activation seems
to be driven by a direct interaction between phosphoinositide 3-kinase and the ligated vitamin D 3 receptor
(VDR)— even though the relevant phosphoinositide
3-kinase is a receptor-coupled plasma membrane enzyme and the VDR is a hormone-regulated nuclear
transcription factor [17]. It is not clear how this activation occurs, but its occurrence is made more provocative by another report of direct phosphoinositide 3-kinase activation by a ligated nuclear receptor (estrogen
receptor-␣) [18].
Now that it is established that differentiating myeloid cells simultaneously proliferate and start expressing differentiated characteristics, the evocation of
“proliferation-related” events by differentiating agents
no longer seems odd. Conversely, however, it becomes
puzzling that multiple CDK inhibitors (p21, p27, p15,
and p18) are fairly quickly up-regulated in response to
differentiating agents, even while the cells continue to
multiply [e.g., 18, 19].
Like many others, we initially expected that the
processes which lead to growth arrest and to cell maturation might be obligatorily linked, but the accumulating evidence— both published [e.g., 5– 8] and newly
reported here– does not support this notion. However,
133
it remains possible that there might be regulatory
events, independent of those that directly initiate maturation divisions or directly activate CD11b expression, which still influence both the number of maturation divisions and the rate of maturation. Noble’s group
recently reported the possible identification of such a
mechanism in neuroglial progenitor cells [21]. Their
conclusion was that those cells that have a relatively
reduced overall redox balance tend to proliferate substantially before maturing, whereas cells whose redox
poise veers toward a more oxidized state soon differentiate and drop out of cycle. Moreover, hormones and
growth factors that promoted the differentiation or
proliferation of these neuroglial cells perturbed their
redox balance in the ways predicted by their effects on
cell behavior. Pharmacological redox manipulations
also had the effects predicted by this model [21]. The
redox-sensing molecule that serves as the fulcrum of
this proliferation/differentiation balance is yet to be
identified.
Some time ago, it was shown that low mitochondrial
oxidative activity (assayed with rhodamine-123) in
Sca-1 ⫹ progenitor cells is predictive of effective longterm murine hemopoietic repopulation [22]. Other
forerunners of this redox control model were the beneficial effects of culture in a reduced oxygen tension or
of the inclusion of ␤-mercaptoethanol (or other cellpermeant reduced thiol; e.g., GSH) on the cloning efficiencies of hemopoietic and related progenitor cells
[23–28]. These observations imply that redox balance
also regulates the balance between growth and differentiation in myeloid cells, but this remains unproven.
Now that it is clear that cell proliferation and the
expression of differentiated characteristics can occur
concurrently in differentiating myeloid cells, but are
independent processes, it should be possible to analyze
systematically how modulation of the cellular redox
status influences each of these processes.
This work was supported by the Leukaemia Research Fund as part
of LRF Specialist Programme Grant 93/96 (to GB) and by the Royal
Society (to R.H.M.).
REFERENCES
1.
Collins, S. J., Ruscetti, F. W., Gallagher, R. E., and Gallo, R. C.
(1978). Terminal differentiation of human promyelocytic leukemia cells induced by dimethyl sulfoxide and other polar compounds. Proc. Natl. Acad. Sci. USA 75, 2458 –2462.
2.
Breitman, T. R., Selonick, S. E., and Collins, S. J. (1980). Induction of differentiation of the human promyelocytic leukemia
cell line (HL-60) by retinoic acid. Proc. Natl. Acad. Sci. USA 77,
2936 –2940.
3.
Fischkoff, S. A., Pollak, A., Gleich, G. J., Testa, J. R., Misawa,
S., and Reber, T. J. (1984). Eosinophilic differentiation of the
human promyelocytic cell leukemic cell line HL-60. J. Exp. Med.
160, 179 –196.
134
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
DRAYSON ET AL.
Hutt-Taylor, S. R., Harnish, D., Richardson, M., Ishizaka, T.,
and Denburg, J. A. (1984). Sodium butyrate and a T lymphocyte
cell line-derived differentiation factor induce basophilic differentiation of the human promyelocytic leukemia cell line HL-60.
Blood 71, 209 –215.
Rots N. Y., Iavarone A., Bromleigh V., and Freedman L. P.
(1999). Induced differentiation of U937 cells by 1,25-dihydroxyvitamin D3 involves cell cycle arrest in G1 that is preceded by a transient proliferative burst and an increase in
cyclin expression. Blood 93, 2721–2729.
Brown, G., Choudhry, M. A., Durham, J., Drayson, M. T., and
Michell, R. H. (1999). Monocytically differentiating HL60 cells
proliferate rapidly before they mature. Exp. Cell Res. 253, 511–
518.
Liu, Q., Van Hoy, R. W., Zhou, J. H., Dantzer, R., Freund, G. G.,
and Kelley, K. W. (1999). Elevated cyclin E levels, inactive
retinoblastoma protein, and suppression of the p27 (KIP1) inhibitor characterize early development of promyeloid cells into
macrophages. Mol. Cell. Biol. 19, 6229 – 6239.
Studzinski, G. P., Rathod, B., Wang, Q. M., Rao, J., and Zhang,
F. (1997). Uncoupling of cell cycle arrest from the expression of
monocytic differentiation markers in HL60 cell variants. Exp.
Cell Res. 232, 376 –387.
Wallington, L. A., Durham, J., Bunce, C. M., and Brown, G.
(1996). Growth of single HL60 cells in liquid culture: Analysis of
the influences of differentiative agents. Leukaemia Res. 20,
821– 829.
Studzinski, G. P., Bhandal, A. K., and Brelvi, Z. S. (1985). Cell
cycle sensitivity of HL-60 cells to the differentiation-inducing
effects of 1-alpha,25-dihydroxyvitamin D 3. Cancer Res. 45,
3989 –3905.
Tamayo, P., Slonim, D., Mesirov, J., Zhu, Q., Kitareewan, S.,
Dmitrovsky, E., Lander, E. S., and Golub T. R. (1999). Interpreting patterns of gene expression with self-organizing maps:
Methods and application to hematopoietic differentiation. Proc.
Natl. Acad. Sci. USA 96, 2907–2912
Liu, T. X., Zhang, J. W., Tao J., Zhang, R. B., Zhang Q. H., Zhao,
C. J., Tong, J. H., Lanotte, M., Waxman, S., Chen, S. J., Mao,
M., Hu, G. X., Zhu, L., and Chen, Z. (2000). Gene expression
networks underlying retinoic acid-induced differentiation of
acute promyelocytic leukemia cells. Blood 96, 1496 –504
Chen, F., and Studzinski, G. P. (1999). Cdk5 activity enhances
monocytic phenotype and cell cycle traverse in 1,25-dihydrocyvitamin D 3-treated HL60 cells. Exp. Cell Res. 249, 422– 428.
Wang, X., Rao, J., and Studzinski, G. P. (2000). Inhibition of
p38 MAP kinase up-regulates multiple MAP kinase pathways
and potentiates 1,25-dihydroxyvitamin D 3-induced differentiation of human leukemia HL60 cells. Exp. Cell Res. 258, 425–
437.
Juan, G., Traganos, F., and Darzynkiewitz, Z. (1999). Histone
H3 phosphorylation in human monocytes and during HL60 cell
differentiation. Exp. Cell Res. 246, 212–220.
Received November 2, 2000
Revised version received January 19, 2001
Published online April 6, 2001
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Liu, Q., Ning, W., Dantzer, R., Freund, G. G., and Kelley, K. W.
(1998). Activation of protein kinase C-␨ and phosphatidylinositol 3-kinase and promotion of macrophage differentiation by
insulin-like growth factor-I. J. Immunol. 160, 1393–1401.
Hmama, Z., Nandan, D., Sly, L., Knutson, K. L., Herrera-Velit,
P., and Reiner, N. E. (1999). 1,25-Dihydroxyvitamin D 3-induced
myeloid cell differentiation is regulated by a vitamin D receptor-phosphatidylinositol 3-kinase signaling complex. J. Exp.
Med. 190, 1583–1594.
Simoncini, T., Hafezi-Moghadam, A., Brazil, D. P., Ley, K.,
Chin, W. W., and Liao, J. K. (2000). Interaction of oestrogen
receptor with the regulatory subunit of phosphatidylinositol3-OH kinase. Nature 407, 538 –541.
Liu, M., Lee, M-H., Cohen, M., Bommakanti, M., and Freedman, L. P. (1996). Transcriptional activation of the Cdk inhibitor p21 by vitamin D 3 leads to the induced differentiation of the
myelomonocytic cell line U937. Genes Dev. 10, 142–153.
Muto, A., Kizaki, M., Yamato, K., Kawai, Y., Kamata-Matsushita, M., Ohguchi, H. M., Nishihara, T., Koeffler, H. P., and
Ikeda, Y. (1999). 1,25-Dihydroxyvitamin D 3 induces differentiation of a retinoic acid-resistant acute promyelocytic leukemia
cell line (UF-1) associated with expression of p21 WAF1/CIP1 and
p27 KIP1. Blood 93, 2225–2233.
Smith, J., Ladi, E., Mayer-Proschel, E. M., and Noble, M.
(2000). Redox state is a central modulator of the balance between self-renewal and differentiation in a dividing glial precursor cell line. Proc. Natl. Acad. Sci. USA 97, 10032–10037.
Kim, M. J., Cooper, D. D., Hayes, S. F., and Spangrude, G. J.
(1998). Rhodamine-123 staining in haemopoietic stem cells of
young mice indicates mitochondrial activation rather than dye
efflux. Blood 91, 4106 – 4117.
Iscove, N. N., and Sieber, F. (1975). Erythroid progenitors in
mouse bone marrow detected by macroscopic colony formation
in culture. Exp. Hematol. 3, 32– 43.
Helgestad, J., Storm-Mathisen, I., and Lie, S. O. (1986). Vitamin C and thiol reagents promote the in vitro growth of murine
granulocyte/macrophage progenitor cells by neutralizing endogenous inhibitor(s). Blut 52, 1– 8.
Katahira, J., and Mizoguchi, H. (1987). Improvement of culture
conditions for human megakaryocytic and pluripotent progenitor cells by low oxygen tension. Int. J. Cell Cloning 5, 412– 420.
Levy, E. M., Wu, J., Salibian, M., and Black, P. H. (1992). The
effect of changes in thiol subcompartments on T-cell colony
formation and cell cycle progression: Relevance to AIDS. Cell.
Immunol. 140, 370 –380.
Ono, K., and Alter, B. P. (1995). Effects of low oxygen tension
and antioxidants on human erythropoiesis in vitro. Exp. Hematol. 23, 1372–1377.
Inui, K., Oreffo, R. O., and Triffitt, J. T. (1997). Effects of
␤-mercaptoethanol on the proliferation and differentiation of
human osteoprogenitor cells. Cell. Biol. Int. 21, 419 – 425.