Interleukin 3 Protects Murine Bone Marrow Cells
from Apoptosis Induced by DNA Damaging Agents
By Mary K. L. Collins, Jacqueline Marvel, Prupti Malde,
and Abelardo Lopez-Kivas*
From the Chester Beatty Laboratories, Institute of Cancer Research, London $W3 6J1~ United
Kingdom; and the *Instituto Lopez Neyra de Parasitologia, 18001 Granada, Spain
Murine bone marrow-derived cells, dependent on interleukin 3 (I1.3) for their growth in culture,
undergo programmed cell, or apoptosis, upon cytokine withdrawal. Here it is reported that a
variety of DNA damaging agents cause a more rapid onset of apoptosis in a factor-dependent
cell line, BAF3, deprived of I1.-3. In contrast, when cultured in the presence of I1.3, or other
growth promoting factors, BAF3 cells are highly resistant to X-irradiation and the cytotoxic
drugs etoposide and cisplatin. Overexpression of the bcl2 gene product also protects BAF3 cells
from DNA damage. The presence of I1-3 is not required during the initial events of DNA damage
or its repair. In the absence of I1-3, cells still complete the repair of DNA breaks within 15
min, and continue to cycle for 5 h. At this time, I1-3 is necessary to prevent the accelerated
onset of DNA cleavage from a G2 arrest point.
toffs, as well as the rate of cell proliferation. For example,
those thymocytes that are triggered by antigen within the
thymus die by apoptosis (1), which provides a mechanism
for the elimination of self-reactive T lymphocytes. In con-
trast, B lymphocytes from the germinal centers of lymph nodes
apoptose if they are not antigen triggered (2), allowing the
selection of cells producing high affinity antibodies. Expan-
sion and maturation of the immature progenitor cells of bone
marrow also appear to be controlled by a balance between
proliferation and apoptosis. The proliferation of such cells
is stimulated by cytokines, including I1.3 and GM-CSF, which
act on the earliest precursor cells (3, 4). Upon withdrawal
of II.-3 or GM-CSF from progenitor cell lines (5, 6) or pri-
mary I1.3-dependent cells from bone marrow (6), a program
of events typical of apoptosis is observed, which includes the
digestion of cellular chromatin to oligonucleosome-length
fragments by an endogenous nuclease. This provides a model
for the regulation of haematopoiesis in mouse bone marrow
where the least mature progenitor cells are not cycling (7),
but after the stimulation of haematopoiesis, activated pro-
genitor cells proliferating in response to cytokines could be
induced to apoptose when cytokines are no longer available,
rapidly terminating their expansion.
The mechanism by which cytokines such as II-3 protect
cytokine-dependent cells from the onset of apoptosis remains
unclear. The initial interaction of II.-3 with its receptor stimu-
lates rapid events such as the tyrosine phosphorylation of cel-
lular proteins (8). However, such signals decay rapidly upon
he development of haematopoietic cells can be controlled
by regulating the rate of programmed cell death, or apop-
I1.3 removal (8), whereas the initiation of DNA deavage does
not occur until 6 h after I1.3 deprivation (6). The protective
action of I1.3 can be mimicked either by other growth
promoting factors that interact with homologous receptors,
such as GM-CSF (5) and I1.4 (9), or by the overexpression
of the product of the bcl2 oncogene (10, 11). However, un-
like cytokines, bcl2 inhibits the onset of apoptosis but does
not cause cells to proliferate.
In addition to its role in the control of haematopoietic cell
development and immune repertoire selection, the induction
of apoptosis is involved in the mechanism of killing of tumor
cells by a number of therapeutic regimes. A variety of agents
used for some time in human tumor therapies, such as antag-
onists of steroid hormones (12), glucocorticoids (13), DNA
damaging drugs (14, 15), or radiation (16) have recently been
demonstrated to kill cells by inducing apoptosis. Again, the
mechanism by which initial DNA damage leads to subse-
quent cleavage of DNA remains unclear. Among normal cells,
those of the haematopoietic system are the most sensitive when
exposed in vivo to DNA damaging agents. Because I1.3 will
protect bone marrow-derived haemopoietic cells from apop-
tosis, we investigated whether cytokines will protect against
apoptosis induced by DNA damaging agents. We show that
I1.-3, other growth promoting factors, and the bcl2 protein
inhibit the rapid apoptosis induced by DNA damaging agents
in murine bone marrow-derived I1.3-dependent cells. The
extent of DNA damage and its repair were not affected by
II.-3. The cytokine acted 6 h after these initial events to allow
cells to pass a restriction point at which DNA damage is
J. Exp. Med.y The Rockefeller University Press i 0022-1007/92/10/1043/09 $2.00
Volume 176 October 1992 1043-1051
Materials and Methods
Cell Culture and Reagents. BAF3 cells (17; obtained from Dr.
Konald Palacios, Basel Institute for Immunology, Basel, Switzer-
land) were maintained in DMEM containing 10% fetal bovine serum
(FBS) 1 and 5% WEHI 3B cell-conditioned medium which was
used throughout as a source of IL-3, at a density of 5 x 104-5
x 10 s cells/ml. Primary IL-3-dependent cultures of murine bone
marrow ceUs were prepared as previously described (6). They were
characterized largely as mast cells by the release of ~-glucosaminidase
upon calcium ionophore stimulation. To remove Ib3, calls were
washed twice in DMEM/10% FBS. Cell viability and viable cell
density were determined by counting cells able to exchde trypan
blue on a haemocytometer. Duplicate samples of at least 50 ceLls
were analyzed. Murine IL-4 was obtained from Genzyme Corp.
(Boston, MA), insulin-like growth factor (IGF-1) from Bachem
(Bubbendorf, Switzerland), A23187 from Calbiochem Corp.
(Boston, MA), etoposide from Bristol-Myers (Slough, UK),
and cisplatin from David Bull Laboratories (Warwick, UK).
X-irradiation was performed using a Pantac x-ray machine (Chester
Beatty Laboratories), with an output of 24 0 keV at a dose rate
of 250-300 rad/min.
DNA Fragmentation. 107 cells were lysed in 0.75 ml 10 mM
Tris (pH 8.0), 150 mM NaC1, 0.1 mM EDTA, 1% SDS, and in-
cubated for 15 h with 200 #g/ml proteinase K at 37~ After ex-
traction with 50% phenol/50% chloroform, total DNA was
precipitated by addition of 1/10 vol 3 M NaAcetate (pH 5.0) and
2.5 vol ethanol and incubation for 15 h at -20~
was resuspended in 50 #1 10 mM Tris (pH 8.0), 0.1 mM EDTA,
and the nucleic acid concentration determined by measurement of
ODz60. 15 #g of nucleic acid was incubated for 1 h at 37~ with
10 U/ml DNAse-fi~ RNAse, then electrophoresed on a 1% agarose
gel in the presence of ethidium bromide. A 1-kb DNA ladder size
marker (Gibco-BRL, Gaithersburg, MD) was included in the last
lane of each gel.
Detection of Single-stranded DNA Breaks. The differential rates
of elution of test sample DNA and an internal standard DNA, at
alkaline pH, were used to detect the presence of single-stranded
DNA breaks in the test samples. DNA was labeled by culture of
test cells for 15 h with 0.5 #Ci/ml [uC]thymidine (Amersham In-
ternational, Amersham, Bucks, UK) and culture of standard cells
with 0.5 #Ci/ml [3H]thymidine (Amersham International). Im-
mediately before irradiation, test cells were washed twice in
DMEM/10% FBS, suspended in DMEM/10% FBS containing 50
mM Hepes pH 7.0, placed on ice, and irradiated. A nonirradiated
and an irradiated control test cell sample were maintained on ice
and other test samples were incubated at 37~ with Ib3 if indi-
cated, for the time shown. An equivalent number (10 ~) of each
test sample and standard cells were then admixed on ice, pelleted
at 4~ resuspended in PBS, and loaded onto polycarbonate filters
at 4~ The rate ofelution of test and standard DNA in each sample
by 2% tetrapropylammoniumhydroxide (pH 12.1) was then mea-
sured, after cell lysis and protease digestion, as previously de-
Cell Cycle Analysis. 5 x 106 cells were pelleted, resuspended
in 0.2 ml PBS, and fixed by the addition of 2 ml of ice-cold 75%
ethanol/25% PBS. Fixed cells were pelleted, vigorously resuspended
in PBS, and incubated for 30 min at 37~ with 100 #g/ml RNase
and 40 #g/ml propidium iodide. The fluorescence of the stained
The DNA pellet
1 Abbreviations used in this paper: FBS, fetal bovine scum; IGF-1, insulin-
like growth factor 1.
cells was analyzed using a FACScan | (Becton Dickinson & Co.,
Mountain View, CA).
Generation ofbd2-expressing Cells. A 1.9-kb human bc12 eDNA
fragment was inserted in the EcoRI restriction site of the retroviral
pM5Gneo (from C. Stocking, Heinrich-Petter Institut, Hamburg,
Germany). After transfection of this vector into the packaging cell
line ~2 (19) and selection in G418 a producer cell clone, which
transferred G418, resistance to National Institutes of Health (NIH)
3T3 cells at a frequency of 104 resistant colonies/ml producer cell
supernatant, was isolated. BAF3 cells were infected by cocultiva-
tion with this producer cell clone, and a cell population which was
100% positive for the expression of human bcl2, assessed by fluores-
cence activated cell analysis using an anti-human bcl2 antibody (20),
was used for further experiments.
IL3 Inhibits Apoptosis Induced by DNA Damaging Agents
in a Murine Bone Marrow-derived Cell Line and in Primary Cul-
tures of Murine Bone Marry.
derived cell line BAF3 has been shown to undergo apoptosis,
characterized by an initial deavage of cellular chromatin and
a subsequent loss of plasma membrane integrity, upon with-
drawal of Ib3 (6). Fig. 1 A demonstrates that such IL-
3-deprived BAF3 cells began to lose viability 13 h after factor
removal, and death was asynchronous within the cell popu-
lation and complete only after 28 h. When BAF3 calls were
exposed to X-irradiation (400 tad) immediately after II,3 with-
drawal, this loss of viability was more rapid and synchronous,
being first observed after 8 h and essentially complete after
18 h (Fig. 1 A). In the presence of II--3, BAF3 cells were
completely resistant to this dose of radiation for 36 h (Fig.
1 A). Some loss of cell viability was observed over the fol-
lowing 36 h, and is shown in Fig. 1 C. During these first
72 h after X-irradiation, the viable cells showed little
[3H]thymidine incorporation (data not shown) and failed to
increase in number (Fig. 1 C). However, exponential growth
was observed after this initial lag period (Fig. 1 C). The viable
calls in the culture were therefore able to divide normally.
Indeed, 72 h after X-irradiation, such cells showed a cloning
efficiency assessed by limiting dilution of 83%, compared with
87% of control cells. We have previously shown that pri-
mary murine bone marrow cells, cultured in the presence of
II,3, also apoptose upon cytokine removal (6). Such primary
IL-3-dependent cultures, which were largely mast cells, also
lost viability rapidly upon exposure to 400 tad of X-irradiation
and could be protected from death by the presence of Ib3
(Fig. 1 D).
The main effect of X-irradiation at a dose of 400 rad is
the induction of single-stranded DNA breaks (18). There-
fore, to investigate whether DNA damage was responsible
for the rapid induction of death in BAF3 cells, other agents
known to induce DNA damage were used to treat BAF3 ceUs
deprived of Ib3. Fig. 1 B demonstrates that etoposide, a
topisomerase II inhibitor that also induces both single-stranded
DNA and double-stranded breaks (21), and cisplatin, which
causes DNA strand crosslinks (22), both induced cell death
with similar kinetics to that observed after X-irradiation. ID3
inhibited the action of these agents for the initial 24 h of
culture (Fig. 1 B). At this time, removal of the drugs by
The murine bone marrow-
1044 Intcrleukin 3 Protection of Cells from DNA Damage
4C~ rr~: §
\ \ _
10 2'0 3'0
----0--- viable cells
0 100 2O0
to 400 tad X-irradiation as indicated, then incubated, in the absence or presence of 1I.-3 as indicated, for the time shown. Cell viability was then deter-
mined. (13) BAF3 cells were incubated for the time shown with II.,3, 1/zg/ml etoposide and 4/~g/ml cisplatin as indicated, after which cell viability
was determined. (C) 2 x 10 s BAF3 cells were exposed to 400 tad X-irradiation, then incubated in the presence of IL-3 for the time shown, when
the number and percentage of viable cells was determined. Cells were maintained at a density of <5 x 10 s cells/ml. (D) Primary cultures of Ib3-de-
pendent mast cells from murine bone marrow were treated as in A.
10 2O 3'0
' '0 ' '
5 10 2 30 40
The induction of apoptosis by DNA damaging agents and its inhibition by IL-3. (A) BAF3 cells were washed to remove Ib3, exposed
washing resulted in the recovery of cells that remained viable
during subsequent culture.
The early loss of viability, observed after treatment of BAF3
cells with various DNA damaging agents, suggested that these
agents might trigger the apoptotic pathway more rapidly than
II.-3 removal alone. To investigate this, the time of initiation
of DNA deavage in X-irradiated, and etoposide- or cisplatin-
treated BAF3 cells was observed. Fig. 2 A shows that
oligonucleosome-length DNA fragments, characteristic of
cleavage of cellular chromatin by an endogenous nuclease (13),
could be detected 4 h after treatment of IL-3-deprived cells
with these agents. In control cultures deprived of II.-3, such
fragmentation was not observed until 8 h (reference 6 and
1045 Collins et al.
Fig. 2 A). DNA cleavage was not detected in cultures treated
with DNA damaging agents in the presence of II.-3 (Fig.
2 A). Thus, initial endogenous nuclease digestion of chro-
matin precedes initial loss of cell viability by 4-5 h in both
DNA-damaged and control Ib3-deprived cells. Unlike this
fixed time period after which cell death occurs once cleavage
is initiated, it is clear that the time of onset of cleavage can
be greatly shortened by DNA damaging agents.
The ability of IL3 to protect BAF3 cells from X-irradiation
or cytotoxic drugs could be demonstrated at very high doses
of the DNA damaging agents. In the absence of IL-3, 100
rad of X-irradiation, 0.5 ~g/ml cisphtin or 0.05 #g/ml etopo-
side were sufficient to induce the rapid onset of apoptosis
Figure 2. Effect of DNA damaging agents on DNA fragmentation. (Uppert, anel ) BAF3 cells were X-irradiated (A and E), treated with 1/~g/ml
etoposide (B and F) or 1/tg/ml cisplatin (C and G), then incubated in the absence (A-C) or presence (E-G) of I1.,3 for 2 h (lanes I), 4 h (lanes
2), 6 h (lanes 3), 8 h (lanes 4), 12 h (lanes 5), and 23 h (lanes 6). (D) DNA prepared from cells at the same time points, after II.-3 removal in the
absence of any DNA damaging treatment. Fragmentation of cellular DNA was then analyzed. (Lower panel) BAF3 cells were incubated for 15 h in
the presence of I1.,3, either after X-irradiation: (A) lanes I. 200 tad; 2. 400 rad; 3. 800 tad; 4. 1,600 tad; 5. 3,200 tad); or in the presence of etoposide:
(B) lanes I. 0.04/~g/ml; 2. 0.1/zg/ml; 3. 0.4/~g/ml; 4. 1/~g/ml; 5. 4/~g/ml); with no addition (C) or in the presence of cisplatin: (D) Lanes l.
1 ttg/ml; 2. 2 #g/ml; 3. 4/~g/ml; 4. 8 #g/ml; and 5. 16/~g/ml). Fragmentation of cellular DNA was then analyzed.
(data not shown). In the presence of IL-3, the calls were resis-
tant to 1,600 tad of X-irradiation. 8 #g/ml cisplatin, or 0.05
/zg/ml 1/zg/ml etoposide (Fig. 2 B) which represent higher
levels than are normally tolerated by eucaryotic cells (14, 21).
Other Growth.promoting Factors and the bcl2 0ncogene Product
also Protect Cells from X-irradiation. Protection of these cells
was not confined to IL-3. We have previously demonstrated
that the growth-promoting factors IGF-1 and II.-4 can pro-
tect BAF3 cells from apoptosis when II.,3 is removed (6, 23).
Fig. 3 demonstrates that either IGF-1 or IL-4 was effective
in protecting the cells from X-irradiation. We have also demon-
strated that the calcium ionophore A23187 can inhibit onset
of apoptosis in the absence of IL,3 (6). Unlike IGF-1 or II.,4,
A23187 does not stimulate proliferation of BAF3 ceils. When
added alone, the ionophore maintains ceUs viable but not cy-
cling, and when added in the presence of IL-3, it arrests their
cell cycle (6). The observation that A23187 alone or in com-
bination with 11,3 could protect BAF3 cells from X-irradiation
(Fig. 3 A) therefore suggested that cells do not need to cycle
to recover from its effect. Because of the difficuhy of main-
taining cells for more than 48 h with calcium ionophores,
this conclusion was best demonstrated when the effect of over-
expression of the bcl2 protein on the sensitivity of BAF3 cells
to radiation was observed. Fig. 3 B shows the behavior of
a population of BAF3 cells, infected with a recombinant
retrovirus encoding human bcl2 protein, when deprived of
Ib3 and exposed to X-irradiation. The cells expressing human
bc12 protein showed a normal cell cycle distribution when
grown in IL-3 (Fig. 3 B, pand a) but arrested with diploid
DNA content 48 h after II,3 was removed (Fig. 3 B, panel
1046 Interleukin 3 Protection of Cells from DNA Damage
IL3 tGF 1 IL4
G G 2
i 400 rads
r = IIII
10 20 30 40 50
Figure 3. Inhibition of the effect of X-irradiation by a variety of agents.
(.4) BAF3 cells were exposed to 400 rad X-irradiation, then incubated for
15 h with IL-3, IGF-1 (1/~g/ml), II.-4 (10 U/ml), and A23187 (1/~M)
as indicated, after which cell viability was determined. (/3) BAF3 cells over-
expressing human bd2 were deprived of IL-3 for 15 h. Cell cycle analysis
at this point (b) shows that the cells are arrested in the G1 phase of the
cell cycle, compared with the same cells cultured in Ib3 (a). (C) The ar-
rested cells were exposed to 400 tad of X-irradiation, and the viability
of irradiated and control arrested human bcl2 cells, and irradiated and con-
trol parental BAF3 cells, was determined at the times indicated.
b). Such arrested cells lost viability after "~100 h total IL-3
deprivation (Fig. 3 C), and their rate of death was not ac-
celerated by X-irradiation (Fig. 3 C).
II~3 Does Not Affect the Extent of DNA Damage or the Rate
of Its Repair. To investigate the mechanism by which IL-3
protected BAF3 cells from X-irradiation, the extent of repair
of DNA damage was measured in cells X-irradiated at 4~
in the absence of I1-3, then incubated in the absence or pres-
ence of IL3 at 37~ Single-stranded DNA breaks after a
dose of 400 tad were detected by alkaline filter elution (18).
No double-stranded DNA breaks could be detected by neu-
tral filter elution (24). Fig. 4 A demonstrates that the initial
single-stranded DNA breaks were undetectable after a 1-h
incubation in either the absence or presence of I1"3. This sug-
gested that DNA damage did not lead directly to the onset
of DNA cleavage in cells deprived of I1"3. However, the sen-
sitivity of this measurement of single-stranded DNA breaks
did not exclude the possibility that a small number of breaks
remained unrepaired in the cells deprived of I1"3. The rate
of DNA repair was therefore measured in cells deprived of
I1-3 for 6 h. After 5 rain, the extent of DNA repair was iden-
tical in 11-3-deprived (Fig. 4 C) and control cells (Fig. 4 B)
1047 Collins et al.
and no breaks could be detected after 15' in either culture
(data not shown). Thus, as the rate of repair was not im-
paired in cells deprived of II-3 for 6 h, a failure to repair a
small number of breaks seemed unlikely to be responsible
for the damage sensitivity of I1"3-deprived cells.
11,3 Allows Cells To Pass a G2 Restriction Point at which DNA
Repair Is Monitored. Because I1"3 did not affect the initial
repair of DNA damage, it was then determined at which
point during induction of the apoptotic process the presence
of I1"3 was necessary. Fig. 5 shows that I1"3 could be re-
moved from control BAF3 cells for 9 h, then readded with
no subsequent loss in cell viability. Further experiments
demonstrated that this period of I1"3 deprivation could be
extended to 12 h (data not shown). X-irradiated cells could
be deprived of I1-3 for the first 2 h after irradiation with no
loss in viability upon I1-3 readdition. Even at 6 h after irradi-
ation, 70% of the cells could be rescued by I1-3 (Fig. 5).
These data demonstrate that the first irreversible event in the
onset of apoptosis occurred close to the time of initiation
of DNA cleavage in both control and X-irradiated ceils.
As the point at which II.,3 protected cells from X-irradiation
was several hours after the initial damage and repair, the cy-
I i i i i i i
Figure 4. Measurement of tingle-stranded DNA breaks and their re-
pair. (A) Comparative rates of elution of test (14C-labeled) and internal
standard (~H-labeled) DNA. Test samples were prepared from BAF3 cells
that were washed to remove I1,3, then analyzed (D), given 400 rad
X-irradiation, and analyzed without further treatment (O), or irradiated
and then incubated for 1 h at 37~ in the absence (&) or presence (m)
of 11,3. (B) Test samples were prepared from control BAF3 cells (V1), cells
given 400 tad X-irradiation in the presence of II.,3 (O), and irradiated
cells subsequently incubated for 5' at 37~ in the presence of I1,3 (I).
(C) Test samples were prepared from BAF3 cells deprived of IL-3 for 6 h
([3), these cells were given 400 tad X-irradiation in the absence of 11,3
(0), and these irradiated cells subsequently incubated for 5' at 37~ in
the absence of I1,3 (I).
cling of the cells during this period in the absence or pres-
ence of IL-3 was observed by cytofluorimetry of propidium
iodide-stained cells. The presence of a subdiploid DNA con-
tent population, with light scatter characteristics identical
IL3 deprivation (h)
Figure 5. Determination of the time at which I1,3 is required to pre-
vent cell death. BAF3 cells were washed to remove I1,3 and irradiated with
400 tad X-rays as indicated. They were then incubated at 37~ for the
time shown in the absence of I1,3. I1,3 was then readded and cell viability
was determined 24 h after the initial I1,3 deprivation.
to the rest of the cells, can be used as a measure of the number
of cells undergoing apoptosis (6). In control cells deprived
of I1,3, such a population was <5% of the total 9 h after
factor removal. At this time the cells were still distributed
throughout the cell cycle very much as were cells cultured
in II-3 (Fig. 6). Thus removal of I1,3 did not result in any
rapid cessation of cell cycle progression in these cells.
X-irradiation of the cells resulted in an accumulation in the
G2 phase of the cell cycle, which was similar in the absence
or presence of IL-3 (Fig. 6). This is characteristic behavior
of eucaryotic cells exposed to DNA-damaging agents (25).
However, in the absence of Ib3, apoptotic cells began to ap-
pear in the culture after 6 h. This appearance of apoptotic
-r L , ,
G1 G 2
G2 G1 G 2 1 G
Figure 6. Cell cycle analysis of X-irradiated cells. The cell cycle distri-
bution of BAF3 cells cultured in the presence of I1,3 (A), in the absence
of I1,3 (B), X-irradiated with 400 rad, then cultured in the presence of
I1,3 (C), or X-irradiated with 400 tad, then cultured in the absence of
IIr (D), for the time shown, was determined.
1048 Interleukin 3 Protection of Cells from DNA Damage
Figure 7. Effect of cydoheximide and
cell cycle inhibitors on radiation-induced
apoptosis. (A) The cell cycle distribu-
tion of BAF3 cells exposed to 400 rad
X-irradiation, incubated for t h at 37°C,
then cultured for a further 5 h in the
absence oflb3 (panel I), with 10 #g/ml
cycloheximide (panel 2), 5 #g/ml aphi-
dicholine (panel 3), or 500 ng/ml col-
cemid (pand 4). (B) DNA fragmenta-
tion after 6 h of the cells of A. Lane
I, no addition; lane 2, aphidicholine;
lane 3, cycloheximide; and lane 4, col-
cemid. The concentration of cyclohex-
imide used inhibited protein synthesis,
measured by incorporation of [3sS]me-
thionine during a 1-h period 5 h after
irradiation, by 91% and that of aphidi-
choline inhibited protein synthesis by
40%. In contrast, cycloheximide in-
hibited [3H]thymidine incorporation
during the 6 h after irradiation by 26%
and aphidicholine by 96%.
cells was accompanied by a loss of ceils from the G2 arrest
point (Fig. 6 B).
To investigate whether a checkpoint associated with G2
arrest was directly responsible for the onset of apoptosis, we
used agents that prevented this G2 arrest. Fig. 7 A demon-
strates that both cycloheximide, an inhibitor of protein syn-
thesis, and aphidicholine, which inhibits DNA polymerase
c~ and therefore blocks DNA replication, prevented the ac-
cumulation of cells in G2 after X-irradiation in the absence
of IL-3. Significantly, both of these agents were able to pre-
vent the rapid onset of DNA deavage observed in control
cells that had accumulated in G2 (Fig. 7 B), whereas cob
cemid, which blocks the cell cycle in mitosis and thus does
not affect the G2 arrest point, had no effect on the radiation-
induced DNA deavage (Fig. 7 B).
The data presented here demonstrate a direct inhibition
by cytokines of apoptosis triggered by DNA damage. Var-
ious cytokines have been shown to be radioprotective in vivo
(for review see reference 26). They can be classified into those
that can be given before radiation and afford subsequent pro-
tection, such as ILl (27) and YNF-ol (27), and those that
have been shown to be myelorestorative when given after ir-
radiation. IL-3 belongs to the latter category (28), and under
these conditions, presumably acts to stimulate the prolifera-
tion of residual undamaged progenitor cells. The direct pro-
tection of BAF3 cells by Ib3, and other growth-promoting
factors, suggests that bone marrow progenitor cells might
be protected by appropriate cytokines administered during
chemo- or radiotherapy of nonhaematopoietic tumors. It also
implies that differential sensitivity of a variety of tumors to
DNA damaging drugs or radiation could be explained by
their production of autocrine growth factors, or overexpres-
sion of the bd2 protein.
The presence of Ib3 does not affect the initial extent of
DNA breaks caused by irradiation, or their rate of repair.
Therefore, the mechanism by which DNA damage leads to
apoptosis in BAF3 cells deprived of Ib3 involves the detec-
tion of DNA repair. The arrest of eucaryotic cells in the G2
phase of the cell cycle subsequent to DNA damage and re-
pair is a well-recognized phenomenon (25) and, although the
molecular basis of this control is not understood, yeast mu-
tants deficient in this regulation have been isolated (29). The
death of BAF3 cells from such a G2 restriction point and
its inhibition by preventing cells from reaching the restric-
tion point, demonstrates that cells are particularly sensitive
to the onset of apoptosis when arrested there in the absence
of IL-3. However, it is dear that arrest in G2 per se is not
necessary for recovery of BAF3 cells from DNA damage in
the presence of IL3. Bcl2 overexpressing cells, which remain
arrested in G1 in the absence of IL-3, are also resistant to
X-irradiation. Such radio-resistance has also been reported
in thymocytes isolated from transgenic mice overexpressing
bcl2 (30, 31), and we have also demonstrated that bcl2 over-
expressing BAF3 cells are resistant to DNA damaging agents
such as etoposide (Ascaso, R., M. Collins, J. Marvel, and
A. Lopez-Rivas, manuscript in preparation). Whether main-
tenance or induction of bcl2 function represents part of the
mechanism of action of Ib3 remains to be investigated. A
decrease in bcl2 mRNA has been reported late after Ib3
removal from a factor-dependent cell line (10).
IL-3 removal from either control, or X-irradiated BAF3
cells does not affect the cycling of the cells. Furthermore,
IL-3 is not required to protect cells from apoptosis until a
time point very close to the initiation of DNA cleavage. This
implies that 11,-3 is not only a growth factor that is required
to maintain cell proliferation of these factor-dependent cells,
but also a survival factor that prevents the onset of DNA
cleavage. The signaling pathway that leads to apoptosis after
II.-3 removal requires at least 8 h for the initiation of this
cleavage. DNA damage either induces a different pathway
or the same pathway with more rapid kinetics, as cleavage
in this case starts after 4 h. The ability of II_-3 to rapidly re-
verse the apoptotic program when added close to the time
of nudease cleavage, argues against a coordinated and irre-
versible program of gene expression leading to apoptosis. One
possibility is that II~3 acts to maintain the level of crucial
1049 Collins et al.
cellular metabolites such as ATP (32). It is known that the
presence of DNA strand breaks causes the activation of
poly(ADP-dbose)polymerase, which consumes cellular NAD +
leading to a fall in ATP (33). This could explain the more
rapid onset of apoptosis after DNA damage. To define the
mechanism of action of IL-3, it is now necessary to identify
the cellular changes that immediately precede onset of DNA
cleavage, determine how they lead to an increase in nuclease
activity, and demonstrate their rapid reversal by Ib3.
The eDNA encoding human bcl2 was a gift from Dr. G. Nunez, University of Michigan, Ann Arbor, and
the anti-human bcl2 antibody was a gift from Dr. D. Mason, Oxford University. Professor C. J. Marshall
and Dr. O. Danos provided valuable suggestions for the manuscript.
This work was supported by the Cancer Research Campaign, the Comision Interministerial de Ciencia
y Tecnologia (SAL91-0411), and the Leukaemia Research Fund.
Address correspondence to Mary K. L. Collins, Chester Beatty Laboratories, Institute of Cancer Research,
237 Fulham Road, London SW3 6JB, UK.
Received for publication 6 May 1992 and in revised form 29 June 1992.
1. Murphy, K., A. Heimberger, and D. LOh. 1990. Induction by
antigen of intrathymic apoptosis of CD4+CD8 + TCR ~~
thymocytes in vivo. Science (Wash. DC). 250:1720.
2. Liu, Y., D. Joshua, G. Williams, C. Smith, J. Gordon, and
I. MacLennan. 1989. Mechanism of antigen-driven selection
in germinal centres. Nature (Lond.). 342:929.
3. Lee, F., T. Yokota, T. Otsuka, L. Gemmell, N. Larson, J. Luh,
K. Arai, and D. Rannick. 1985. Isolation ofcDNA for a human
granulocyte-macrophage colony-stimulating factor by functional
expression in mammalian cells. Proc. Natl. Acad. Sci. USA.
4. Sonoda, Y., Y. Lang, G. Wong, S. Clark, and M. Ogawa. 1988.
Erythroid burst-promoting activity of purified recombinant
human GM-CSF and interleukin-3: studies with anti-GM-CSF
and anti-IL-3 sera and studies in serum-free cultures. Blood.
5. Williams, G., C. Smith, E. Spooncer, T. Dexter, and D. Taylor.
1990. Haemopoietic colony stimulating factors promote cell
survival by suppressing apoptosis. Nature (Lond.). 343:76.
6. Rxxlriguez-Tarduchy, G., M. Collins, and A. Lopez-Rivas. 1990.
Reguhtion of apoptosis in interleukin-3-dependent hemopoietic
cells by interleukin-3 and calcium ionophores. EMBO (Fur. Mol.
Biol. Organ.) J. 9:2997.
7. Van-Zant, G. 1984. Studies of hematopoietic stem cells spared
by 5-fluorouracil. J. Extx Med. 159:679.
8. Koyasu, S., A. Tojo, A. Miyajima, T. Akiyama, M. Kasuga,
A. Urabe, and J. Schreurs. 1987. Interleukin 3-specific tyro-
sine phosphoryhtion of a membrane glyr.oprotein of Mr 150,000
in multi-factor-dependent myeloid cell lines. EMBO (Fur. Mol.
Biol. Organ.) J. 6:3979.
9. Rodriguez-Tarduchy, G., P. Malde, A. Lopez-Rivas, and M.
Collins. 1992. Inhibition of apoptosis by calcium ionophores
in IL3-dependent bone marrow cells is dependent upon produc-
tion of IL4. J. Immunol. 148:1416.
10. Nunez, G., K. London, D. Hockenberry, M. Alexander, J.
McKearu, and S. Korsmeyer. 1990. De-regulated bd2 gene ex-
pression selectively prolongs survival of growth factor-deprived
hemopoietic cell lines. J. Immunol. 144:3602.
11. Hockenberry, D., G. Nunez, C. Milliman, R. Schreiber, and
S. Korsmeyer. 1990. Bcl-2 is an inner mitochondrial protein
that blocks programmed cell death. Nature (Lond.). 348:334.
12. Bardon, S., F. Vignon, P. Montcourrier, and H. Rochefort.
1987. Steroid receptor-mediated cytotoxicity of an antiestmgen
and an antiprogestin in breast cancer cells. Cancer Res. 47:1441.
13. Wyllie, A. 1980. Glucocorticoid-induced thymocyte apoptosis
is associated with endogenous endonudease activation. Nature
14. Sorenson, C., M. Barry, and A. Eastman. 1990. Analysis of
events associated with cell cycle arrest at G2 phase and cell death
induced by cisplatin. J. Natl. Cancer Inst. 82:749.
15. Shimizu, T., M. Kubota, A. Tanizawa, H. Sano, Y. Kasai, H.
Hashimoto, Y. Akiyama, and H. Mikawa. 1990. Inhibition
of both etoposide-induced DNA fragmentation and activation
of poly(ADP-ribose) sythesis by zinc ions. Biochem. Biophys.
Res. Commun. 169:1172.
16. Sellins, K., and J. Cohen. 1987. Gene induction by gamma-
irradiation leads to DNA fragmentation in lymphocytes.J. Im-
17. Palacios, R., and M. Steinmetz. 1985. IL3-dependent mouse
clones that express B-220 surface antigen, contain Ig genes
in germ-line configuration and generate B lymphocytes in vivo.
18. Kemp, L., S. Sedgwick, and P. Jeggo. 1984. X-ray sensitive
mutants of Chinese hamster ovary cells defective in double-
strand break rejoining. Mutation Res. 132:189.
19. Mann, R., R. Mulligan, and D. Baltimore. 1983. Construc-
tion of a retrovirus packaging mutant and its use to produce
helper-free defective retrovirus. Cell. 33:153.
20. Pezzella, F., A. Tse, J. Cordell, K. Pulford, K. Garter, and D.
Mason. 1990. Expression of the bd-2 oncogene protein is not
specific for the 14;18 chromosomal translocation. Am.J. Pathol.
21. Liu, L. 1989. DNA topoisomerase poisons as antitumor drugs.
Annu. Rev. Biochem. 58:351.
22. Eastman, A. 1987. The formation, isolation and characterisa-
tion of DNA adducts produced by anticancer platinum com-
pounds. Pharmacol. & Ther. 34:155.
23. Rodriguez-Tarduchy, G., M. Collins, I. Garcia, and A. Lopez-
1050 Interhukin 3 Protection of Cells from DNA Damage
Rivas. 1992. Insulin-like growth factor-I inhibits apoptosis in
mouse hematopoietic cells. J. Imraunol. In press.
24. Bradley, M., and K. Kohn. 1979. X-ray induced DNA double-
strand break production and repair in mammalian ceils as mea-
sured by neutral filter elution. Nucleic Acids Res. 7:793.
25. Tobey, K. 1975. Different drugs arrest ceUs at a number of
distinct stages in G2. Nature (Lond.). 254:245.
26. Neta, K., and J. Oppenheim. 1991. Radioprotection with
cytokines- learning from nature to cope with radiation. Cancer
Cells. (Cold Spring Harbor). 3:391.
27. Neta, K., S. Douches, and J. Oppenheim. 1986. Interleukin
1 is a radioprotector, f Iramunol. 136:2483.
28. GlasFy, J., and D. Golde. 1990. Clinical trials of myeloid growth
factors. Ext~ Heraatol. (Oak Ridge). 18:1137.
29. Weinart, T., and L. HartweU. 1988. The RAD9 gene controls
the cell cycle response to DNA damage in Saccharorayres
cerevisiae. Science (Wash. IX2). 241:317.
30. Strasser, A., A. Harris, and S. Cory. 1991. bd-2 transgene in-
hibits T cell death and perturbs thymic self-censorship. Cell.
31. Sentman, C., J. Shutter, D. Hockenberry, O. Kanagawa, and
S. Korsmeyer. 1991. bd-2 inhibits multiple forms of apoptosis
but not negative selection in thymocytes. Cell. 67:879.
32. Whetton, A., S. Huang, and P. Monk. 1988. Adenosine
triphosphate can maintain multipotent haemopoietic stem cells
in the absence of interleukin 3 via a membrane permeabiliza-
tion mechanism. Biochera. Biophys. Res. Coraraun. 152:1173.
33. Berger, N. 1985. Poly(ADP-ribose) in the cellular response to
DNA damage. Radiation Res. 101:4.
1051 Collins et al.