Immunity, Vol. 9, 47±57, J uly, 1998, Copyright 1998 by Cell Press
Mice Defective in Two Apoptosis Pathways
in the Myeloid Lineage Develop
Acute Myeloblastic Leukemia
Recent work has shown that the AML1/ETO fusion pro-
teinacts as a dominantnegative inhibitorofnormalAml1
function (Yergeauet al., 1997)and also up-regulates the
expression of the bcl-2 protooncogene (Klampfer et al.,
Bcl-2 was discovered upon characterization of a
t(14;18) chromosomal translocation found in follicular
lymphoma (Bakhshiet al., 1985; Cleary and Sklar, 1985)
and was subsequently shown to protect cells from a
wide variety of apoptotic cues (Vaux et al., 1988; re-
viewed in Yang and Korsmeyer, 1996). The role of Bcl-2
in oncogenesis has been confirmed mainly in lymphoid
cells, but recent work has suggested that deregulation
of the bcl-2gene may be important inthe transformation
of myeloid cells. Leukemic cells from mosthuman AMLs
have been found to express Bcl-2 at levels muchhigher
thantheirnormalcellularcounterparts (Delia et al., 1992;
Bensi et al., 1995), suggesting that deregulation of the
bcl-2 gene may be one of the critical events in the my-
eloid transformation process. The extension of cellular
survival by Bcl-2 may allow sufficient time forthe acqui-
sition of additional oncogenic mutations. This mecha-
nismmay underly the progression from chronic to acute
leukemia (Rabbitts, 1991, 1994).
To test directly the role of bcl-2 in the development
of myeloid leukemia, we have created a transgenic
mouse inwhichconstitutive expressionof thebcl-2gene
is targeted exclusively to myeloid cells by the hMRP8
promoter (Lagasse and Weissman, 1994). hMRP8bcl-2
mice develop a disease that in many ways is analogous
to human chronic myelomonocytic leukemia (CMML)
(Lagasse and Weissman, submitted). hMRP8bcl-2 trans-
genic animals develop monocytosis with age and be-
come neutropenic dueto ashiftingranulopoiesis toward
immature celltypes. Whereas monocytes fromwild-type
mice rapidly undergoapoptosis invitro, monocytes from
hMRP8bcl-2 mice are able to survive in the absence of
serum or exogenous growth factors. We have recently
foundthat monocytes frompatients withCMML similarly
have extended survival in vitro while monocytes from
healthy volunteers rapidly undergo apoptosis, suggesting
that a disorderin programmed cell death of monocytes
is a critical feature of CMML (Lagasse and Weissman,
submitted). hMRP8bcl-2 mice show decreased survival
compared to their littermate controls but rarely develop
acute malignancies (Lagasse and Weissman, submit-
ted). Despite the deregulation of Bcl-2 expressionfound
in many human cancers, overexpression of Bcl-2 alone
has beenfoundto be relativelybenign interms ofcellular
transformation in transgenic mouse models (Cory et al.,
1994). However, deregulated expression of Bcl-2 cou-
pled to additional mutations such as enforced expres-
sion of c-Myc can rapidly lead to the transformation of
cells of the B lymphoid lineage (Vaux et al., 1988). We
thus reasoned that additional mutations are likely to
act in concert with bcl-2 deregulation to promote acute
Loss-of-functionmutations inthe Fas receptor(CD95)
or in components of the Fas signaling pathway have
been implicated in several human AMLs (Robertson et
David Traver,*Koichi Akashi,
Irving L. Weissman, and Eric Lagasse
Department of Pathology
Department of Developmental Biology
Stanford University School of Medicine
Stanford, California 94305
Fas-deficient (Faslpr/lpr) mice constitutively expressing
Bcl-2 in myeloid cells by the hMRP8 promoter often
develop a fatal disease analogous to human acute
myeloblastic leukemia (AML-M2). Hematopoietic cells
from leukemic Faslpr/lprhMRP8bcl-2 animals form clo-
nogenic blastcolonies invitro andcan transferdisease
to wild-type mice. In vitro ligation of Fas on Fas?/?
hMRP8bcl-2 marrow cells depletes approximately 50%
of myeloid progenitor activity, demonstrating that Bcl-2
can only partially block Fas-mediated death signals in
myelomonocytic progenitors. Inaddition, Faslpr/lprmar-
row contains greatly increased numbers of myeloid
colony-forming cells as compared to Fas?/?controls.
Taken together, these data suggest that Fas has a
novel role in the regulation of myelopoiesis and that
Fas may act as a tumor suppressor to control leuke-
mogenic transformation in myeloid progenitor cells.
Acute myeloid leukemia (AML) accounts for over 80%
of all adult acute leukemias (Schiffer, 1997)and is char-
acterized by a clonal expansion of immature myeloid
cells in all hematopoietic tissues. Many patients pro-
gress to AML from preleukemic myelodysplastic syn-
dromes (MDS) or from chronic myelogenous leukemia
(CML).This progressionto blastcrisisis thought toresult
from the accumulation of genetic lesions in a single
self-renewing progenitor cell that prevents the normal
maturation of the mutant cell and its progeny (reviewed
in Sawyers et al., 1991).
Acutemyeloblastic leukemia withmaturation, orAML-
M2 as classified by the French-American-British (FAB)
system (Bennett et al., 1976), is the most common sub-
type of AML and is characterized by an accumulation of
granulocyte precursors in all hematolymphoid tissues.
Surveys aimed at identifying molecular abnormalities in
AML-M2 have found that approximately 40% of AML-
M2 patients possess a t(8;21) chromosomal transloca-
tion that juxtaposes the aml1 and eto genes (Miyoshi et
al., 1991).Aml1 is a transcriptionfactor of the core bind-
ing factor (CBF) family and has been shownto up-regu-
late the expression of several growth factor genes in-
cluding m-csf, gm-csf, and il-3 (reviewed in Tenen et
al., 1997), and is necessary forthe formationof allhema-
topoietic lineages (Okudaet al., 1996; Wang etal., 1996).
*To whom correspondence should be addressed (e-mail: dtraver@
Figure 1. Morphology of Leukemia in Faslpr/lprhMRP8bcl-2 Mice
(A±C) Preparations of marrow (A), spleen (B), and blood (C) cells of leukemic mice compared to controls. Note predominance of myeloblasts
in all hematopoietic tissues in leukemic Faslpr/lprhMRP8bcl-2 mouse (May-Gru Ènwald/Giemsa, all photos ?187.5 original magnification).
(D) Expression of human Bcl-2 in splenocytes. Expression is seen in all myeloid cells, including intermediate staining in leukemic blasts (?125
(E) Staining for NCAE in control (left) versus leukemic (right) splenocytes. Tumor cells stain slightly positive (red) for the granulocytic NCAE
marker (?125 original magnification).
Blockade of Myeloid Cell Death Leads to AML
al., 1995;Bouscary etal., 1997).Despite this, nocompel-
ling association between loss of Fas function and sub-
type of AML has been found. Granulocytes and their
myeloblastic progenitors are knownto express highlev-
els of Fas (Liles et al., 1996), and several patients with
AML-M2 have been shown to have myeloblasts with
functional deficiencies in the Fas signaling pathway
(Robertson et al., 1995; Dirks et al., 1997). In light of
these clinical findings, we bred Fas-deficient Faslpr/lpr
mice to hMRP8bcl-2 transgenic animals to determine
whether these two mutations that block programmed
cell deathcan cooperate to generate malignantmyeloid
Here, we present an in vivo model of murine acute
myeloblastic leukemia inFas-deficientanimals constitu-
tively expressing Bcl-2 in myeloid cells. Approximately
15% of Faslpr/lprhMRP8bcl-2 mice reproducibly develop
disease similar to human AML-M2, suggesting that de-
regulated expression of Bcl-2 and loss of Fas function
may be crucial events in the transformation of myelo-
blasts in both mouse and human. These results demon-
strate a previously unknown role for Fas in protecting
myeloid precursors from leukemogenic transformation.
In addition, predisposition to AML in Faslpr/lprhMRP8bcl-2
mice but not in Faslpr/lprorhMRP8bcl-2 mice alone sug-
gests that Fas and Bcl-2 regulate distinct pathways of
programmed cell death in the myeloid lineage. Finally,
because not all mice bearing these mutations develop
AML, itis likely that otheroncogenic events are involved
in the progression to acute myeloblastic leukemia.
observation period of approximately 7 months. Of the
9 leukemic Faslpr/lprhMRP8bcl-2 mice, 7 became mori-
bund by 3±8 weeks of age and the other 2 at 4 and 7
months. Marrow in leukemic animals was hypercellular
and was largely composed of immature granulocytic
cells, of which myeloblasts predominated (Figure 1A;
Table 1). Extramedullary hematopoiesis and prolifera-
tion of leukemic cells occurred in the spleen (Figure
1B; Table 1), leading to splenomegaly in leukemic mice.
Greaterthan30% of peripheralblood mononuclearcells
in leukemic animals were myeloblasts (Figure 1C; Table
1). Invasion of leukemic cells into the liver, kidneys, and
lungs was also observed (data not shown). Leukemic
blasts were Mac-1int, Gr-1int, c-Kithiand negative for
lymphoid and erythroid markers by flow cytometry, con-
firming their myeloid origin (data not shown). Leukemic
cells expressed moderate to high levels of human Bcl-2
(Figure 1D)and stained positively fornaptholAS-D chlo-
roacetate esterase (NCAE) (Figure 1E), a diagnostic
marker of acute myeloblastic leukemia in humans. Leu-
kemic myeloblasts were immature in appearance with
open chromatin structure and few granules (Figures
1A±1C and 1G). Histochemical stains formonocytic and
lymphocytic markers were uniformly negative in leuke-
mic blast cells (data not shown).
In summary, leukemic Faslpr/lprhMRP8bcl-2 mice had
an expansionof myeloblasts inallhematopoetic tissues.
Leukemic myeloblasts appeared partially able to give
rise to more mature granulocytic elements, but the sub-
stantially lowered numbers of granulocytes in the mar-
row and blood of leukemic animals suggests that granu-
locytic maturation is impaired (Table 1). This phenotype
is remarkably similarto humanAML-M2, oracute myelo-
blastic leukemia withmaturation, whichis characterized
by ?30% myeloblasts, ?10% maturing granulocytic ele-
ments, and ?20% monocytic cells in the nonerythroid
marrow compartment (NEC). Averages from six leuke-
mic Faslpr/lprhMRP8bcl-2mice showed29% myeloblasts,
46% maturing granulocytes, and 15% monocytic cells
in the NEC (Table 1). A diagnostic hallmark of AML-M2
is the presence of stacked granule components known
as Auer rods in the cytoplasm of leukemic myeloblasts.
We have never observed Auer rods in the blast cells of
any leukemic Faslpr/lprhMRP8bcl-2 mice. To our knowl-
edge, Auer rods have yet to be described in any murine
myeloid leukemia, possibly reflecting a difference be-
tween species. It is also possible, due to the apparent
immaturity of leukemic myeloblastsinFaslpr/lprhMRP8bcl-2
mice (Figure 1G),thatmaturationarrest occurs ata stage
in myeloblastic differentiation before sufficient granule
components have been synthesized.
No AML was observed in Faslpr/?hMRP8bcl-2 animals
or in Faslpr/lpranimals crossed to the independently-
derived hMRP8bcl-2/1 transgenic founderline, in which
the level of Bcl-2 expression is substantially lower and
difficult to detect in myeloid progenitor cells (data not
shown). Faslpr/lprhMRP8bcl-2 mice that did not develop
Generation of Faslpr/lprhMRP8bcl-2 Mice
hMRP8bcl-2 mice overexpressing human Bcl-2 under
the control of the human MRP8 promoter were gener-
ated as previously described (Lagasse and Weissman,
1994). All transgenic animals used in this study were
derived from the high-expressing hMRP8bcl-2/2 founder.
Transgenic animals were crossed with B6.MRL-Faslpr/lpr
mice. Resulting F1Faslpr/?animals werebledandscreened
forhuman Bcl-2 expressionby flow cytometry. The sta-
tus of the fas gene in F2animals was determined by
PCR analysis (see ExperimentalProcedures for details).
No deviations from expected genotype percentages
Faslpr/lprhMRP8bcl-2 Mice Are Predisposed
to Acute Myeloid Leukemia
hMRP8bcl-2 mice progressively develop monocytosis
and associated neutropenia that leads to decreased
survival compared to littermate controls after 1 year of
age (Lagasse and Weissman, submitted). Despite this
CMML-like disease, hMRP8bcl-2 mice very rarely de-
velop AML. We have also not observed any myeloid
perturbations in Faslpr/lprmice. When these two geno-
types were crossed,however, 9 of60 Faslpr/lprhMRP8bcl-2
mice developed acute myeloblastic leukemia over an
(F±G) Electronphotomicrographs of splenocytes from a leukemic Faslpr/lprhMRP8bcl-2 mouse. Field view (F)shows myeloblasts (closedarrows)
and lymphocytes (open arrows). Note immaturity of myeloblast in closeup (G) as evidenced by the paucity of granules and open chromatin
structure (Field view ?3000, blast closeup ?7100 original magnification).
Table 1. Hematological Parameters of Leukemia in Faslpr/lprhMRP8bcl-2 Mice
Percentage of Cellsa
ErythrocyteTissue Genotypen Blast MetamyelocyteGranulocyte Lymphocyte
1.4 ? 0.8
1.9 ? 1.4
4.7 ? 4.4
3.3 ? 2.0
21.7 ? 9.5
5.2 ? 2.5
5.0 ? 1.7
7.6 ? 3.0
6.3 ? 1.2
17.2 ? 7.3
4.5 ? 1.5
5.5 ? 1.4
9.2 ? 2.3
12.7 ? 7.0
11.1 ? 7.1
40.4 ? 6.1
41.1 ? 4.9
21.5 ? 10.3
31.0 ? 6.8
6.1 ? 3.4
7.6 ? 2.4
6.5 ? 3.1
20.0 ? 3.1
22.9 ? 7.8
10.9 ? 6.2
21.4 ? 4.8
19.2 ? 3.5
17.9 ? 4.9
13.1 ? 4.0
7.3 ? 4.4
19.4 ? 3.7
20.7 ? 9.1
19.2 ? 5.9
10.7 ? 6.3
25.7 ? 3.1
2.2 ? 1.9
1.8 ? 1.8
15.9 ? 7.4
2.7 ? 1.9
4.4 ? 1.1
9.8 ? 2.8
4.2 ? 2.8
5.8 ? 3.4
7.5 ? 5.0
3.3 ? 1.3
6.4 ? 1.9
7.2 ? 4.9
8.1 ? 4.4
2.3 ? 1.4
4.6 ? 2.8
7.2 ? 1.8
11.7 ? 4.7
14.6 ? 4.6
23.1 ? 14.4
89.7 ? 2.8
83.0 ? 3.0
58.7 ? 15.3
48.3 ? 9.2
24.7 ? 13.3
2.4 ? 1.2
2.5 ? 1.1
13.2 ? 10.2
17.0 ? 4.0
16.7 ? 14.3
34.9 ? 23.0
9.0 ? 7.1
1.7 ? 1.3
1.0 ? 1.1
1.0 ? .02
2.1 ? 2.2
11.4 ? 4.2
19.5 ? 5.3
7.4 ? 4.6
13.9 ? 4.2
7.1 ? 5.8
6.1 ? 2.1
3.9 ? 3.1
16.0 ? 7.8
35.4 ? 16.0
15.6 ? 9.0
81.0 ? 7.2
74.9 ? 6.3
72.1 ? 10.2
49.0 ? 13.0
28.7 ? 22.2
3.1 ? 3.1
2.7 ? 2.5
aDifferential cell counts were performed by identifying at least 300 cells per marrow and spleen cytospin and 100 cells per blood smear.
Numbers are presented as means plus or minus standard deviations.
AML displayed differential cell counts similar to those
from transgenic hMRP8bcl-2 animals (Table 1).
weeks following transplantation. The majority of mar-
row, spleen, and blood cells in leukemic recipient ani-
mals were immature myeloid elements, with a prepon-
deranceof myeloblasts (Figure 2D).FACS analysis of the
hematopoietic systems of leukemic recipients showed
cells to be exclusively derived from leukemic CD45.2?
donoranimals (Figure 2E, insets)and that nearly allcells
were myeloid (Figure 2E). Cell cycle analyses showed
transplantedcells to berapidly cycling, with29.0?2.2%
(n ? 3) of marrow and 44.5 ? 9.0% (n ? 3) of spleen
cells in S/G2/M phase. Analysis of nonleukemic Faslpr/lpr
hMRP8bcl-2miceshowed 13.7?1.3% (n?6)ofmarrow
and 14.9 ? 3.7% (n? 6) of spleen cells in S/G2/Mphase.
Both of the transplanted leukemias also transferred le-
thal disease to secondary recipients (data not shown).
Thus, the transplantability of disease suggests that leu-
kemic cells are sufficiently transformed to be inherently
Leukemic Blasts Are Cytokine-Dependent and Form
Clonogenic Blast Colonies In Vitro
The inability of blast cells to differentiate into more ma-
ture cells is a hallmark of acute leukemia. To assess
the differentiation capacity of blast cells from leukemic
Faslpr/lprhMRP8bcl-2 mice, we transferred limiting num-
bers of splenocytes into methylcellulose cultures. As
with most human myeloid leukemias, blast cells were
unable to proliferate in culture without the addition of
exogenous growth factors (Collins, 1987; Akashi et al.,
1991) (Figure 2A). In the presence of individual myeloid
growth factors such as IL-3, GM-CSF, and G-CSF, how-
ever, leukemic blasts rapidly gave rise to clonogenic
colonies (Figure 2A). Of the cytokines tested, IL-3 and
GM-CSF exhibited a synergistic effect on colony forma-
tion. Cultures with both IL-3 and GM-CSF gave rise to
251 ? 10 colonies/5 ? 104cells, while cultures in GM-
CSF orIL-3alone produced 179? 12 or21? 5 colonies,
respectively. Colonies displayed tightly grouped clus-
ters of cells indicative of leukemic blast cell colonies
(Akashi et al., 1991) (Figure 2B), and morphological ex-
amination showed individual colonies to be composed
entirely of immature myeloblasts (Figure 2C).
Fas-Deficient Marrow Contains Increased
Myeloid Progenitor Potential
We next sought to determine whetherFas-deficient he-
matopoietic progenitors possess intrinsic defects that
might predispose Faslpr/lpranimals toward leukemia. To
assess changes in the multipotent hematopoietic stem
cell (HSC) compartment, we compared the frequencies
of Thy-1.1loSca-1?c-Kithilin-/locells, which are highly en-
riched for HSC activity, between all mouse genotypes
used in this study by flow cytometry according to pre-
viously published techniques (Spangrude et al., 1988;
Morrison and Weissman, 1994). While hMRP8bcl-2 ani-
mals displayed significantly increased percentages of
HSCs compared to wild-type animals, no significant dif-
ferences were observed in HSC frequency between
hMRP8bcl-2 and Faslpr/lprhMRP8bcl-2 mice (data not
shown). To test primitive multipotent progenitors func-
tionally, we performed spleen colony-forming unit assays
(CFU-S) with marrow donor cells from all genotypes
used in this study. Day 12 CFU-S activity is correlated
Acute Myeloid Leukemia Can Be Transplanted
to Recipient Animals
To test further the transformation of cells from leukemic
Faslpr/lprhMRP8bcl-2 mice, mixtures of 5 ? 106marrow
cells and 1 ? 107splenocytes were injected into lethally
irradiated or nonirradiated wild-type recipient animals
(see ExperimentalProcedures fordetails). Onlyrecipient
animals whose hematolymphoid systems had been ab-
lated by lethal irradiation supported donor-derived leu-
kemic cells. Cells from two of six leukemic Faslpr/lpr
hMRP8bcl-2 mice used as donors transferred disease,
and recipients developed fatal blast crises at 3 and 15
Blockade of Myeloid Cell Death Leads to AML
Figure 2. Transformation Assays
(A) Colony production from leukemic Faslpr/lpr
hMRP8bcl-2 splenocytes in methylcellulose
supplemented with growth factors. Numbers
of colonies presented are per 5 ? 104input
(B)Clonogenic day 3 colonyfrom culture con-
taining IL-3/GM-CSF. Note tight grouping of
cells and homogeneity of cell size indicative
of leukemic blast colonies.
(C) Morphology of cells from colony in (B).
All cells retain the myeloblastic morphology
observed inthe leukemic donormouse (May-
Gru Ènwald/Giemsa staining).
(D)Morphology of marrow,spleen, and blood
cells from a C57BL/Ka-Thy1.1, CD45.1 re-
cipient mouse that developed AML 3 weeks
following transplantation from a leukemic
Faslpr/lprhMRP8bcl-2 donor animal (?187.5
(E)FACS analysis of hematopoietic tissues of
the leukemic recipient in (D). Almost all cells
are myeloidinmarrow, spleen, and blood and
are exclusively derived from the leukemic
Faslpr/lprhMRP8bcl-2 (CD45.2) donor animal
(insets). Shaded histograms represent CD45.2
profiles fromaC57BL/Ka-Thy1.1,CD45.1 ani-
mal that did not receive transplanted cells.
with long-term and transient multipotent progenitor cell
frequencies (Till and McCulloch, 1961; Wu et al., 1967).
Marrow cells (1 ? 105) were injected into lethally irradi-
ated syngeneic recipients, and spleen colonies were
counted on day 12. Only modest increases in colony
numbers were seen in animals receiving marrow cells
from Faslpr/lpr, hMRP8bcl-2, and Faslpr/lprhMRP8bcl-2
mice compared to wild-type marrow (Figure 3A). Thus,
no significant increases in HSC frequency or in day 12
CFU-S multipotent progenitoractivity were seen in Fas-
To compare functional differences in myeloerythroid
progenitor activity, we performed day 8 CFU-S experi-
ments using 1 ? 105input marrow cells (Becker et al.,
1963; Wu et al., 1967). In contrast to the day 12 CFU-S
results, Faslpr/lprmice displayed large differences in day
8 colony numbers compared to controls (Figure 3B).
Whereas hMRP8bcl-2 marrow produced an approxi-
mate 2-fold increase in colony numbers, Faslpr/lprmarrow
contained a 5-fold increase in myeloid progenitoractiv-
ity (Figure 3B). Marrow from Faslpr/lprhMRP8bcl-2 mice
also showed a 5-fold increase in myeloid progenitor
activity (Figure 3B). Morphological examination of colo-
nies derived from each donor genotype showed most
cells to be undifferentiated myeloid blasts and showed
no significant differences among remaining cell types
(data not shown).
Both day 12 and day 8 CFU-S assays using 1 ? 106
donor spleen cells showed increased colony numbers
only in recipients injected with hMRP8bcl-2?spleno-
cytes (Figures 3A and 3B, respectively). This is due to
increased numbers of progenitors in the spleen that
results from extramedullary hematopoiesis occurring in
these animals (Lagasse and Weissman, submitted).
Blastic Myeloid Cells Express High Levels of Fas
The above results suggest that signaling through the
Fas receptor negatively regulates myelopoiesis by con-
trolling homeostasis of myeloid progenitorpopulations.
We thus predicted that Fas expression would be high
on myeloid progenitor cells in the marrow of wild-type
mice. Marrow cells were divided by flow cytometry into
Figure 3. CFU-S Assays
(A) Colony numbers from spleens harvested
12 days after injection of limiting numbers of
marrow or spleen cells. Numbers are pre-
sented as colonies/1 ? 106input cells.
(B) Colony numbers fromspleensharvested on
day 8. Note increased numbers of colonies
derived from the marrow of Fas-deficient
mice. Heterogeneity in splenic colony num-
bers from hMRP8bcl-2 transgenics is due to
variable extramedullary hematopoiesis oc-
curring in these mice.
three populationsÐ low, intermediate, and highÐ based
onexpressionofthe Fas receptor(Figure 4A).Increasing
levels of Fas correlated with increasing levels of the
myeloid activation marker Mac-1 (Figure 4B). Scatter
profiles of marrow myeloid cells were then obtained by
gating on cells positive for both Mac-1 and Gr-1 (Figure
4C). Blastic myeloid cells (72%) were found to express
high levels of Fas (Figure 4C). Cells sorted from this
population displayed the morphology of immature my-
eloid cells (data not shown).
similar up-regulation, we plated marrow cells in media
containing GM-CSF and IL-3,potentstimulatorsof gran-
ulopoiesis (Metcalf, 1989). Compared to cultures con-
taining serum alone, Fas levels rose significantly on my-
eloid cells in cultures supplemented with GM-CSF/IL-3
over 2 days (Figures 5A and 5B, respectively). Myeloid
cells from hMRP8bcl-2 marrow showed an especially
large increase in Fas expression(Figure 5B). We cannot
rule out the possibility, however, that selective survival
ofsubpopulations ofcells may accountforthe increased
numbers of cells with higher Fas expression.
Fas Expression Is Up-Regulated by Myeloid
Fas is knownto be up-regulated on lymphoid cells upon
antigen receptor triggering and/or upon stimulation by
activating cytokines (Daniel and Krammer, 1994; Alder-
son et al., 1995). To determine if myeloid cells exhibit
Myeloid Progenitors Are Susceptible
to Fas-Mediated Killing
That myeloid progenitor cells express high levels of Fas
(Figure 4C)and that myeloid activation is coupled to Fas
up-regulation(Figure 5B) suggests that the Fas receptor
Figure 4. Myeloid Blasts Express High Levels of Fas
(A) Histogram showing Fas expression on marrow cells from a Fas?/?mouse. Cells were gated based on Fas expression for further analysis
in (B). Cells within the Faslo, Fasint, and Fashigates represent 36%, 42%, and 22% of the total population, respectively. Lighter histogram
represents Fas staining on marrow cells from a Faslpr/lprmouse.
(B) Myeloid markers were analyzed on populations of marrow cells subdivided by Fas expression levels. Myeloid cells with the highest levels
of Mac-1/Gr-1 expression were found in the Fashifraction.
(C) Myeloid cells were gated from total marrow cells based on expression of Mac-1/Gr-1 markers from (B) (boxes) and analyzed for scatter
characteristics. Marrow cells from the Fashicompartment contained the majority of blastic myeloid cells.
Blockade of Myeloid Cell Death Leads to AML
Figure 5. Up-Regulation of Fas by Growth
(A) Marrow cells from Faslpr/lpr, Fas?/?, and
hMRP8bcl-2 mice were cultured for 2 days in
media containing 10% FCS. Cells were sub-
sequently removed, washed, stained for Fas,
and analyzed by flow cytometry. Gr-1?cells
are displayed in histograms. Mean fluores-
cence intensities for Faslpr/lpr, Fas?/?, and
hMRP8bcl-2 plots are 25.2, 44.5, and 67.6,
(B) Marrow cells were cultured as in (A) with
the addition of GM-CSF and IL-3. Cytokine
stimulation up-regulated Fas expression on
myeloid cells; the largest increase occurred
incells fromhMRP8bcl-2 mice.Mean fluores-
cence intensities are 23.5, 55.5, and 91.9 for
Faslpr/lpr, Fas?/?, and hMRP8bcl-2 plots, re-
may functionally regulate myelopoiesis. To test this di-
rectly, we performed Fas-killing assays on marrow my-
eloid cells. Marrow cells were harvested from control,
hMRP8bcl-2, Faslpr/lpr, and Faslpr/lprhMRP8bcl-2mice and
activated in vitro for 2 days with GM-CSF and IL-3. The
Fas receptorwas thenligated onthese cells by antibody
cross-linking; cells were subsequently plated in methyl-
cellulose containing myeloid growthfactors (see Experi-
mental Procedures for details). Following Fas ligation,
colony counts dropped approximately 5-fold from wild-
typemarrow and 2-fold fromhMRP8bcl-2 marrow, com-
pared to respective marrow samples without antibody
(Table 2). Marrow cells from Fas-deficient animals treated
in colony numbers (Table 2).
least partially nonoverlapping, pathways of programmed
cell death in myeloid progenitors. This is in agreement
with previous evidence showing that overexpression of
Bcl-2 in FDC-P1 murine myeloid leukemia cells could
only partially block Fas-mediated killing signals (Itoh et
al., 1993). Likewise, Fas and Bcl-2 apparently regulate
independentapoptotic pathways incells of thelymphoid
system (Strasser et al., 1995).
The remarkable conservation in disease phenotype
among numerous leukemic Faslpr/lprhMRP8bcl-2animals
suggests that myeloblasts are prone to transformation
by this combinationof mutations.Thatthis murineleuke-
mia so closely resembles human AML-M2also suggests
that the molecular bases for these diseases may be
conserved. Several clinical studies support our finding
in this mouse model that blockade of the cell death
pathways normally regulated by Bcl-2 and Fas may lead
to AML-M2. Leukemic blasts from patients with AML-
M2 have been shown to express high levels of Bcl-2
relative to normal blasts from healthy volunteers (Delia
et al., 1992; Bensi et al., 1995). Additional studies have
provided evidence thatblastcells fromseveralsubtypes
of AML, including AML-M2, often express decreased
levels of Fas receptorprotein ontheirsurface compared
to normal cellular counterparts (Komada et al., 1995;
Robertsonetal., 1995). Also, leukemic blasts expressing
Fas are often unable to respond to stimulation of the
Fas receptor in vitro (Dirks et al., 1997), suggesting that
aberrant signaling elements downstream of the Fas re-
ceptor may functionally mimic the absence of the Fas
receptor in our model.
The t(8;21) chromosomal translocation that creates
the AML1/ETO fusion protein in myeloid progenitors is
Inthis study, we have generatedFas-deficient mice with
constitutive expression of Bcl-2 targeted specifically to
myeloid cells by the hMRP8 promoter. The fact that
many Faslpr/lprhMRP8bcl-2 mice develop AML provides
novel evidence that Fas not only plays a role in myelo-
poietic homeostasis butalso that Fas canact as a tumor
suppressor to control leukemogenic transformation in
granulocyte progenitors. The development of AML in
Faslpr/lprhMRP8bcl-2 animals, but not in either single-
mutantgenotype, shows that gain-of-functionbcl-2 mu-
tations collaborate synergistically with loss-of-function
Fas mutations inthe transformation of primitive myeloid
cells. We further show that constitutive expression of
Bcl-2 only partially blocks Fas-mediated death signals,
suggesting that Fas and Bcl-2 regulate distinct, or at
Table 2. Myeloid Progenitor Assay
30 ? 3
6 ? 2a
82 ? 3
73 ? 6b
50 ? 3
24 ? 6a
35 ? 4
40 ? 6b
Fas receptor cross-linking was performed on marrow cells activated for 2 days in culture. 105marrow cells were then plated in triplicate in
methylcellulose containing GM-CSF/IL-3. Colonies were counted at day 6 and are presented as means plus or minus standard deviations.
Representative experiment of three shown.
ap ? 0.05.
bNo significant difference.
currently the best knownmolecular anomaly associated
withAML-M2 (Miyoshi etal., 1991; Ericksonetal., 1992).
Patients in long-term remission following treatment for
t(8;21)?AML-M2 have been found to maintain produc-
tion of the AML1/ETO fusion protein in myeloerythroid
progenitor cells (Nucifora et al., 1993; Miyamoto et al.,
1996). Thissuggests thatt(8;21)?progenitorsrequire addi-
tional genetic lesions for transformation. Similarly, in-
complete incidence of disease in Faslpr/lprhMRP8bcl-2
mice suggests that somatic mutations are necessary
for leukemogenesis. It is thus possible that a similar
translocation or deregulation of the aml1 or eto genes
may collaborate with mutagenesis of the apoptosis
pathways regulated by Fas and Bcl-2 to lead to murine
AML-M2. Karyotypic analyses of leukemic cells from
Faslpr/lprhMRP8bcl-2 mice revealed no gross chromo-
somal abnormalities (data not shown). Since detection
of subtle abnormalities in murine chromosomes is diffi-
cult, we are currently searching for alterations in the
aml1 and eto genes in leukemic blasts from Faslpr/lpr
havelow remissionrates following inductivechemother-
apeutic regimens (Min et al., 1996). Taken together,
these clinical data support our finding that the loss of
Fas functionis important inthe etiology ofacute myeloid
leukemia. Mutations inFas signaling pathways may also
provide a mechanism by which leukemic cells avoid
immune surveillance and resist cytoreductive therapies.
Role of Fas in Granulopoiesis
Transformation of the earliest defined granulocytic pro-
genitor in Faslpr/lprhMRP8bcl-2 mice prompted us to
question the role of the Fas receptor in normal granulo-
poiesis. As AML is often associated with deranged
multipotent progenitor cells, we first sought to deter-
mine whether lack of the Fas receptor might perturb
this population in the marrow of Faslpr/lprmice. Previous
research in our laboratory has shown that Fas is not
expressed in HSCs (Aguila and Weissman, 1996). Ac-
cordingly, HSC frequency did not significantly differ be-
tween Faslpr/lprmice and Fas?/?mice (data not shown).
Day 12 CFU-S colony numbers also showed no signifi-
cant differences between these two genotypes. These
data suggest that the Fas receptor does not play an
important role in regulating hematopoiesis by primitive,
multipotent progenitor cells.
We next assessed the role of Fas in granulocyte pro-
genitors. In vivo, day 8 CFU-S colony production showed
dramatic increases from Faslpr/lprmarrow over Fas?/?
marrow, suggesting that Fas normally limits the prolifer-
ative potential of myeloerythroid progenitors. In vitro,
we show that Fas is expressed on blastic granulocyte
progenitor cells, is up-regulated on myeloid cells cul-
tured with IL-3 and GM-CSF, and can functionally de-
plete myeloid progenitor activity upon receptor cross-
linking. No significant decrease in myeloid progenitor
activity was observed when fresh, uncultured marrow
was usedinsimilarexperiments. Thatsubstantialdeple-
tion of myeloid progenitoractivity requires prior activa-
tion by cytokines suggests that signaling through the
Fas receptor may regulate homeostasis in proliferating
myeloidprogenitors by limiting average cellularlifespan.
A model of homeostatic control of myeloid progeni-
tors through the Fas receptor requires the presence
of Fas ligand±bearing cells. Cytotoxic T lymphocytes
(CTLs) express high levels of FasL and can efficiently
killFas-bearing cells (Kagietal., 1994;Lowinetal., 1994;
Suda and Nagata, 1994). CTLs are rare in bone marrow,
however, and studies in T cell-deficient Faslpr/lprmice
have shown no significant alteration in myelopoiesis or
myeloid leukemogenesis. Recent work has shown that
neutrophils express high levels of FasL (Liles et al.,
1996), suggesting that granulopoiesis may be partially
regulated by a negative feedback mechanism through
interactions between progeny and progenitor.
Faslpr/lprmice display normal steady-state granulo-
poiesis and show no obvious accumulation of myeloid
progenitors, suggesting that additionalmechanisms ex-
ist to eliminate superfluous cells in the myeloid lineage.
Differences in myeloid progenitoractivity are apparent,
however,inspleencolony-forming assays wherelimiting
numbers of Fas-deficient marrow cells produce several
times the number of colonies from wild-type marrow
Fas as a Tumor Suppressor
Oncogenesis is kept in check by tumorsuppressorpro-
teins that induce apoptosis in aberrant cells (reviewed
in Marshall, 1991; Weinberg, 1991). The results pre-
sented in this study suggest that the Fas receptor pos-
sesses a tumorsuppressorfunctionin myeloid progeni-
tor cells. hMRP8bcl-2 mice very rarely develop acute
myeloid malignancies; when both alleles of the fas gene
are inactivated, however,these transgenic animals have
a greatly increased incidence of AML.
Trimerization of the Fas receptor transmits potent
death signals in most hematopoietic cells (reviewed in
Nagata, 1997)and induces apoptosis in activatedT lym-
phocytes in a process known as activation-induced cell
death (AICD) (Singer and Abbas, 1994; Alderson et al.,
1995). The accumulation of large numbers of T cells in
Faslpr/lprmice presumably results from the inability of
previously activated peripheral T cells to be cleared (Rus-
sellet al., 1993). Despite this buildup of cells that pheno-
typically resembles a preleukemic condition, Faslpr/lpr
mice are not prone to develop T cell leukemia. Evidence
that Fas may have a tumor suppressor function in B
cells, however, was revealed in T cell±deficient Faslpr/lpr
mice (Peng et al., 1996). Over70% of Faslpr/lprmice lack-
ing T cells developed lethal B cell lymphoma, while T
cell±intact Faslpr/lprmice showed only background inci-
dence. T cell±deficient Fas?/?animals (10%) developed
B lymphoma.These data suggest that T cellsurveillance
normally controls B cell transformation through Fas-
independent mechanisms and that stimulationof Fas by
non±T cells prevents the majority of B cell lymphomas.
A recent clinical study in patients with myelodysplas-
tic syndromes has shown that myeloid progenitor cells
often lose expression of the Fas receptoruponprogres-
sion to AML (Bouscary et al., 1997). This observation
was supported functionally by the finding that cells from
MDS patients in transition to more advanced disease
become increasingly resistant to Fas-mediated killing
in vitro. Studies in patients with AML have shown that
individuals with blast cells expressing low levels of Fas
Blockade of Myeloid Cell Death Leads to AML
AS-D choloracetate esterase was performed according to manufac-
turer's instructions (Sigma Diagnostics). Antibody staining for hu-
man Bcl-2 was done as previously described (Lagasse and Weiss-
cells. This may suggest that expansion of myeloid col-
ony-forming cells is normally regulated through Fas re-
ceptor/Fas ligand interactions in a mannersimilarto the
process of AICD that limits the expansion of activated
lymphoid populations. The greatest expansion of the
myeloid population takes place during the initiation of
myelopoiesis in embryonic development. The fact that
leukemic Faslpr/lprhMRP8bcl-2 mice become consis-
tentlymoribund approximately 5±10weeks following the
initiation of myelopoiesis in utero suggests that trans-
forming events may occur during the early amplification
of the myeloid lineage. It is likely, since hMRP8bcl-2
animals rarely develop acute malignancies, thatimmune
surveillance targets aberrant cells through the Fas re-
ceptor/Fas ligand system and controls similarpreleuke-
mic or leukemic clones from expanding in transgenic or
The results presented in this study have elucidated a
role for the Fas receptor in the complex process of
homeostasis in myeloid progenitorcells. Although other
genes are clearly involved in the regulation of granulo-
poiesis,signaling throughtheFas receptorcaneliminate
activated myeloid progenitors, which may serve to keep
the great proliferative potential of these cells in check.
Significant predisposition to acute myeloblastic leuke-
mia in Faslpr/lprhMRP8bcl-2 mice suggests that Fas and
Bcl-2 regulate distinct pathways of programmed cell
deathinmyeloidprogenitors. Thefactthat notallFaslpr/lpr
hMRP8bcl-2micedevelop AML indicates thatatleastone
more oncogenic event is required in primitive myeloid
cells or their progenitors for the progression to acute
myeloblastic leukemia. Thestriking similarity of this mu-
rine disease to human AML-M2 suggests that similar
defects may underlie transformation of myeloblasts in
both mouse and human and provides a model in which
to study further oncogenic alterations leading toward
Marrow cells were harvested, counted, and plated in RPMI medium
containing 10% fetal calf serum (FCS; Gibco BRL), 50 U/ml rmGM-
CSF (Genzyme), and 200 U/ml rmIL-3 (Genzyme). A fractionof cells
were removed after 36±48 hr and analyzed for Fas expression by
flow cytometry. Remaining cells were split and processed for Fas-
killing assays. All cultures were maintained at 37?C in incubators
containing 7% CO2.
Clonogenic Progenitor Assays
Splenocytes (1 ? 105) from leukemic Faslpr/lprhMRP8bcl-2 mice were
plated in alpha-Modified Eagle Medium±based methylcellulose me-
dia (Methocult M3100; Stem Cell Technologies) that was supple-
mented with 30% FCS, 1% bovine serum albumin, 2 mM L-gluta-
mine, and 50?M 2-mercaptoethanol.The following cytokines, alone
or in combination, were added at the beginning of culture: rhM-CSF
(25 U/ml; Genzyme), rhG-CSF (100 ng/ml; Genzyme), rmGM-CSF
(10 ng/ml), rmIL-3 (30ng/ml), and rmSlf (100 ng/ml;Immunex Corpo-
ration).Colonies were countedonday 3.Allcultures were maintained
at 37?C/7% CO2in humidified chambers.
For Fas-killing assays, marrow cells were incubated on ice with
J o-2 (hamster anti-mouse Fas mAb; 50 ?g/ml; low endotoxin/no
sodium azide; Pharmingen) for 45 min. Cells were washed in PBS,
centrifuged, and resuspended in media containing biotinylated anti-
hamster IgG antibody for 30 min followed by streptavidin (10 ?g/
ml) for 20 min. Cells were washed, counted, and plated in triplicate
in Methocult M3100 supplemented as above and containing rmGM-
CSF (10 ng/ml) and rmIL-3 (30 ng/ml). Colonies were counted on
Mixtures of 5 ? 106marrow cells and 1 ? 107splenocytes from
leukemic animals (C57BL/Ka-Thy1.1, CD45.2)were injected into the
retroorbital sinusoids of congenic recipient animals (C57BL/Ka-
Thy1.1, CD45.1) having previously received either 920 rad in a split
dose or no irradiation. Recipient mice were 8 to 14 weeks of age at
the time of transplantation. Marrow cells (2 ? 105) from C57BL/
Ka-Thy1.1, CD45.1 mice were added to ensure radioprotection of
recipient animals. Mice were monitored twice a week for sickness
and sacrificed when moribund.
Faslpr/lprhMRP8bcl-2animals were generated by breeding hMRP8bcl-2
transgenic mice (C57Bl/Ka-Thy1.1) (Lagasse and Weissman, 1994)
to B6.MRL-Faslpr/lprmice obtained from the J ackson Laboratory (Bar
Harbor, ME). Faslpr/lprhMRP8bcl-2 animals were then bred with
B6.MRL-Faslpr/lpranimals. Offspring were screened as outlined be-
low. All genotypes presented were backcrossed onto the C57Bl/
Ka-Thy1.1 background (referred to in the text as Fas?/?) for at least
three generations to minimize strain differences. All hMRP8bcl-2
animals presented in this study are from the hMRP8bcl-2/2 high-
expressing founderline. Mice presented in this study were generally
between 3±10 weeks of age. Faslpr/lprhMRP8bcl-2 mice showed no
signs of AML unless noted as leukemic. Allanimals were maintained
inStanford University's Research Animal Facilityin accordance with
Single-cell suspensions were prepared from hematopoietic tissues
and stained for FACS analyses using the following monoclonalanti-
bodies (mAbs): M1/70 (anti-Mac-1); 8C5 (anti-Gr-1); ALI-4A2 (anti-
CD45.1); A20.1 (anti-CD45.2); 2B8 (anti-c-Kit); and J o-2 (anti-Fas).
All antibodies except J o-2 (Pharmingen) were prepared in our labo-
ratory and directly conjugated to phycoerythrin, fluorescein-isothio-
cyanate, allophycocyanin, or biotin. Biotinylated MAbs were de-
tected with streptavidin coupled to Texas red (Cappel). Nonspecific
binding of MAbs was prevented by preincubation with 2.4G2, an
MAb which blocks high-affinity Fc receptor subtypes. Viable cells
gated by light scatterand exclusion of propidium iodide (0.5 ?g/ml)
were analyzed on a FACS Vantage (Becton Dickinson). Cell cycle
analyses were performed as previously described (Fleming et al.,
1993). Data analyses were performed using software developed by
the Stanford University Shared FACS Facility.
Faslpr/lprhMRP8bcl-2 mice were screened for human Bcl-2 expres-
sion byflow cytometry as previouslydescribed (Lagasseand Weiss-
man, 1994). Screening for the Faslprmutant allele was carried out
by a PCR protocol published previously (Singer and Abbas, 1994).
Recipient C57BL/Ka mice having previously received a split dose
of 920 rad were retroorbitally injected with 1 ? 105marrow or 1 ?
106spleencells fromcontrol,Faslpr/lpr,hMRP8bcl-2, orhealthyFaslpr/lpr
hMRP8bcl-2 mice. Five wild-type recipient animals were used per
donor mouse; three donors were used per genotype. Recipients
were sacrificed at 8 or 12 days postinjection. Spleen colonies were
visualized and counted as described (Till and McCulloch, 1961).
Histological and Immunohistological Methods
Cells (2 ? 105) from marrow or spleen suspensions were cytocentri-
fuged onto glass slides according tothe manufacturer's instructions
(Shandon). Cellular morphology was assessed following staining in
May-Gru Ènwald and Giemsa solutions (Fluka). Staining for napthol
Acknowledgments isolation of a fusion transcript, AML1/ETO, with similarity to Dro-
sophila segmentation gene, runt. Blood 80, 1825±1831.
Fleming, W.H., Alpern, E.J ., Uchida, N., Ikuta, K., Spangrude, G.J .,
and Weissman, I.L. (1993). Functional heterogeneity is associated
with the cellcycle status of murine hematopoietic stem cells. J . Cell
Biol. 122, 897±902.
Itoh, N., Tsujimoto, Y., and Nagata, S. (1993). Effect of bcl-2 on Fas
antigen-mediated cell death. J . Immunol. 151, 621±627.
Kagi, D., Vignaux, F., Ledermann, B., Burki, K., Depraetere, V., Na-
gata, S., Hengartner, H., and Golstein, P. (1994). Fas and perforin
pathways as major mechanisms of T cell-mediated cytotoxicity.
Science 265, 528±530.
Klampfer, L., Zhang, J ., Zelenetz, A.O., Uchida, H., and Nimer, S.D.
(1996). The AML1/ETO fusion protein activates transcription of
BCL-2. Proc. Natl. Acad. Sci. USA 93, 14059±14064.
Komada,Y., Zhou,Y.W.,Zhang, X.L.,Xue,H.L.,Sakai, H., Tanaka,S.,
Sakatoku, H., and Sakurai, M. (1995).Fas receptor(CD95)-mediated
apoptosis is induced in leukemic cells entering G1B compartment
of the cell cycle. Blood 86, 3848±3860.
Lagasse, E., and Weissman, I.L. (1994). Bcl-2 inhibits apoptosis of
neutrophils but not their engulfment by macrophages. J . Exp. Med.
Lagasse,E., and Weissman, I.L.(1997).Enforced expressionof Bcl-2
in monocytes rescues macrophages and partially reverses os-
teopetrosis in op/op mice. Cell 89, 1021±1031.
Liles, W.C., Kiener, P.A., Ledbetter, J .A., Aruffo, A., and Klebanoff,
S.J . (1996). Differential expression of Fas (CD95) and Fas ligand
on normal human phagocytes: implications for the regulation of
apoptosis in neutrophils. J . Exp. Med. 184, 429±440.
Lowin, B., Hahne, M., Mattmann, C., and Tschopp, J . (1994). Cyto-
lytic T-cell cytotoxicity is mediated through perforin and Fas lytic
pathways. Nature 370, 650±652.
Marshall, C.J . (1991). Tumor suppressor genes. Cell 64, 313±326.
Metcalf, D. (1989). The molecular controlof cell division, differentia-
tion commitment and maturation in haemopoietic cells. Nature 339,
Min, Y.H., Lee, S., Lee, J .W., Chong, S.Y., Hahn, J .S., and Ko, Y.W.
(1996). Expression of Fas antigen in acute myeloid leukaemia is
associated with therapeutic response to chemotherapy. Br. J .
Haematol. 93, 928±930.
Miyamoto, T., Nagafuji, K., Akashi, K., Harada, M., Kyo, T., Akashi,
T., Takenaka, K., Mizuno, S., Gondo, H., and Okamura, T. (1996).
Persistence of multipotent progenitors expressing AML1/ETO tran-
scripts in long-term remission patients with t(8;21) acute myeloge-
nous leukemia. Blood 87, 4789±4796.
Miyoshi, H., Shimizu, K., Kozu, T., Maseki, N., Kaneko, Y., and Ohki,
M. (1991). t(8;21) breakpoints on chromosome 21 in acute myeloid
leukemia are clustered within a limited region of a single gene,
AML1. Proc. Natl. Acad. Sci. USA 88, 10431±10434.
Morrison, S.J ., and Weissman, I.L. (1994). The long-term repopulat-
ing subset of hematopoietic stem cells is deterministic and isolat-
able by phenotype. Immunity 1, 661±673.
Nagata, S. (1997). Apoptosis by death factor. Cell 88, 355±365.
Nucifora, G., Larson, R.A., and Rowley, J .D. (1993). Persistence of
the 8;21 translocation in patients with acute myeloid leukemia type
M2 in long-term remission. Blood 82, 712±715.
Okuda, T., van Deursen, J ., Hiebert, S.W., Grosveld, G., and Down-
ing, J .R.(1996).AML1, the targetofmultiple chromosomaltransloca-
tions in human leukemia, is essential for normal fetal liver hemato-
poiesis. Cell 84, 321±330.
Peng, S.L., Robert,M.E., Hayday, A.C., and Craft, J .(1996). A tumor-
suppressor functionfor Fas (CD95) revealed inT cell-deficient mice.
J . Exp. Med. 184, 1149±1154.
Rabbitts, T.H. (1991). Translocations, master genes, and differences
between the origins of acute and chronic leukemias. Cell 67,
Rabbitts, T.H. (1994).Chromosomaltranslocations inhuman cancer.
Nature 372, 143±149.
Robertson, M.J ., Manley, T.J ., Pichert, G., Cameron, C., Cochran,
We thank Amy Kiger, Douglas Wright, and Annette Schlageter for
critical evaluation of the manuscript, Libuse J erabek for capable
laboratory management and assistance with animal procedures,
Marilyn Masek for electron microscopy, J os Domen for helpful dis-
cussions and technical advice, Veronica Braunstein for antibody
production, Dana Bangs for karyotypic analyses, Laura Olteanu for
technicalassistance, and LucinoHidalgoand Bert Lavarroforanimal
maintenance. D. T. is supported by NationalInstitute of Allergy and
Infectious Diseases Training Grant # 5T32 AI-07290. This work was
supported by a grant from Systemix/Sandoz.
Received March 31, 1998; revised May 22, 1998.
Aguila, H.L., and Weissman, I.L. (1996). Hematopoietic stem cells
are not direct cytotoxic targets of natural killer cells. Blood 87,
Akashi, K., Harada, M., Shibuya, T., Eto, T., Takamatsu, Y., Teshima,
T., and Niho, Y. (1991). Effects of interleukin-4 and interleukin-6 on
the proliferation of CD34? and CD34? blasts from acute myeloge-
nous leukemia. Blood 78, 197±204.
Alderson, M.R., Tough, T.W., Davis-Smith, T., Braddy, S., Falk, B.,
Schooley, K.A., Goodwin, R.G.,Smith, C.A., Ramsdell,F., and Lynch,
D.H. (1995). Fas ligand mediates activation-induced cell death in
human T lymphocytes. J . Exp. Med. 181, 71±77.
Bakhshi, A., J ensen, J .P., Goldman, P., Wright, J .J ., McBride, O.W.,
Epstein, A.L., and Korsmeyer, S.J . (1985). Cloning the chromosomal
breakpoint of t(14;18) human lymphomas: clustering around J H on
chromosome 14 and near a transitionalunit on18. Cell 14, 899±906.
Becker, A.J ., McCulloch, E.A., and Till, J .E. (1963). Cytologicaldem-
onstrationof the clonalnature ofspleencolonies derived fromtrans-
planted mouse marrow cells. Nature 197, 452±454.
Bennett, J .M., Catovsky, D., Daniel, M.T., Flandrin, G., Galton, D.A.,
Gralnick, H.R., and Sultan, C. (1976). Proposals for the classification
of the acute leukaemias. French-American-British (FAB) co-opera-
tive group. Br. J . Haematol. 33, 451±458.
Bensi, L.,Longo,R., Vecchi,A., Messora, C.,Garagnani, L.,Bernardi,
S., Tamassia, M.G., and Sacchi, S.(1995). Bcl-2 oncoproteinexpres-
sion in acute myeloid leukemia. Haematologica 80, 98±102.
Bouscary, D., De Vos, J ., Guesnu, M., J ondeau, K., Viguier, F., Melle,
J ., Picard, F., Dreyfus, F., and Fontenay-Roupie, M. (1997). Fas/
Apo-1 (CD95) expression and apoptosis in patients with myelodys-
plastic syndromes. Leukemia 11, 839±845.
Cleary, M.L., and Sklar, J . (1985). Nucleotide sequence of a t(14;18)
chromosomal breakpoint in follicular lymphoma and demonstration
of a breakpoint-cluster region near a transcriptionally active locus
on chromosome 18. Proc. Natl. Acad. Sci. USA 82, 7439±7443.
Collins, S.J . (1987). The HL-60 promyelocytic leukemia cell line:
proliferation, differentiation, and cellular oncogene expression.
Blood 70, 1233±1244.
Cory, S., Harris, A.W., and Strasser, A. (1994). Insights from
transgenic mice regarding the role of bcl-2 innormal and neoplastic
lymphoid cells. Philos.Trans.R. Soc.Lond.B Biol.Sci.345, 289±295.
Daniel, P.T., and Krammer,P.H. (1994).Activationinduces sensitivity
toward APO-1 (CD95)-mediated apoptosis in human B cells. J . Im-
munol. 152, 5624±5632.
Delia, D., Aiello, A., Soligo, D., Fontanella, E., Melani, C., Pezzella,
F., Pierotti, M.A., and Della Porta, G. (1992). bcl-2 proto-oncogene
expression in normal and neoplastic human myeloid cells. Blood
Dirks, W., Schone, S., Uphoff, C., Quentmeier, H., Pradella, S., and
Drexler, H.G. (1997). Expression and function of CD95 (FAS/APO-1)
in leukaemia-lymphoma tumour lines. Br. J . Haematol. 96, 584±593.
Erickson, P.,Gao, J ., Chang, K.S.,Look, T., Whisenant,E., Raimondi,
S., Lasher, R., Trujillo, J ., Rowley, J ., and Drabkin, H. (1992). Identifi-
cation of breakpoints in t(8;21) acute myelogenous leukemia and
Blockade of Myeloid Cell Death Leads to AML
K.J ., Levine, H., and Ritz, J . (1995). Functional consequences of
APO-1/Fas (CD95) antigen expression by normal and neoplastic
hematopoietic cells. Leuk. Lymphoma 17, 51±61.
Russell, J .H., Rush, B., Weaver, C., and Wang, R. (1993). Mature T
cells ofautoimmune lpr/lprmice havea defectinantigen- stimulated
suicide. Proc. Natl. Acad. Sci. USA 90, 4409±4413.
Sawyers, C.L., Denny, C.T., and Witte, O.N. (1991). Leukemia and
the disruption of normal hematopoiesis. Cell 64, 337±350.
Schiffer, C.A. (1997). Acute myeloid leukemia in adults. In Cancer
Medicine, J .F. Holland, R.C Bast, D.L. Morton, E. Frei, D.W. Kufe,
and R.R. Weichselbaum, eds. (Baltimore, MD: Williams and Wilkins),
Singer, G., and Abbas, A. (1994). The fas antigen is involved in
peripheralbutnot thymic deletion ofT lymphocytes inT cellreceptor
transgenic mice. Immunity 1, 365±371.
Spangrude, G.J ., Heimfeld, S., and Weissman, I.L. (1988). Purifica-
tion and characterization of mouse hematopoietic stem cells Sci-
ence 241, 58±62.
Strasser, A., Harris, A.W., Huang, D.C., Krammer, P.H., and Cory,
S. (1995). Bcl-2 and Fas/APO-1 regulate distinct pathways to lym-
phocyte apoptosis. EMBO J . 14, 6136±6147.
Suda, T., and Nagata, S. (1994). Purification and characterization of
the Fas-ligand that induces apoptosis. J . Exp. Med. 179, 873±879.
Tenen, D.G., Hromas, R., Licht, J .D., and Zhang, D.E. (1997). Tran-
scription factors,normalmyeloiddevelopment, and leukemia.Blood
Till, J .E., and McCulloch, E.A. (1961). A direct measure of the radia-
tion sensitivity of normal mouse bone marrow cells. Radiat. Res.
Vaux, D.L., Cory, S., and Adams, J .M. (1988). Bcl-2 gene promotes
haemopoietic cell survivaland cooperates with c- myc to immortal-
ize pre-B cells. Nature 335, 440±442.
Wang, Q., Stacy, T., Binder, M., Marin-Padilla, M., Sharpe, A.H., and
Speck, N.A. (1996). Disruption of the Cbfa2 gene causes necrosis
and hemorrhaging in the central nervous system and blocks defini-
tive hematopoiesis. Proc. Natl. Acad. Sci. USA 93, 3444±3449.
Weinberg, R.A. (1991).Tumorsuppressorgenes. Science 254, 1138±
Wu, A.M., Till, J .E., Siminovitch, L., and McCulloch, E.A. (1967). A
cytological study of the capacity for differentiationof normal hemo-
poietic colony-forming cells. J . Cell. Physiol. 69, 177±184.
Yang, E., and Korsmeyer, S.J . (1996). Molecular thanatopsis: a dis-
course on the BCL2 family and cell death. Blood 88, 386±401.
Yergeau, D.A.,Hetherington, C.J ., Wang, Q., Zhang,P.,Sharpe, A.H.,
Binder, M., Marin-Padilla, M., Tenen, D.G., Speck, N.A., and Zhang,
D.E. (1997). Embryonic lethality and impairment of haematopoiesis
in mice heterozygous for an AML1-ETO fusion gene. Nat. Genet.