(Received for publication, October 1, 1996, and in revised form, February 17, 1997)
From the Department of Biochemistry, Trinity College, Dublin, Ireland and § CAM Research Department, ZENECA Pharmaceuticals, Mereside, Alderley Park, Macclesfield, Cheshire SK10 4TG, United Kingdom
The anthracycline antibiotic,
daunorubicin, can induce programmed cell death (apoptosis) in cells.
Recent work suggests that this event is mediated by ceramide via
enhanced ceramide synthase activity. Since the generation of ceramide
has been directly linked with the activation of the transcription
factor, NFB, this was investigated as a novel target for the action
of daunorubicin. Here we describe how treatment of HL-60 promyelocytes
and Jurkat T lymphoma cells with daunorubicin results in the activation
of the transcription factor NF
B. The effect of daunorubicin was evident following 1-2 h treatment, which was in contrast to the time
course of activation obtained with the cytokine, tumor necrosis factor,
where NF
B activation was detected within minutes of cellular stimulation. Activated complexes were shown to contain predominantly p50 and p65/RelA subunit components. Daunorubicin also induced I
B
degradation and increased the expression of an NF
B-linked reporter
gene. In addition, the drug was found to strongly potentiate the
ability of tumor necrosis factor to induce an NF
B-linked reporter
gene, suggesting a synergy between these two agents in this response.
These events were sensitive to the iron chelator, deferoxamine mesylate
(desferal), and the anti-oxidant and metal chelator pyrrolidine
dithiocarbamate. A structurally related compound, mitoxantrone, which,
unlike daunorubicin, is unable to undergo redox cycling in cells, also
activated NF
B in a pyrrolidine dithiocarbamate-sensitive manner. A
specific inhibitor of ceramide synthase, fumonisin B1, had no effect on
daunorubicin induced NF
B activation at a range of concentrations
previously reported to block apoptosis induced by this drug. However,
this agent could inhibit increases in ceramide induced by daunorubicin,
in addition to blocking ceramide synthase activity from HL-60 cells
which was activated in response to daunorubicin treatment. These data
therefore suggest that the effect of daunorubicin on NF
B is unlikely
to involve ceramide, but may involve reactive oxygen species generated
as a result of endogenous cellular processes rather than reductive
metabolism of the drug. As NF
B may be involved in apoptosis, this
effect may be an important aspect of the cellular responses to this
agent.
The anthracycline antibiotic, daunorubicin, is widely used in cancer chemotherapy with proven therapeutic benefit in the treatment of a variety of neoplasia (1). Although its mechanism of anti-tumor action is uncertain, DNA is believed to be a primary target (2). Its ability to cause strand scission may be mediated by stabilizing a cleavable complex between DNA and the enzyme, topoisomerase II, and/or oxygen radicals arising from redox cycling following its bioreduction. Additionally, bioreduction products and reactive oxygen species have been associated with anthracycline induced alkylation of cellular macromolecules, DNA intercalation and cross-linking, lipid peroxidation, and cell membrane damage (2). Irrespective of the initial insult, anthracyclines, along with a variety of agonists, ultimately activate the event of programmed cell death or apoptosis in cells (3). Their ability to induce this pathway may be a mechanism underlying their therapeutic efficacy in certain tumor types.
The development of the apoptotic morphology is well defined; however,
signaling pathways that may act as primary mediators of apoptosis and
growth suppression are poorly characterized (4). Some of those relevant
to the cytotoxic action of chemotherapeutic drugs include the
triggering of CD95/CD95-L interaction resulting in a type of autocrine
suicide (5), and the activation of effector molecules such as
interleukin-1 converting enzyme-like proteases (6). Recent studies
suggest that the neutral lipid ceramide may also play a role in
mediating drug-induced apoptosis (7, 8). Ceramide is a putative second
messenger which can also be generated following the activation of
distinct sphingomyelinase activities in response to a range of
extracellular agents including TNF-,1
interleukin-1
,
-interferon, nerve growth factor, Fas ligand, and
1,25-dihydroxyvitamin D3 (9). A general role in growth arrest and suppression is suggested by its ability to induce cell differentiation, cell-cycle arrest, apoptosis, or cell senescence (9),
although a mitogenic role has been demonstrated in certain cell types
(10). In a recent study, daunorubicin was shown to increase ceramide
levels in cells following induction of the enzyme ceramide synthase
(7). Furthermore, inhibition of this enzyme by the mycotoxin, fumonisin
B1, blocked apoptosis induced by daunorubicin. The regulated
biosynthesis of ceramide may represent a signaling mechanism by which
apoptotic events are induced by this drug.
The generation of ceramide following activation of lysosomal acid
sphingomyelinase has not only been linked with TNF-induced apoptosis,
but also with the activation of NFB (11). This inducible transcription factor has been implicated in the regulation of many
genes which code for mediators of the immune, acute phase, and
inflammatory responses (12). The DNA-binding protein complex recognizes
a discrete nucleotide sequence (5
-GGGACTTTCC-3
) in the upstream
regions of a variety of responsive genes. Subunits belonging to the
NF
B family comprise five members in mammals: p50, p65 (RelA), c-Rel,
p52, and RelB. These proteins share a conserved 300-amino acid sequence
in the N-terminal portion, termed the Rel homology domain, which
mediates DNA binding, protein dimerization, nuclear localization, and
binding of the inhibitor protein I
B (either
or
) (12).
Various dimer combinations of these proteins have distinct DNA binding
specificities and may serve to activate specific sets of genes (13). In
resting cells, the NF
B dimer is sequestered in the cytosol by
associating with I
B, and can be "liberated" from this complex by
a variety of inducers. A simple model for NF
B activation is as
follows. Phosphorylation of I
B by specific activated protein
kinase(s) tags it for proteolytic degradation (14). This facilitates
the nuclear translocation of activated NF
B complexes, whereupon
binding to cognate sequences, gene expression is activated.
The signaling pathways linking receptor stimulation to NFB
activation are poorly defined. A number of kinases have been implicated in the phosphorylation of I
B, the most notable being a recently identified ubiquitination-dependent multisubunit protein
kinase (14). Phosphorylation at two specific residues, serine 32 and 36, on I
B
, is thought to lead to its ubiquitination and
subsequent degradation by the proteosome, facilitating NF
B release
and translocation into the nucleus (15). Mutants lacking these residues
cannot undergo phosphorylation (and subsequent proteolytic degradation) in response to a variety of stimuli (16) suggesting that many signal
transduction pathways converge on the putative I
B kinase(s).
In addition to ceramide being implicated as an upstream regulator of
IB
phosphorylation, a model has been proposed whereby reactive
oxygen species (ROS) act as second messengers in this event (17).
Evidence to support this model is based on the ability of
H2O2 to activate NF
B (18) and the inhibitory
effects of antioxidants such as N-acetylcysteine (a thiol
antioxidant and glutathione precursor) and pyrrolidine
dithiocarbamate (PDTC), which is also a metal chelator, on NF
B
activation (18-21). The link between ceramide and ROS in signaling
either transcriptional activation or apoptotic events is unclear (22,
23). In a model for NF
B proposed by Baeuerle and Henkel (12),
agonist-stimulated ceramide generation lies upstream of an event which
triggers H2O2 production, leading to the
activation of this transcription factor.
Because previous work demonstrated the production of ROS and ceramide
in response to daunorubicin (2, 7), we investigated the effect of this
agent on NFB activation. We have found that daunorubicin signals
NF
B activation in HL-60 promyelocytes and Jurkat T cells by a
PDTC-sensitive mechanism, suggesting the involvement of ROS, and not
activated ceramide synthase. The importance of this signal for
daunorubicin-mediated apoptosis is discussed.
HL-60 and Jurkat T cells (both obtained from the
European Collection of Animal Cell Culture (ECACC, Salisbury, United
Kingdom) were grown in suspension culture in RPMI 1640 supplemented
with 10% fetal calf serum, penicillin/streptomycin (100 units/ml and 100 mg/ml, respectively), and L-glutamate (2 mM
final concentration), all obtained from Life Technologies, Inc.
(Paisley, United Kingdom). Recombinant human TNF was a gift from
Zeneca Pharmaceuticals Ltd., Macclesfield, United Kingdom. Mitoxantrone
was also a generous gift from Wyeth-Ayerst Research (United Kingdom).
Poly(dI·dC) was from Pharmacia Biosystems (Milton Keynes, United
Kingdom), T4 polynucleotide kinase and oligonucleotide containing the
consensus sequence (5
-GGGACTTTCC-3
), corresponding to the
-light
chain enhancer motif, were purchased from Promega (Southampton, United Kingdom). [
-32P]ATP (3000 Ci/mmol),
[14C]chloramphenicol (56 mCi/mmol),
[1-14C]palmitoyl-coenzyme A (55 mCi/mmol), and ECL
reagent were from Amersham (Aylesbury, United Kingdom). Diacylglycerol
kinase was from Calbiochem (United Kingdom). Rabbit polyclonal antibody
preparations to the DNA-binding subunits of NF
B (c-Rel and RelA) and
the inhibitor protein I
B
were from Santa Cruz Biotechnology Inc.
Mutant NF
B oligonucleotide was also from Santa Cruz. An antiserum to
the p50 subunit of NF
B was a generous gift from Dr. Jean Imbert
(INSERM, Marseille). All other reagents were purchased from Sigma
(Poole, Dorset, United Kingdom) unless otherwise stated.
For treatments, cells in late log phase of growth were resuspended in fresh medium at a concentration of 1 × 106/ml and incubated at 37 °C in a humidified atmosphere of 5% CO2, 95% air. Where required, cells were preincubated with inhibitors (fumonisin B1 and PDTC for 60 min, and desferal for 16 h) prior to the addition of drug (4 h). Following stimulation, incubations were discontinued by the addition of ice-cold phosphate-buffered saline, and either nuclear or whole cell extracts were prepared as described previously (24). Protein determinations were made using the Bradford assay with bovine albumin as standard.
Transfection StudiesThe transactivating potential of
activated NFB complexes was assessed following transfection of cells
(25) with a plasmid containing five NF
B consensus sequences upstream
of a chloramphenicol acetyltransferase reporter gene
(pCATTM-Promoter plasmid, a gift from Dr. Tim Bird, Immunex
Corp., Seattle, WA). Following treatment (indicated in legends),
extracts prepared from harvested cells were assayed for CAT activity as
described previously (26). Statistical significance was evaluated by
employing Student's t test for unpaired data.
Nuclear NFB was
assessed by the electrophoretic mobility shift assay using a 22-base
pair oligonucleotide containing the human
-light chain enhancer
motif, which had previously been end-labeled with
[
-32P]ATP as described (24). Typically, 4 µg of
nuclear extract protein was incubated with radiolabeled oligonucleotide
(10,000 cpm) at room temperature for 30 min using conditions as
described previously (24). NF
B complexes were resolved on 5%
acrylamide gels and identified following autoradiography. To identify
the subunit components of activated NF
B complexes, supershift
analysis was carried out where extracts from treated cells were
preincubated with antibody preparations to p50, RelA (p65), and c-Rel
subunit components on ice for 30 min prior to the addition of labeled probe. A similar protocol was employed in competition studies (incubations were at room temperature), where mutant and wild type
NF
B consensus sequence were assessed for their ability to block
binding of activated complexes to labeled wild type NF
B probe.
Equal amounts of whole cell lysate
protein (as indicated) were resolved by SDS-polyacrylamide gel
electrophoresis, transferred onto nitrocellulose, and IB
immunoblot analysis was performed as described previously (24).
Secondary antibody was used at a dilution of 1:400. The blots were
developed by ECL according to the manufacturers recommendations.
Ceramide was quantified by the diacylglycerol
kinase assay as described (7), with some modifications. In brief,
following stimulation, cell pellets were extracted with 600 µl of
chloroform, methanol, 1 N HCl (100:100:1, v/v/v). Following
alkaline hydrolysis (1 h at 37 °C), re-extracted samples were dried
down and redissolved in 50 µl of reaction buffer (7). The reaction
was started by the addition of 40 µg/ml (4 milliunits/ml)
Escherichia coli diacylglycerol kinase followed closely by
10 µCi of [-32P]ATP. Reaction termination was as
described following incubation at room temperature for 90 min. The
level of ceramide was determined by comparison with a standard curve
generated with known amounts of ceramide (ceramide type III;
Sigma).
This activity was measured in HL-60 microsomal membranes as described previously (7). In general, 50 × 106 cells were pelleted following drug treatment and disrupted in 300 µl of homogenization buffer by repeatedly passing through a 26-gauge needle. Microsomal membrane protein (37.5 µg) was incubated in a 250-µl reaction volume/mixture as described (7) with dihydrosphingosine as substrate. The reaction was started by the addition of 3.6 µM (0.2 µCi) [1-14C]palmitoyl-coenzyme A, the incubation was allowed to proceed for 1 h at 37 °C, and stopped by extraction with an equal volume of chloroform/methanol (2:1, v/v). The substrate concentrations chosen were based on those reported to allow maximal enzyme activity to be monitored (7), with 100 µM dihydrosphingosine being optimal. Following TLC as described (7), radioactivity corresponding to synthesized dihydroceramide was determined using an InstantImagerTM (Packard Instrument Co., Meriden, CT).
Treatment of both HL-60 and Jurkat T
cells with the anthracycline antibiotic, daunorubicin, resulted in the
activation of NFB, which was dose-dependent and
time-responsive (Fig. 1A, C, and
D). Fig. 1A illustrates data obtained with HL-60
cells, where activation of NF
B is demonstrated by the appearance of
DNA-protein complexes. The activation was time-dependent
occurring from 1 to 4 h and was sustained up to 24 h (Fig.
1A). This was in contrast to that seen with TNF, where the
activation of NF
B was rapid occurring within minutes of cellular
stimulation (Fig. 1B). Concentrations of daunorubicin
employed were similar to those previously reported to induce apoptosis
in this cell line (7), with activation being apparent at 0.05 µM, and peaking at 0.5 µM (Fig.
1C). These concentrations also paralleled that reported for
ceramide elevation induced by this drug (7). Data for TNF (0.6 and 2.5 ng/ml) are shown in Fig. 1C for comparison purposes. A less
potent induction was observed in Jurkat T cells (Fig. 1D)
where weak activation could be detected at 0.125 µM and a
strong signal observed at 2.5 µM daunorubicin.
The binding specificity of activated complexes was demonstrated by
competition studies in which unlabeled oligonucleotide containing
NFB consensus sequence inhibited the appearance of retarded
complexes, whereas a mutant oligonucleotide had no effect at equivalent
concentrations (Fig. 2A). Analysis of
specific subunit components in activated NF
B complexes revealed the
presence of p50 and to a lesser extent p65/RelA as indicated by
enhanced retardation of labeled complexes following gel electrophoresis
(Fig. 2B). Although supershifted complexes were not seen
with anti-c-Rel antibodies, a weaker signal when compared with control
lanes suggested the presence of this subunit component in the complex
(Fig. 2B, lane 2).
Daunorubicin Induces I
HL-60 cells treated
with daunorubicin at doses which resulted in NFB activation were
examined for degradation of the inhibitor protein, I
B
, a critical
event in the activation of this transcription factor (12). A marked
degradation of this inhibitor protein was observed which was
dose-responsive (Fig. 3). In support of the proposed
model for I
B
degradation, a doublet was observed prior to
degradation (open arrow), most probably corresponding to the
phosphorylated form of this protein, which is a signal for its
degradation (15). At the highest concentrations of daunorubicin employed (2.5 µM), I
B
degradation was complete
(Fig. 3, lane 6).
Daunorubicin Stimulates
Following
transfection of Jurkat T cells with a CAT reporter gene construct
containing five NFB sites upstream of a chloramphenicol acetyltransferase (CAT) reporter gene, the effects of daunorubicin on
B-dependent gene expression were investigated.
Daunorubicin induced expression of CAT activity in a dose dependent
fashion (Fig. 4A). At 0.25 µM
daunorubicin, a concentration previously shown to activate NF
B, CAT
activity was increased 4-fold over control values (unstimulated cells),
indicating that induced complexes were transcriptionally active. In
addition, daunorubicin and TNF were found to synergize in this response
(Fig. 4B). Combining a concentration of TNF which was
marginally inducing (1.2-fold over control) with a concentration of
daunorubicin inducing a 6-fold increase in CAT activity, resulted in a
14-fold induction. This synergy suggests that TNF and daunorubicin
activated NF
B by different pathways, as was previously suggested
from the different time courses of activation observed for these two
agonists (Fig. 1, A and B).
Activation of NF
We next investigated the mechanism by which daunorubicin
activates NFB. In cells pretreated with the metal chelators desferal and PDTC (which also has anti-oxidant properties), activation was
inhibited (Fig. 5A), as indicated by a
diminished signal corresponding to activated complexes. Desferal (1 mM) inhibited the response by 38% with no further
inhibition being observed at higher doses (lanes 3-5). PDTC
was the more potent inhibitor of the two at equivalent concentrations
employed, inhibiting the response by 63% at 1 mM
(lane 8). PDTC and desferal were also found to inhibit the
daunorubicin-mediated increase in
B-dependent CAT
expression (Fig. 5B).
PDTC and desferal inhibit
daunorubicin-mediated NFB activation and
B-linked gene
expression. A, HL-60 cells (1 × 106/ml)
were preincubated with either desferal for 16 h (lanes
3-5) or PDTC for 1 h (lanes 6-8) (concentrations
indicated), prior to the addition of daunorubicin (0.25 µM). Nuclear extracts were prepared and assessed for
NF
B as described under "Experimental Procedures." NF
B-DNA
complexes are shown. Densitometric analysis was performed on
autoradiograms corresponding to the experiment shown by UVP
transillumination scanning. Quantitation was by Gelworks 1D
AdvancedTM software where inhibition is represented as % of control value (daunorubicin only). B, Jurkat T cells
(1 × 106/ml) transfected with an NF
B-linked
reporter plasmid were incubated with either PDTC (1 mM) or
desferal (2.5 mM) for 30 min prior to the addition of daunorubicin (24 h). Cell lysates were prepared and assayed for
CAT activity as described under "Experimental Procedures." Results
are representative of three separate experiments (triplicate samples),
and inhibition is expressed as percentage (mean ± S.D.) of
control (daunorubicin only).
Mitoxantrone Activates NF
We next determined whether the mechanism of NFB
activation involved redox cycling of daunorubicin. For this purpose, we
used a closely related anthraquinone, mitoxantrone, which does not undergo redox cycling (27, 28). Mitoxantrone was found to be as
potent an activator of NF
B in HL-60 as daunorubicin (Fig. 6A), with an effect being evident from 0.05 µM. In addition, PDTC was found to inhibit this
activation, with 1 mM completely blocking the response
induced by 1.0 and 0.2 µM mitoxantrone (Fig.
6B). These results suggested that activation of NF
B by
daunorubicin involved the generation of ROS via endogenous cellular
processes rather than through redox cycling of the drug per
se.
The Ceramide Synthase Inhibitor Fumonisin B1 Does Not Block NF
Finally, we tested the effect of a
specific ceramide synthase inhibitor, fumonisin B1, for its ability to
block daunorubicin-induced NFB activation at a range of
concentrations previously reported to block apoptosis induced by this
drug (7). Fumonisin B1 failed to inhibit NF
B activation at all
concentrations tested (Fig. 7A) and,
furthermore, no inhibition of daunorubicin-mediated CAT activation was
observed (data not shown). However, treatment of HL-60 cells with
daunorubicin (0.5 and 10 µM) for 4 h increased ceramide levels, as shown in Fig. 7B, with 10 µM causing a 2.7-fold increase over controls. This effect
was inhibited by fumonisin B1 (300 µM), with ceramide
levels being reduced to that in unstimulated cells. Furthermore,
treatment of HL-60 cells with daunorubicin (10 µM) for
4 h was found to increase microsomal ceramide synthase activity
more than 2-fold over controls. 300 µM fumonisin B1
inhibited this activity, reducing it to below control levels, which was consistent with its ability to act as a competitive inhibitor toward
dihydrosphingosine (Fig. 7B). This confirmed a previous report where induction of ceramide synthase was found to mediate increases in ceramide levels in cells in response to daunorubicin (7).
However, our data indicated that this process was not involved in
NF
B activation here.
We therefore concluded that increases in ROS, generated by endogenous
cellular processes, but not ceramide synthase induction, were mediating
the effect of daunorubicin on NFB activation and
B-driven gene
expression.
In this study, we present the report that daunorubicin, an
anthracycline antibiotic, activates NFB. The significance of this effect was first examined in terms of its transcriptional potential. In
this regard, it is established that the dimer composition of the NF
B
complex determines its fine DNA-binding specificity (13), giving rise
to selective transcriptional activation or attenuation, the latter of
which has been observed with non-transactivating p50 homodimeric forms
(29). Transcriptional activation of specific sets of genes will
primarily depend on various dimer combinations being activated
distinctly, or whether their relative amounts in cell types and tissues
are subject to regulation. p50, RelA (p65), and c-Rel are the major
components of NF
B complexes (12), binding to most of the identified
cis-acting
B sites. Analysis of subunits present in activated
complexes from daunorubicin-treated cells indicated the presence of
both RelA and p50 components, which together constitute the predominant
transcriptionally active NF
B dimer combination. The ability of these
activated complexes to promote transactivation was confirmed in a
reporter gene assay, where treatment of cells with daunorubicin
stimulated activity from a transfected NF
B-linked reporter plasmid
in a dose-responsive manner. Furthermore, in line with the
classical model of NF
B activation (12), I
B phosphorylation
and degradation was demonstrated at concentrations which
paralleled those shown to activate NF
B and stimulate
transcription.
We considered the role of oxidative stress and ROS as second messengers
in daunorubicin-mediated NFB activation. In this regard, NF
B is
considered to be an oxidative stress-responsive transcription factor
(17) and interestingly, daunorubicin itself can be reductively
metabolized to a semi-quinone radical intermediate (2) which further
participates in reactions which give rise to ROS. The single-electron
reduction of daunorubicin is catalyzed by a number of cellular enzymes,
including cytochrome P450, the flavin NADPH-cytochrome p450 reductases,
NADH-cytochrome b5 reductase, and mitochondrial
NADH dehydrogenase (2). The resulting semiquinone is highly reactive
and in the presence of O2 rapidly autoxidizes to the parent
quinone with concomitant production of superoxide anion radical.
Furthermore, ROS are generated through its participation in
futile/redox cycling. Reactive oxygen species have previously been
implicated in the mechanism of NF
B activation in response to
cytokines, phorbol esters, and bacterial toxins (17, 18). Conclusions
have been drawn from studies using modulators of signaling events with
antioxidant and metal chelating properties (18, 19, 21), and the
overexpression of enzymes such as catalase and superoxide dismutase
(20). To investigate the mechanism of NF
B activation by
daunorubicin, compounds were employed which included PDTC which has
both antioxidant and metal chelating properties, and previously has
been shown to inhibit the activation of NF
B mediated by
H2O2 (19). Another inhibitor utilized,
deferoxamine (desferal), can interfere with the production of oxygen
radicals, in particular OH radicals generated in the presence of
catalytic amounts of transition metals (the Fenton reaction), by
preferentially chelating iron ions (30). Both compounds inhibited
daunorubicin-induced NF
B activation and
B-linked gene expression.
The more potent inhibition observed with PDTC at equivalent
concentrations may be due to additional antioxidant properties.
Interestingly, desferal has previously been shown to reduce the growth
inhibitory effects of daunorubicin to cells (2), possibly suggesting
that NF
B-mediated transcriptional regulation might in some way
participate in the growth inhibitory effects of this compound.
Further evidence for ROS involvement in the effect of daunorubicin came
from time course studies where NFB activation was significantly
slower when compared with that obtained with the cytokine, TNF, with
enhanced nuclear complexes only being detected 1-2 h
post-treatment. However, this time course for NF
B activation is
similar to that exhibited by H2O2 in
endothelial and epithelial cells (31, 32), suggesting that a similar
signaling pathway for NF
B activation might be mediated by agents
which directly generate ROS. Furthermore, we found that daunorubicin
and TNF could synergize in the induction of a
B-linked reporter
gene. The basis for this is unclear, but suggests that both agents
activate NF
B by different mechanisms, as was also indicated from
time course studies.
Interestingly, the photosensitizer, proflavine, can activate NFB in
a time course which mirrors that previously evinced by H2O2 (32) and it has been suggested that DNA
oxidative damage might initiate a signaling event (distinct from that
initiated by cytokines) which promotes translocation of NF
B
complexes resident in the cytoplasm into the nucleus. A similar effect
may be occurring with daunorubicin-mediated NF
B activation. A number
of agonists which have also been shown to activate NF
B directly, for
example, ionizing radiation, TNF, and oxidative stress, can also induce DNA strand breakage in treated cells (33). It is possible that this
damage may be a signal for the later activation of specific NF
B
complexes, as has been previously suggested (32).
Although enzyme-mediated daunorubicin free-radical formation may have
been the source of ROS which activated NFB, it was also possible
that endogenous cellular processes were responsible for ROS generation.
To test this, we examined a closely related compound, mitoxantrone,
which, although structurally similar to daunorubicin, does not undergo
redox cycling (27, 28). It was developed with the intention of
maintaining the DNA-complexing ability of doxorubicin, but reducing
systemic side effects such as cardiotoxicity, the cause of which is
purported to be ROS generation via redox cycling (1). Mitoxantrone was
found to be as potent at activating NF
B as daunorubicin, and was
even more susceptible to inhibition by PDTC. This suggested that the
mechanism of NF
B activation by daunorubicin (and mitoxantrone)
involved generation of ROS, not through redox cycling of the drug, but
through cellular events activated by the compounds leading to oxidative
stress. The precise source of the ROS awaits determination, as is the case for several other activators of NF
B which are sensitive to
antioxidant inhibition.
Other studies have questioned the role of redox cycling in
daunorubicin's cytotoxic effects, with mitoxantrone being more potent
than the daunorubicin analogue, doxorubicin, in this regard (27). It
therefore appears that redox cycling may not be critical for either
cytotoxicity or NFB activation induced by such anthraquinones. Mitoxantrone is a potent inducer of DNA strand breakage (27). While
this has been implicated in mediating its cytotoxic effects (27), it
may also be a signaling event leading to NF
B activation, analogous
to that proposed above for daunorubicin.
The time course of NFB activation by daunorubicin paralleled that
reported in a previous study for ceramide elevation induced by this
drug (7). Bose et al. (7), employed a mycotoxin, fumonisin
B1, which is a specific inhibitor of ceramide synthase, to demonstrate
that this enzymic activity was responsible for daunorubicin-induced
ceramide elevation. However, in our studies, fumonisin B1 did not block
daunorubicin-mediated NF
B activation. This result suggested that
ceramide reportedly generated during apoptotic induction by this drug
was not responsible for NF
B activation, although it is an
established inducer of this event (11). We were, however, able to
demonstrate that under conditions where fumonisin B1 failed to
inhibit NF
B activation, increases in ceramide induced by
daunorubicin in HL-60 were blocked. In addition, fumonisin B1 inhibited
ceramide synthase activity in microsomal extracts from
daunorubicin-treated HL-60 cells. This confirmed results from a
previous study which demonstrated that daunorubicin increases ceramide
in cells through an induction of ceramide synthase activity (7). In
another study, it has been shown that daunorubicin activates neutral
sphingomyelinase activity and that this is responsible for the ceramide
increase in response to clinically relevant doses of daunorubicin (8). Furthermore, in direct contradiction to the report by Bose et al. (7), they failed to demonstrate inhibition by fumonisin B1 of
apoptosis induced by daunorubicin. The basis for these inconsistencies is unclear. In our study, any increases in ceramide were abolished by
fumonisin B1, implying that ceramide synthase is the enzyme responsible
for such increases. As it was only with higher doses of daunorubicin
that we observed increases in ceramide and ceramide synthase activity,
it was possible, although unlikely, that the NF
B activation evident
at lower doses of daunorubicin was mediated via an undetectable rise in
ceramide occurring as a result of sphingomyelinase activation. We have
concluded from our findings that ceramide was unlikely to be important
in the effect of daunorubicin on NF
B. This conclusion is consistent
with observations indicating that ceramide is unlikely to be an
important signal for other activators of NF
B such as TNF (34).
Another recent study questions the importance of ceramide synthase in
the apoptotic effect of daunorubicin. Doses of this agent used to
induce the enzyme are suggested to be above therapeutic concentrations
(5) and in the same paper, the investigators show that a closely
related analogue, doxorubicin, induced apoptosis via FAS ligand (5)
which like daunorubicin has been shown to activate sphingomyelinase
(35). The effective concentration range of daunorubicin employed in our
study was in agreement with that reported to induce apoptosis (7) and
the suggested therapeutic plasma concentrations for the closely related
analogue, doxorubicin (5), underlining the potential clinical relevance
of our observations. The precise role of ceramide synthase in the
induction of apoptosis by daunorubicin therefore awaits clarification,
although as stated our study indicates that it is not involved in
NFB activation.
It is tempting to speculate that increased expression of genes
regulated by NFB in response to daunorubicin may be involved in
daunorubicin-mediated apoptosis. One candidate gene would be c-myc which plays a pivotal role in the induction of
apoptosis (36). Its expression has been shown to be regulated in
response to different hetero- and homodimeric NF
B complexes (29) and studies are consistent with the possibility that its overexpression might be related to apoptotic induction (37). It has also been proposed
that c-Rel, which is present in the NF
B complex, may function in the
activation of a set of death genes where its elevated expression was
shown to coincide with the onset of apoptosis (38).
Other chemotherapeutic agents have been shown to activate NFB. For
example, the deoxycytidine analogue, ara-C, has been reported to
activate NF
B, via neutral sphingomyelinase (39). Similar to
daunorubicin, it was also found to induce NF
B-linked gene expression
independently at a concentration which correlated with its ability to
activate this transcription factor. The DNA alkylating agents,
mitomycin C, has recently been shown to activate NF
B (40, 41) by a
novel mechanism involving enhanced nuclear processing of p105 in
Epstein-Barr Virus-immortalized B cells (40). NF
B activation may,
therefore, be a common mechanism for apoptosis-inducing anti-neoplastic
agents. It is also possible, however, that NF
B activation represents
an anti-apoptotic response. This has been convincingly demonstrated in
three recent reports. Studies in cells from transgenic mice deficient
in p65/RelA, or in cells where NF
B is inhibited demonstrated
enhanced apoptosis in response to a range of agents including
daunorubicin (42, 43, 44). In addition, the p65/RelA-deficient mice
exhibited massive liver degeneration by apoptosis (45). Significantly,
other mouse tissues did not show enhanced apoptosis. Drug resistance is
frequently associated with altered expression of certain
xenobiotic-metabolizing enzymes in the liver and NF
B may play a role
in regulating the expression of such proteins, as has been suggested
(46). A shift in the balance between apoptosis and NF
B could
therefore determine whether cells survive or die and so the study of a
functional link between NF
B-mediated transcriptional activation and
apoptotic induction or inhibition may provide important information on
anthracycline antitumor efficacy.
Anti-sera to the NFB p50 subunit component
was generously provided by Dr. Jean Imbert, INSERM, Marseille. M. P. B. thanks Andrew Bowie and E. R. Boland for helpful discussions.