ã
Oncogene (2001) 20, 7178 ± 7185
2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00
www.nature.com/onc
Function of RARa during the maturation of neutrophils
Philippe Kastner1 and Susan Chan*,1
1
Institut de GeÂneÂtique et de Biologie MoleÂculaire et Cellulaire, CNRS-INSERM-ULP, 1 rue Laurent Fries, BP163, 67404 Illkirch
Cedex, C.U. de Strasbourg, France
The retinoic acid receptor a gene is the target of
chromosomal rearrangements in all cases of acute
promyelocytic leukemia (APL). This recurrent involvement of RARa in the pathogenesis of APL is likely to
re¯ect an important role played by this receptor during the
di€erentiation of immature myeloid cells to neutrophils.
RARa is a negative regulator of promyelocyte di€erentiation when not complexed with RA, and stimulates this
di€erentiation when bound to RA. Since RARs are
dispensable for the generation of mature neutrophils, their
role thus appears to be to modulatory, rather than
obligatory, for the control of neutrophil di€erentiation.
In vitro, retinoic acid is also a potent inducer of neutrophil
cell fate, suggesting that it might play a role in the
commitment of pluripotent hematopoietic progenitors to
the neutrophil lineage. Thus, the APL translocations target
an important regulator of myeloid cell di€erentiation.
Oncogene (2001) 20, 7178 ± 7185.
Keywords: retinoids; granulocyte; di€erentiation
The RARa gene is the target of chromosomal rearrangements in all cases of acute promyelocytic leukemia
(APL). These rearrangements lead to the production of
fusion proteins containing a variable N-terminal portion
originating from the PML, PLZF, NuMA, nucleophosmin or STAT5b proteins (reviewed in Melnick and Licht,
1999), and a ®xed C-terminal portion corresponding to
most of the RARa polypeptide (including the DNA
binding domain and ligand binding domain). That RARa
is consistently altered in APL raises the fundamental
question of whether, and how, the rearrangements may
perturb a previously unrecognized function of normal
RARa in controlling the proliferation of promyelocytes
and their di€erentiation into neutrophils. We will review
here the accumulating evidence suggesting that RAR and
retinoids may play several roles in the maturation of a
hematopoietic stem cell to a mature neutrophil.
Molecular aspects of RAR function
Although a detailed analysis of the molecular understanding of RAR function is not within the scope of
*Correspondence: P Kaster and S Chan;
E-mail: [email protected]
this review (the reader is referred to Chambon, 1996),
below is a brief summary of the current knowledge
which may be helpful in understanding the role of
RAR in the hematopoietic system.
RARa is a nuclear receptor for all-trans (ATRA)
and 9-cis (9C-RA) retinoic acid, and belongs to a
family of three closely-related receptors (RARa, b and
g). RARs are transcriptional regulators which bind to
speci®c retinoic acid response elements (RAREs)
present in the promoters of their target genes. RARs
do not bind RAREs on their own, but require
heterodimerization with an RXR (a, b or g) to do so.
RXRs are also nuclear receptors but bind only the 9cis isomer of retinoic acid. Extensive genetic analyses
both in cell lines and mice have demonstrated that the
RAR/RXR heterodimer is the functional unit responsible for the transduction of retinoid signals (Kastner et
al., 1997; Chiba et al., 1997; Mascrez et al., 1998).
The RAR/RXR complex has two main functions.
In the absence of ligand, the heterodimer binds the
corepressors SMRT or NCoR, which in turn interacts
with histone deacetylase-containing Sin3A complexes,
leading to the silencing of target genes (Nagy et al.,
1997). Binding of ligand provokes the dissociation of
these complexes and promotes the interaction of
RAR/RXR with coactivators that serve to recruit
complexes involved in transcriptional activation (for
review see Minucci and Pelicci, 1999; Glass and
Rosenfeld, 2000). The coactivators target the liganddependent activation function AF-2. How these
activators function initiating transcription is only
now becoming clear (Dilworth et al., 2000). Importantly, the AF-2 functions present in the RAR and
RXR moieties can synergize and play a crucial role in
mediating retinoid function in vivo, particularly during
development (Mascrez et al., 1998). In addition to
AF-2, the N-terminal portions of RAR and RXR
contain ligand-independent activation functions (AF-1;
Nagpal et al., 1993). With RXRa, genetic studies have
shown that the AF-1 function participates in mediating RAR/RXR activity in vivo (Mascrez et al., 2001).
The AF-1 function of RARa and RARg, which
involves the phosphorylation of a conserved serine
residue, also appears to be important for RAmediated di€erentiation of F9 embryonic carcinoma
cells (Taneja et al., 1997).
Each RAR encodes for at least two major isoforms
(RARa1 and RARa2 in the case of RARa, Leroy et
al., 1991) which di€er in their N-terminal A region.
These A regions contribute to the AF-1 function
RARa and neutrophil differentiation
P Kastner and S Chan
(Nagpal et al., 1992) and may therefore confer distinct
transcriptional properties to each RAR isoform.
RAR and RXR knock-out mice
Mice de®cient for each RAR gene have been produced.
RARa7/7 mice are viable and exhibit few obvious
defects, aside from abnormal spermatogenesis (Lufkin
et al., 1993). RARb7/7 and RARg7/7 mice also have
relatively few defects (Lohnes et al., 1993; Ghyselinck
et al., 1997). Animals bearing the targeted disruption of
single isoforms have also been generated, all of which,
with the exception of RARg17/7 mice, appear normal
(Lohnes et al., 1993; Lufkin et al., 1993; Subbarayan et
al., 1997). In contrast to the single mutants, double
homozygote animals exhibit a much larger array of
developmental defects, in most cases leading to
lethality, either in utero (the majority of
RARa7/7RARg7/7 mutants die during fetal development) or shortly after birth (RARa7/7/RARb7/7 and
RARb7/7RARg7/7 mutants; Lohnes et al., 1994;
Mendelsohn et al., 1994). Together, the defects in
these mutants recapitulate the complete spectrum of
the fetal vitamin A de®ciency syndrome established by
Warkany and coworkers more than half a century ago
(Kastner et al., 1995 for review and refs). However,
none of the RAR double mutants show the early
embryonic defects that occur when retinoic acid
synthesis is prevented, such as that seen after the
inactivation of the crucial retinaldehyde dehydrogenase
RALDH2 gene (Niederreither et al., 1999). These
results suggest that there is a large degree of functional
redundancy during development.
All three RXRs have also been knocked-out.
RXRa7/7 mice die during fetal development of cardiac
dysfunction (Kastner et al., 1994; Sucov et al., 1994),
but RXRb7/7 and RXRg7/7 animals are viable and
mostly normal (Kastner et al., 1996; Krezel et al.,
1996). The production of compound RXR/RAR
mutants (Kastner et al., 1994, 1997) has shown that
there is a strong synergy between RXRa and RAR
mutations, as RXRa/RAR double mutants recapitulate
most of the defects observed in RAR-only double
mutants. Recent work using spatio-temporally controlled mutagenesis of RXRa has also demonstrated an
important role for this receptor in several adult tissues,
including skin, liver and adipose tissue (Imai et al.,
2001a,b; Li et al., 2000). RXRb and RXRg appear to
be more limited in their function than RXRa and, to
date, have been implicated in spermatogenesis (Kastner
et al., 1996) and dopamine signaling in the striatum
(Krezel et al., 1998).
RARs are dispensable for granulopoiesis
RARa-de®cient mice have normal numbers of neutrophils in their peripheral blood and hematopoietic
organs (Kastner et al., 2001), demonstrating that RARa
is not required for neutrophil di€erentiation. In fact,
neutrophil development may take place in the
absence of any RAR, as granulopoiesis appears
to occur normally in RARa7/7RARg7/7 and
RARa17/7RARg7/7 double-de®cient fetuses (Kastner
et al., 2001; Labrecque et al., 1998). In vitro,
morphologically normal neutrophils can be generated
from the fetal liver cells of RARa7/7RARg7/7 mutants
(but apparently not from the bone marrow cells of
RARa17/7RARg7/7 fetuses, see below). As RARb is
not expressed in myeloid cells, the neutrophils in these
mutant animals do not express any RAR. It is
presently unknown if RAR function is dispensable
during adult granulopoiesis; the answer to this question
will come from the study of mice whose hematopoietic
system has been reconstituted with RARa7/7RARg7/7
double-mutant fetal liver cells.
Despite the lack of neutrophil defects in RARa7/7
mice, RARa appears to mediate some of the e€ects of
exogenous retinoids on myeloid cells (Kastner et al.,
2001). (i) RARa7/7 bone marrow cells produce normal
numbers of colonies when cultured in the presence of
9C-RA (with Epo, I1-3, IL-6 and SCF) while the
number of WT colonies is reduced by twofold. (ii) In
these same assays, the addition of 9C-RA also does not
induce the expected reduction in colony size. (iii)
Exogenous RA does not enhance granulocyte di€erentiation in RARa7/7 mice, both in vitro and in vivo.
These observations therefore suggest that the retinoid
signal is largely transduced through RARa in myeloid
cells. Hence, the presence of a normal neutrophil
population in RARa7/7 mice further suggests that
retinoic acid is not essential for neutrophil di€erentiation in adult mice.
7179
A role of RA in the commitment of hematopoietic
progenitors to the neutrophil lineage?
RA has the capacity to drive pluripotent hematopoietic
progenitors into the granulocyte lineage, as seen in
assays using the multipotent cell lines FDCPmix A4
(Tsai and Collins, 1993) and EML (a line that
overexpresses a dominant-negative RARa; Tsai et al.,
1994). RA can also suppress the formation of erythroid
and macrophage colonies, and induce a concomitant
increase in the number of neutrophil colonies, from
primary human progenitors (Gratas et al., 1993; van
Bockstaele et al., 1993; Labbaye et al., 1994; Tocci et
al., 1996). Importantly, Tocci et al. (1996) have shown
that this e€ect directly results from a cell-fate
conversion by RA. In their experiments, single
CD34+ fetal liver cells were plated in methylcellulosecontaining medium in the presence of Epo, IL3 and
SCF; 2 days later, individual daughter cells were
replated in a similar medium in the presence or absence
of RA. Invariably, progenitors that normally gave rise
to erythroid or macrophage colonies in the absence of
RA di€erentiated into granulocytes in the presence of
all-trans RA. Similarly, Paul et al. (1995) showed that
cells from the eosinophil precursor cell line
AML14.3D10 are redirected to become neutrophils
Oncogene
RARa and neutrophil differentiation
P Kastner and S Chan
7180
by RA. These results indicate that RA can reprogram
hematopoietic progenitors already engaged in a nongranulocyte lineage to the neutrophil fate.
It is unclear whether RA acts endogenously as a
speci®cation factor for the neutrophil lineage. To date,
no evidence for such a role has been described. In
contrast, several observations suggest otherwise: (i)
vitamin A-de®ciency leads to an increased number of
neutrophils in mice (Kuwata et al., 2000; Kastner et al.,
2001), an observation incompatible with a major role
for retinoids in specifying neutrophil fate; (ii) similar
numbers of granulocyte, macrophage and mixed
granulocyte/macrophage colonies are grown from
progenitors of normal or VAD mice (Kuwata et al.,
2000), (iii) competitive reconstitution experiments with
50 : 50 mixes of RARa7/7 and WT bone marrow cells
do not reveal a disadvantage of RARa7/7 cells in
contributing to the neutrophil lineage (our unpublished
results). Nonetheless, the potent activity of RA in
inducing neutrophil commitment in vitro remains
intriguing, and further research is required to assess
its possible relevance in vivo.
RARa is a bidrectional regulator of neutrophil
di€erentiation
RA has also been shown to promote the di€erentiation
of normal, as well as leukemic, immature myeloid cells
already engaged in the neutrophil lineage. This was
®rst demonstrated for the HL60 promyelocytic cell line
(Breitman et al., 1980) and later extended to primary
promyelocytes (Gratas et al., 1993). Moreover, the
addition of a RAR antagonist (BMS493) to cultures of
bone marrow cells in the presence of G-CSF and SCF
results in a slowdown of neutrophil di€erentiation
(Kastner et al., 2001), suggesting that serum retinoids
normally play a role in the positive control of this
di€erentiation. The role of endogenous retinoids in
controlling neutrophil di€erentiation in vivo has been
more elusive. The best evidence has come from data
looking at the short-term e€ects of BMS493 in vivo
which, when systemically administered to mice for 3
days, leads to an accumulation of immature granulocytes in the bone marrow, as assessed by Mac-1 and
Gr-1 expression, and cell morphology (Kastner et al.,
2001). A similar increase in the proportion of immature
neutrophils was seen in vitamin A-de®cient mice
(Kastner et al., 2001). Even though VAD is long to
establish, often leading to secondary phenotypes, the
similarity between the e€ects of VAD and BMS493
administration strongly suggests that retinoids are
involved in controlling the di€erentiation of neutrophil
under physiological steady state settings.
If indeed RA plays an important role in controlling
neutrophil di€erentiation, it might seem surprising,
even contradictory, that this di€erentiation occurs
rather normally in RARa7/7 mice. This paradox may
be explained by the observation that unliganded RARa
is an active mediator of the e€ects of RA de®ciency.
Several experiments have demonstrated that unli-
Oncogene
ganded RARa can antagonize di€erentiation. Overexpression of WT RARa1 in immature hematopoietic
cells readily blocks cytokine-induced neutrophil di€erentiation at the promyelocyte stage in the absence of
exogenous RA (Onodera et al., 1995; Du et al., 1999).
The inhibition observed here is reminiscent of the
block seen when a C-terminally-truncated `dominantnegative' RARa (RARa403) lacking the ligand-dependent AF-2 function is overexpressed in immature
myeloid cells. Collins and coworkers have shown that
overexpression of RARa403 can immortalize hematopoietic progenitors at the promyelocyte stage (leading
to the generation of the MPRO cell line; Tsai et al.,
1994) and at the multipotent progenitor stage (leading
to the generation of the EML cell line; Tsai et al.,
1994). The interpretation of these results has generally
pointed to a role for RA in neutrophil di€erentiation.
However, it is equally possible that the dominantnegative RARa described above functions like WT
RARa, although the activity of the former cannot be
alleviated by the low levels of retinoids present in the
culture medium. In addition, ex vivo neutrophil
di€erentiation occurs faster for BM cells derived from
RARa7/7 compared to WT mice (Kastner et al., 2001).
Thus, loss of RARa removes an activity that maintains
myeloid cells at an immature stage. Collectively, these
observations indicate that RARa is a negative
regulator of granulocyte di€erentiation. Of special
signi®cance, treatment of WT mice with a saturating
dose of ATRA leads to a decrease in the size of the
most immature (Mac-1+Gr-17/lo) and an increase in
that of the intermediate (Mac-1+Gr-1lo/+) granulocyte
populations (Kastner et al., 2001), suggesting that not
all of the RAR molecules are liganded under normal
circumstances in vivo, and are therefore free to act as
negative regulators.
Thus, RARa is never a neutral bystander during
neutrophil di€erentiation: it inhibits di€erentiation
when not bound to retinoic acid and promotes
di€erentiation when bound to RA. It is therefore clear
why ligand de®ciency (as in VAD) or the use of RAR
antagonists is not functionally equivalent to lack of the
RAR itself. In the former, only the ligand-dependent
di€erentiating activity is abolished while the blocking
activity is maintained and perhaps even enhanced. This
model predicts that a functional RAR is required for
the full e€ects of vitamin A de®ciency on di€erentiation, a hypothesis that we are currently testing by
generating VAD RARa7/7 mice. Interestingly, a
similar scenario has been observed with thyroid
hormone receptor b (TRb). Mice de®cient for TRb
do not exhibit signs of hypothyroidism, while animals
carrying a point mutation speci®cally abbrogating the
ability of TRb to bind thyroid hormone do (Hashimoto et al., 2001). Together, these results show that
nuclear receptor knock-outs may not always be the
best models for studying the role of the cognate ligand,
especially when the ligand functions in part to relieve
the repression exerted by the unliganded receptor.
Improved mutations that knock-out the speci®c ability
to carry out ligand-dependent and -independent
RARa and neutrophil differentiation
P Kastner and S Chan
functions are required to understand the physiological
roles of these receptors.
Functional speci®city for RARa in granulopoiesis?
Although both RARa and RARg are expressed in
granulocytes (Labrecque et al., 1998), our data suggest
that RARa7/7 cells lack both the negative and positive
RAR-dependent activities in this lineage. RARa7/7
cells are resistant to RA and to the RAR antagonist
BMS493. RARa therefore seems to be the major RAR
operating during granulocyte di€erentiation. Is this due
to a functional speci®city of this particular receptor or
does it merely re¯ect its relative abundance in myeloid
cells?
There is evidence that RARa is functionally speci®c.
Du et al. (1999) have shown, using retroviral delivery
of various RARs into immature hematopoietic progenitors (i.e. lineage-depleted bone marrow cells from
5-¯uorouracil-treated mice), that overexpression of
RARa, but not RARb or RARg, inhibits granulocyte
di€erentiation of transduced cells cultured in the
presence of IL-3, IL-6, GM-CSF and SCF. In addition,
only RARa is able to induce proliferation of primary
myeloid progenitors (as measured by the potential to
form colonies after three passages in methylcellulose
culture). These observations show that only unliganded
RARa can inhibit granulocyte di€erentiation.
RARa-speci®city appears to extend to retinoic acidmediated e€ects as well. Mehta and coworkers (Drach
et al., 1994; Mehta et al., 1997) studied RA-dependent
transcription of the CD38 gene in RAR-transduced
HL60R cells (which are resistant to retinoids due to the
expression of a truncated dominant-negative RARa);
only transfected RARa, not RARb or g, was able to
rescue RA-induced CD38 transcription. In addition,
Du et al. (1999) have shown that immature bone
marrow progenitors transduced with high amounts of
RARa and RARg (but not RARb) are highly sensitive
to the antiproliferative activity of RA. Together, these
observations indicate that RARa is functionally the
best RAR to mediate both the ligand-dependent and independent activities of RARs during neutrophil
maturation.
The apparent high degree of speci®city in the
granulocyte lineage stands in sharp contrast to the
large functional redundancy observed from extensive
genetic studies of RAR function during development,
where most RAR activities require the combined
removal of two receptors to be abbrogated (reviewed
in Kastner et al., 1995). These observations have
suggested that RARs may perform broadly similar
functions and can readily substitute for each other,
provided their expression levels are adequate. A similar
redundancy has been demonstrated in F9 embryonic
carcinoma cells (Rochette-Egly and Chambon, 2001 for
review). Interestingly, some functional speci®city is
observed in this system as well. Although RARg (the
most abundantly expressed RAR in F9 cells) is
required for all retinoid functions, RARa plays a
crucial and unique role, which cannot be substituted
for by RARg, during the parietal di€erentiation
produced by the combined action of RA and cAMP
(Taneja et al., 1997). Furthermore, RARa and RARg
are also markedly di€erent in their ability to induce
transcription of certain target genes in F9 cells (Chiba
et al., 1997).
7181
Do RARa1 and RARa2 play unique roles during
granulocyte di€erentiation?
Both RARa1 and RARa2 are expressed in cells of the
neutrophil lineage (Labrecque et al., 1998; A Zelent,
personal communication), although it is unclear what
their speci®c roles are in these cells. One clue may lie in
the con¯icting data produced from our laboratory
(Kastner et al., 2001) and those of Labrecque et al.
(1998). We reported that myeloid progenitors from
RARa7/7 adult bone marrow or RARa7/7RARg7/7
fetal liver are not blocked in their di€erentiation (and
in fact di€erentiate with a faster kinetics than WT
cells). In contrast, Labrecque et al. (1998) showed that
granulocyte progenitors from RARa17/7RARg7/7
18.5 d.p.c. fetal bone marrow are mostly blocked at
the myelocyte/metamyelocyte stage of di€erentiation
when cultured with SCF, IL3 and Epo; further
di€erentiation was induced by the addition of RA.
Since RARa2 is presumably still expressed in
RARa17/7RARg7/7 myeloid cells, RARa1 and RARa2
might perform distinct functions in promyelocytes. The
results from Labrecque et al. (1998) suggest that RARa2
is a more potent inhibitor of di€erentiation than
RARa1, but that RARa1 may be better at mediating
RA-induced di€erentiation. It might therefore be
interesting to investigate whether the two RARa
isoforms di€erentially regulate di€erentiation and
proliferation in the presence and absence of ligand.
Apoptosis
Retinoids have been implicated in the apoptotic
process of myeloid cell lines like HL60 (Ueno et al.,
1998; Monczak et al., 1997). They may also control the
rate of granulocyte apoptosis in vivo; Kuwata et al.
(2000) have shown that there is a clear reduction in the
rate of apoptosis of granulocytes from vitamin Ade®cient mice. However, as apoptosis is the endpoint
of neutrophil maturation, the retinoid-triggered apoptotic response might just correspond to one facet of the
di€erentiation program controlled by retinoids. Some
studies, using myeloid cell lines, have shown, however,
that the di€erentiation and apoptotic responses elicited
by RA can be separated. In the leukemic cell line PLB985, Monczak et al. (1997) showed that RA provokes
apoptosis without triggering di€erentiation, and that
activation of both RAR and RXR was required.
Likewise, activation of RXR leads to apoptosis in the
absence of di€erentiation in HL60R cells (Mehta et al.,
1996). Interestingly, these studies have also pointed to
Oncogene
RARa and neutrophil differentiation
P Kastner and S Chan
7182
a critical role for RXR activation, in agreement with
data from Boehm et al. (1995) who reported that
RXR-selective agonists trigger apoptosis in human
myeloid cell lines.
At the mechanistic level, it was recently shown that
RA induction of TRAIL, a cell surface ligand for
death receptors, is reponsible for the RA-induced
apoptosis of NB4 and PLB-985 cells (Altucci et al.,
2001). As TRAIL induction is a rather late event
occurring only after 48 h of RA stimulation, it is
unlikely that TRAIL is directly induced by the RAR/
RXR complex. Curiously, earlier events following RA
stimulation are the induction of several anti-apoptotic
pathways, including the strong induction of the BCL2
homolog BCL2A1, and induction of NFkB activity
(Altucci et al., 2001; Liu et al., 2000). Thus, even
though the endpoint will be death, retinoids may
modulate survival and death of myeloid cells in a
temporally complex manner.
Stem cell maintenance
Purton et al. (1999, 2000) have recently shown that RA
is potent in maintaining the `stem cell-ness' of murine
Lin7c-kit+Sca-1+ cells (a population highly enriched
for hematopoietic stem cells (HSC)) in vitro, mainly by
blocking their maturation. Cells cultured in the
presence of ATRA for 14 days retain their short- and
long-term capacity to repopulate lethally-irradiated
recipients, while those cultured without RA do not.
Moreover, adding an RAR antagonist to the cultures
arrests all repopulating activity, its primary e€ect being
to induce maturation, as the number of more mature
CFU-S cells increased in these cultures. Studies by
Zauli et al. (1995) have suggested that RA also acts
eciently to promote stem cell survival. Using puri®ed
CD34+ human cord blood cells, these authors showed
that ATRA (at both low (10710 M) or high (1076 M)
doses) can prevent apoptosis induced by serum or
growth factor starvation. It is currently unknown
which RARs are involved in these processes. Interestingly, the data presented by the studies suggest that
RA inhibits maturation in early pluripotent stem cells
in a manner opposite to that observed at later stages
where it acts to promote di€erentiation and apoptosis.
An important question is whether retinoic acid plays a
physiological role in controlling HSC maturation;
quanti®cation of these cells in RAR knock-out mice,
as well as in VAD mice, will be required to assess the
role of retinoid signaling in the maintenance of the
HSC pool in vivo.
Target genes and mechanisms
Understanding how RARa controls granulopoiesis at
the molecular level will require identi®cation of critical
target genes regulated both by the unliganded and
liganded receptor. Several candidate target genes have
been identi®ed.
Oncogene
CAAT/enhancer binding proteins (C/EBPs)
Perhaps one of the most promising candidates is
C/EBPe. C/EBPe transcription is rapidly induced by
retinoic acid in the absence of protein synthesis in
HL60 and NB4 cells (Chih et al., 1997; Liu et al.,
2000), and a functional RARE has been identi®ed in
the C/EBPe promoter (Park et al., 1999). C/EBPe is
therefore likely to be a direct RAR target gene in
granulocytes. In addition, C/EBPe is a critical regulator
of terminal granulocyte di€erentiation (Yamanaka et
al., 1997). However, C/EBPe is expressed normally in
RARa7/7 mutants (our unpublished results), a fact
which is not too surprising given the normal neutrophil
di€erentiation that occurs in these mice. It is possible
that C/EBPe transcription is repressed by unliganded
RARa, an interesting possiblity that deserves further
investigation. Since C/EBPe is required at a rather late
stage in the di€erentiation (around the metamyelocyte
stage), this target gene cannot account for the block at
the promyelocyte stage that results from overexpressed
RARa. Interestingly, another member of the CAAT/
enhancer binding protein family, GADD153 (also
named CHOP) has been found to be induced by RA
in a cycloheximide-insensitive manner in NB4 cells (Liu
et al., 2000). These results point to genes encoding C/
EBP factors as important RAR target genes during
granulocyte di€erentiation.
c-myc
As c-myc expression is down-regulated by RA in
NB4 cells in a cycloheximide-independent fashion
(Liu et al., 2000), it may therefore be a primary
RAR target gene in early myeloid cells. Interestingly,
the activities mediated by RARa is reminiscent of
that exerted by the Myc/Mad system. A switch from
the binding of the Myc/Max to the Mad/Max
heterodimer on target genes has been shown to be
critical in a cell's decision to proliferate or di€erentiate, particularly in the myeloid lineage (Ayer and
Eisenman, 1993; Hurlin et al., 1995; Cultraro et al.,
1997; Foley et al., 1998). In Mad17/7 mice,
granulocyte di€erentiation proceeds with slower kinetics than in WT animals, opposite to that seen in
RARa7/7 mutants (Foley et al., 1998). A direct role for
c-myc in mediating RA function has been shown in F9
cells where c-myc transcription is down-regulated by
RA. This down-regulation appears to be essential for
di€erentiation, as inhibition of c-myc expression by
antisense oligonucleotides results in the same e€ect
(Griep and Westphal, 1988).
p21
The cyclin-dependent kinase inhibitor p21 (WAF/
CIP1) is a well-studied target gene of p53 and is
important in mediating cell cycle arrest. In the
monocyte cell line U937, p21 is induced by vitamin
D3; this induction is essential for vitamin D3-induced
di€erentiation (Liu et al., 1996b). p21 is probably
RARa and neutrophil differentiation
P Kastner and S Chan
also a direct target gene of RA, and is induced by
retinoids in U937 (Liu et al., 1996a), HL60 (Jiang et
al., 1994) and NB4 cells (Liu et al., 2000). In NB4
cells, p21 induction does not require protein synthesis
(Liu et al., 2000). Furthermore a functional RARE
has been found in the p21 promoter (Liu et al.,
1996a). Given the proposed role of p21 in vitamin
D3-induced monocytic di€erentiation, it is tempting
to speculate that it might be equally important in
mediating RA activity during neutrophil di€erentiation.
Other genes
Other genes have been identi®ed to be regulated by
RA in a number of myeloid cell lines. Liu et al. (2000)
have come up with a large number of genes which are
regulated by RA, using the combined approaches of
suppression-subtractive hybridization, di€erential display and microarrays. Other types of studies have also
identi®ed target genes (Matikainen et al., 1996;
Lawson et al., 1999; for review, see Lawson and
Berliner, 1999). Interestingly, a large number of genes
encode for proteins of regulatory pathways common
to all cells. However, since many of them were
identi®ed in cells expressing abnormal RARs (NB4,
MPRO), caution should be exercised with respect to
their relevance in the control of granulopoiesis by
endogenous RARs. Furthermore, all of the target
genes identi®ed to date are regulated by RA. To our
knowledge, no gene has been shown to be repressed
by the unliganded receptor. In fact, data obtained
from the study of F9 cells de®cient for RAR/RXR
have not shown an upregulation of target genes in the
absence of RA (Chiba et al., 1997). Repression by
unliganded RAR may therefore a€ect only a small
fraction of RAR target genes. Nonetheless the
identi®cation of these genes is critical to understanding how RARa regulates granulopoiesis.
Implications for APL and beyond
The present knowledge of the role of endogenous
RARs in the control of granulocyte di€erentiation may
shed some light on the function of translocated RARa
in APL. There is now solid evidence that the fusion
proteins generated in APL possess strong repressive
properties in the absence of RA (He et al., 1998; Hong
et al., 1997; Lin et al., 1998; Grignani et al., 1998;
Guidez et al., 1998), and that these properties are
crucial for the leukemic phenotype. Since repression by
RARa plays a normal function in granulopoiesis, it
appears that the fusion proteins merely enhance this
physiological activity. This enhancement may be
brought about by two distinct mechanisms. First, the
fusion proteins are intrinsically more potent at
repression than WT RARa, as they are either
completely refractory to (PLZF-RAR), or require
higher concentrations of (PML-RAR), RA, thus
making physiological levels of RA ine€ective at
overcoming the di€erentiation block exerted by these
proteins. Second, the fusion proteins are usually
expressed at higher levels than WT RARa in APL
cells (Kastner et al., 1992; Pandol® et al., 1992), a
factor that probably contributes to their dominant
activity. Since RARa is the only RAR that can mediate
all the ligand-independent and -dependent activities
during granulocyte di€erentiation, it is likely that the
APL fusion proteins e€ectively antagonize di€erentiation by amplifying a normal function of RARa as a
negative regulator.
Understanding how RARa controls granulocyte
di€erentiation may also help in understanding how
RARs, or related nuclear receptors, control di€erentiation in other cellular systems. An altered thyroid
hormone receptor a (TRa), v-erbA, is an oncogene
carried by the avian erythroblastosis virus; it functions
to inhibit the di€erentiation of erythroid progenitors
(Metz and Graf, 1992 for review). Overexpression of
WT TRa inhibits the di€erentiation of erythroblasts
(Bauer et al., 1998). Thyroid hormone may also be
implicated in erythroid di€erentiation as it has been
identi®ed as the serum component required for
spontaneous di€erentiation of an erythroleukemia cell
line (Dinnen et al., 1994). However, TRa7/7 mice
appear to have normal erythropoiesis (Fraichard et al.,
1997), similar to the normal granulopoiesis seen in
RARa7/7 mice. These observations suggest that TRa
might play a similar dual role in the control of
erythroid di€erentiation, for which the negative
regulatory part may be enhanced by v-erbA. PPARg,
a nuclear receptor closely related to RAR, may play a
similar role in adipocyte di€erentiation (for review see
Fajas et al., 1998); similarities in this system include a
dynamic regulation of the Myc/Max and Mad/Max
complexes (Pulverer et al., 2000) and the involvement
of C/EBP proteins (Lane et al., 1999). Finally,
oligodendrocyte di€erentiation (involving RAR and
TR, see Rodriguez-Pena, 1999 for review), and
monocyte di€erentiation (involving VDR) may involve
the same mechanisms. Nuclear receptors may thus be
part of a conserved molecular framework adopted by
di€erent cell types to modulate their di€erentiation.
Unraveling how RARa controls granulocyte di€erentiation at the molecular level is likely to have a broad
signi®cance.
7183
Acknowledgments
The work of our laboratory was supported by the
Association pour la Recherche sur le Cancer, the Centre
National pour la Recherche Scienti®que, the Institut
National de la Sante et de la Recherche MeÂdicale, the
HoÃpital Universitaire de Strasbourg, the ColleÁge de France
and Bristol-Myers Squibb.
Oncogene
RARa and neutrophil differentiation
P Kastner and S Chan
7184
References
Altucci L, Rossin A, Ra€elsberger W, Reitmair A,
Chomienne C and Gronemeyer H. (2001). Nature Med.,
7, 680 ± 686.
Ayer DE and Eisenman RN. (1993). Genes Dev., 7, 2110 ±
2119.
Bauer A, Mikulits W, Lagger G, Stengl G, Brosch G and
Beug H. (1998). EMBO J., 17, 4291 ± 4303.
Boehm MF, Zhang L, Zhi L, McClurg MR, Berger E,
Wagoner M, Mais DE, Suto CM, Davies JA and Heyman
RA. (1995). J. Med. Chem., 38, 3146 ± 3155.
Breitman TR, Selonick SE and Collins SJ. (1980). Proc. Natl.
Acad. Sci. USA, 77, 2936 ± 2940.
Chambon P. (1996). FASEB J., 10, 940 ± 954.
Chiba H, Cli€ord J, Metzger D and Chambon P. (1997).
Mol. Cell Biol., 17, 3013 ± 3020.
Chih DY, Chumakov AM, Park DJ, Silla AG and Koe‚er
HP. (1997). Blood, 90, 2987 ± 2994.
Cultraro CM, Bino T and Segal S. (1997). Mol. Cell Biol., 17,
2353 ± 2359.
Dilworth FJ, Fromental-Ramain C, Yamamoto K and
Chambon P. (2000). Mol. Cell, 6, 1049 ± 1058.
Dinnen RD, White SR, Elsayed S, Yeh YI and Ebisuzaki K.
(1994). Cell Growth Di€er., 5, 855 ± 861.
Drach J, McQueen T, Engel H, Andree€ M, Robertson KA,
Collins SJ, Malavasi F and Mehta K. (1994). Cancer Res.,
54, 1746 ± 1752.
Du C, Redner RL, Cooke MP and Lavau C. (1999). Blood,
94, 793 ± 802.
Fajas L, Fruchart JC and Auwerx J. (1998). Curr. Opin. Cell.
Biol., 10, 165 ± 173.
Foley KP, McArthur GA, Queva C, Hurlin PJ, Soriano P
and Eisenman RN. (1998). EMBO J., 17, 774 ± 785.
Fraichard A, Chassande O, Plateroti M, Roux JP, Trouillas
J, Dehay C, Legrand C, Gauthier K, Kedinger M, Malaval
L, Rousset B and Samarut J. (1997). EMBO J., 16, 4412 ±
4420.
Ghyselinck NB, Dupe V, Dierich A, Messaddeq N, Garnier
JM, Rochette-Egly C, Chambon P and Mark M. (1997).
Int. J. Dev. Biol., 41, 425 ± 447.
Glass CK and Rosenfeld MG. (2000). Genes Dev., 14, 121 ±
141.
Gratas C, Menot ML, Dresch C and Chomienne C. (1993).
Leukemia, 7, 1156 ± 1162.
Griep AE and Westphal H. (1988). Proc. Natl. Acad. Sci.
USA, 85, 6806 ± 6810.
Grignani F, De Matteis S, Nervi C, Tomassoni L, Gelmetti
V, Cioce M, Fanelli M, Ruthardt M, Ferrara FF, Zamir I,
Seiser C, Lazar MA, Minucci S and Pelicci PG. (1998).
Nature, 391, 815 ± 818.
Guidez F, Ivins S, Zhu J, Soderstrom M, Waxman S and
Zelent A. (1998). Blood, 91, 2634 ± 2642.
Hashimoto H, Curty F, Borges P, Lee CE, Abel ED,
Elmquist JK, Cohen RN and Wondisdorf F. (2001). Proc.
Natl. Acad. Sci. USA, 98, 3998 ± 4003.
He LZ, Guidez F, Tribioli C, Peruzzi D, Ruthardt M, Zelent
A and Pandol® PP. (1998). Nat. Genet., 18, 126 ± 135.
Hong SH, David G, Wong CW, Dejean A and Privalsky ML.
(1997). Proc. Natl. Acad. Sci. USA, 94, 9028 ± 9033.
Hurlin PJ, Queva C, Koskinen PJ, Steingrimsson E, Ayer
DE, Copeland NG, Jenkins NA and Eisenman RN.
(1995). EMBO J., 14, 5646 ± 5659.
Imai T, Jiang M, Chambon P and Metzger D. (2001a). Proc.
Natl. Acad. Sci. USA, 98, 224 ± 228.
Imai T, Jiang M, Kastner P, Chambon P and Metzger D.
(2001b). Proc. Natl. Acad. Sci. USA, 98, 4581 ± 4586.
Oncogene
Jiang H, Lin J, Su ZZ, Collart FR, Huberman E and Fisher
PB. (1994). Oncogene, 9, 3397 ± 3406.
Kastner P, Grondona JM, Mark M, Gansmuller A, LeMeur
M, Decimo D, Vonesch JL, Dolle P and Chambon P.
(1994). Cell, 78, 987 ± 1003.
Kastner P, Lawrence HJ, Waltzinger C, Ghyselinck N,
Chambon P and Chan S. (2001). Blood, 97, 1314 ± 1320.
Kastner P, Mark M and Chambon P. (1995). Cell, 83, 859 ±
869.
Kastner P, Mark M, Ghyselinck N, Krezel W, Dupe V,
Grondona JM and Chambon P. (1997). Development, 124,
313 ± 326.
Kastner P, Mark M, Leid M, Gansmuller A, Chin W,
Grondona JM, Decimo D, Krezel W, Dierich A and
Chambon P. (1996). Genes Dev., 10, 80 ± 92.
Kastner P, Perez A, Lutz Y, Rochette-Egly C, Gaub MP,
Durand B, Lanotte M, Berger R and Chambon P. (1992).
EMBO J., 11, 629 ± 642.
Krezel W, Dupe V, Mark M, Dierich A, Kastner P and
Chambon P. (1996). Proc. Natl. Acad. Sci. USA, 93,
9010 ± 9014.
Krezel W, Ghyselinck N, Samad TA, Dupe V, Kastner P,
Borrelli E and Chambon P. (1998). Science, 279, 863 ± 867.
Kuwata T, Wang IM, Tamura T, Ponnamperuma RM,
Levine R, Holmes KL, Morse HC, De Luca LM and
Ozato K. (2000). Blood, 95, 3349 ± 3356.
Labbaye C, Valtieri M, Testa U, Giampaolo A, Meccia E,
Sterpetti P, Parolini I, Pelosi E, Bulgarini D and Cayre
YE. (1994). Blood, 83, 651 ± 656.
Labrecque J, Allan D, Chambon P, Iscove NN, Lohnes D
and Hoang T. (1998). Blood, 92, 607 ± 615.
Lane MD, Tang QQ and Jiang MS. (1999). Biochem.
Biophys. Res. Commun., 266, 677 ± 683.
Lawson ND and Berliner N. (1999). Exp. Hematol., 27,
1355 ± 1367.
Lawson ND, Zain M, Zibello T, Picciotto MR, Nairn AC
and Berliner N. (1999). Exp. Hematol., 27, 1682 ± 1690.
Leroy P, Krust A, Zelent A, Mendelsohn C, Garnier JM,
Kastner P, Dierich A and Chambon P. (1991). EMBO J.,
10, 59 ± 69.
Li M, Indra AK, Warot X, Brocard J, Messaddeq N, Kato S,
Metzger D and Chambon P. (2000). Nature, 407, 633 ±
636.
Lin RJ, Nagy L, Inoue S, Shao W, Miller Jr WH and Evans
RM. (1998). Nature, 391, 811 ± 814.
Liu M, Lavarone A and Freedman LP. (1996a). J. Biol.
Chem., 271, 31723 ± 31728.
Liu M, Lee MH, Cohen M, Bommakanti M and Freedman
LP. (1996b). Genes Dev., 10, 142 ± 153.
Liu TX, Zhang JW, Tao J, Zhang RB, Zhang QH, Zhao CJ,
Tong JH, Lanotte M, Waxman S, Chen SJ, Mao M, Hu
GX, Zhu L and Chen Z. (2000). Blood, 96, 1496 ± 1504.
Lohnes D, Kastner P, Dierich A, Mark M, LeMeur M and
Chambon P. (1993). Cell, 73, 643 ± 658.
Lohnes D, Mark M, Mendelsohn C, Dolle P, Dierich A,
Gorry P, Gansmuller A and Chambon P. (1994).
Development, 120, 2723 ± 2748.
Lufkin T, Lohnes D, Mark M, Dierich A, Gorry P, Gaub
MP, LeMeur M and Chambon P. (1993). Proc. Natl. Acad.
Sci. USA, 90, 7225 ± 7229.
Mascrez B, Mark M, Dierich A, Ghyselinck NB, Kastner P
and Chambon P. (1998). Development, 125, 4691 ± 4707.
Mascrez B, Mark M, Krezel W, Dupe V, Lemeur M,
Ghyselinck N and Chambon P. (2001). Development,
128, 2049 ± 2062.
RARa and neutrophil differentiation
P Kastner and S Chan
Matikainen S, Ronni T, Hurme M, Pine R and Julkunen I.
(1996). Blood, 88, 114 ± 123.
Mehta K, McQueen T, Manshouri T, Andree€ M, Collins S
and Albitar M. (1997). Blood, 89, 3607 ± 3614.
Mehta K, McQueen T, Neamati N, Collins S and Andree€
M. (1996). Cell Growth Di€er., 7, 179 ± 186.
Melnick A and Licht JD. (1999). Blood, 93, 3167 ± 3215.
Mendelsohn C, Lohnes D, Decimo D, Lufkin T, LeMeur M,
Chambon P and Mark M. (1994). Development, 120,
2749 ± 2771.
Metz T and Graf T. (1992). Oncogene, 7, 597 ± 605.
Minucci S and Pelicci PG. (1999). Semin. Cell. Dev. Biol., 10,
215 ± 225.
Monczak Y, Trudel M, Lamph WW and Miller Jr WH.
(1997). Blood, 90, 3345 ± 3355.
Nagpal S, Friant S, Nakshatri H and Chambon P. (1993).
EMBO J., 12, 2349 ± 2360.
Nagpal S, Zelent A and Chambon P. (1992). Proc. Natl.
Acad. Sci. USA, 89, 2718 ± 2722.
Nagy L, Kao HY, Chakravarti D, Lin RJ, Hassig CA, Ayer
DE, Schreiber SL and Evans RM. (1997). Cell, 89, 373 ±
380.
Niederreither K, Subbarayan V, Dolle P and Chambon P.
(1999). Nat. Genet., 21, 444 ± 448.
Onodera M, Kunisada T, Nishikawa S, Sakiyama Y and
Matsumoto S. (1995). Oncogene, 11, 1291 ± 1298.
Pandol® PP, Alcalay M, Fagioli M, Zangrilli D, Mencarelli
A, Diverio D, Biondi A, Lo Coco F, Rambaldi A and
Grignani F. (1992). EMBO J., 11, 1397 ± 1407.
Park DJ, Chumakov AM, Vuong PT, Chih DY, Gombart
AF, Miller Jr WH and Koe‚er HP. (1999). J. Clin. Invest.,
103, 1399 ± 1408.
Paul CC, Mahrer S, Tolbert M, Elbert BL, Wong I,
Ackerman SJ and Baumann MA. (1995). Blood, 86,
3737 ± 3744.
Pulverer B, Sommer A, McArthur GA, Eisenman RN and
Luscher B. (2000). J. Cell. Physiol., 183, 399 ± 410.
Purton LE, Bernstein ID and Collins SJ. (1999). Blood, 94,
483 ± 495.
Purton LE, Bernstein ID and Collins SJ. (2000). Blood, 95,
470 ± 477.
Rochette-Egly C and Chambon P. (2001). Histol. Histopathol., 16, 902 ± 922.
Rodriguez-Pena A. (1999). J. Neurobiol., 40, 497 ± 512.
Subbarayan V, Kastner P, Mark M, Dierich A, Gorry P and
Chambon P. (1997). Mech. Dev., 66, 131 ± 142.
Sucov HM, Dyson E, Gumeringer CL, Price J, Chien KR
and Evans RM. (1994). Genes Dev., 8, 1007 ± 1018.
Taneja R, Rochette-Egly C, Plassat JL, Penna L, Gaub MP
and Chambon P. (1997). EMBO J., 16, 6452 ± 6465.
Tocci A, Parolini I, Gabbianelli M, Testa U, Luchetti L,
Samoggia P, Masella B, Russo G, Valtieri M and Peschle
C. (1996). Blood, 88, 2878 ± 2888.
Tsai S, Bartelmez S, Sitnicka E and Collins S. (1994). Genes
Dev., 8, 2831 ± 2841.
Tsai S and Collins SJ. (1993). Proc. Natl. Acad. Sci. USA, 90,
7153 ± 7157.
Ueno H, Kizaki M, Matsushita H, Muto A, Yamato K,
Nishihara T, Hida T, Yoshimura H, Koe‚er HP and
Ikeda Y. (1998). Leuk. Res., 22, 517 ± 525.
van Bockstaele DR, Lenjou M, Snoeck HW, Lardon F,
Stryckmans P and Peetermans ME. (1993). Ann. Hematol.,
66, 61 ± 66.
Yamanaka R, Barlow C, Lekstrom-Himes J, Castilla LH,
Liu PP, Eckhaus M, Decker T, Wynshaw-Boris A and
Xanthopoulos KG. (1997). Proc. Natl. Acad. Sci. USA,
94, 13187 ± 13192.
Zauli G, Visani G, Vitale M, Gibellini D, Bertolaso L and
Capitani S. (1995). Br. J. Haematol., 90, 274 ± 282.
7185
Oncogene