ABSTRACT
The nuclear factor
[kappa]
B (NF-
[kappa]
B) is involved in T cell activation and enhances HIV-1 gene expression. It is activated in response to numerous stimuli,
including oxidative stress. Oxidative stress damages membrane lipids, proteins
and nucleic acids. We have shown previously that oxidative DNA damage generated
by photosensitization could trigger activation of NF-
[kappa]
B. We now show that a series of topoisomerase poisons (actinomycin D,
camptothecin, daunomycin and etoposide) also activate NF-
[kappa]
B (NFKB1/RelA dimer) in ACH-2 and CEM cells. This activation is inhibited by pyrrolidine
dithiocarbamate. In ACH-2 cells latently infected by HIV-1, camptothecin, daunomycin and etoposide are able to enhance virus
production. Since topoisomerase poisons cause the formation of single- and double-strand breaks in DNA, these lesions might be capable of triggering
NF-
[kappa]
B activation. Indeed, DNA damaging agents generating adducts (
trans
-platin and 4-nitroquinoline 1-oxide) and/or crosslinks in DNA (
cis
-platin and mitomycin C) do not or only weakly activate NF-
[kappa]
B in T cell lines.
The transcription factor NF-[kappa]B is involved in HIV-1 gene expression and T cell activation. Its main form is a
dimer of two subunits called NFKB1 (p50) and RelA (p65) (
1
,
2
). Activation of NF-[kappa]B is controlled by phosphorylation and proteolysis of an inhibitory
subunit called I[kappa]B. In resting T cells, I[kappa]B retains the NFKB1/RelA dimer in the cytoplasm. Upon stimulation,
I[kappa]B is phosphorylated by an as yet unknown kinase and then degraded by the
ubiquitin-dependent activity of the 26S proteasome (
3
,
4
). The free NF-[kappa]B dimer then migrates to the nucleus and binds to specific
sequences. Such sequences ([kappa]B elements) are found in promoter or enhancer regions of genes that are
crucial for immune and acute phase responses, i.e. those coding for adhesion
proteins, membrane receptors, cytokines, inducible nitric oxide synthase, etc.
(reviewed in
1
). The HIV-1 long terminal repeat (LTR) has two [kappa]B sequences in its enhancer region. The importance of those two
sites for HIV-1 gene expression has been extensively demonstrated (
5
-
7
).
NF-[kappa]B is activated by various agents (
1
), including cytokines (e.g. tumor necrosis factor, TNF), mitogens (e.g. lectins
or phorbol 12-myristate 13-acetate, PMA) or UV light. The redox status of the cell modulates
the activation of NF-[kappa]B and oxidative stress is itself an activator of NF-[kappa]B (
8
-
11
). Oxidative stress is defined as the exposure of cells to abnormally high
concentrations of reactive oxygen species (ROS), such as superoxide anion (O
2
--
), hydrogen peroxide (H
2
O
2
), hydroxyl radicals (OH
-
), singlet oxygen (
1
O
2
), nitric oxide (NO
-
) and hypochlorous acid (HClO). All these species can be produced
in vivo
(for a review see
12
). Oxidative stress causes lipid peroxidation, protein oxidation and
crosslinking through disulfide bridges and DNA damage (
13
,
14
). Oxidative stress is detectable in HIV-infected individuals and may affect progression of the HIV-associated disease (
15
and references therein).
In DNA, the main forms of damage caused by oxidative stress are single-strand breaks and oxidized bases (
14
). We have previously shown that generation of oxidative DNA damage localized in
DNA and consisting mainly of strand breaks and oxidation of guanines (generated
with a reaction photosensitized by an intercalating agent, proflavine) was able
to activate NF-[kappa]B and the replication of HIV-1 in non-productively infected lymphocytic (ACH-2) or promonocytic (U1) cells (
16
).
Several chemical DNA damaging drugs activate NF-[kappa]B and/or transcription directed by the HIV-1 LTR: mitomycin C (MMC), ethyl methanesulfonate (EMS),
methyl methanesulfonate (MMS), psoralen+UVA,
cis
-diamine dichloroplatinum (
cis
-Pt), 4-nitroquinoline 1-oxide (4-NQO), 5-fluorouracile and 1-[beta]-D- arabinofuranosyl cytosine
(araC) (
9
,
17
-
20
). UV light and ionizing radiation are inducers of NF-[kappa]B and of expression of HIV-1 genes (
21
,
22
) and DNA damage has often been pointed to as the cause of activation, but these
agents also damage other components of the cell by producing free radicals. UVC-induced NF-[kappa]B activation may not even depend on DNA damage (
23
-
25
).
In this work we have compared the effects of drugs generating different types of
DNA damage. 4-NQO, MMC,
cis
-Pt and
trans
-diamine dichloroplatinum (
trans
-Pt) react by addition on DNA (
26
,
27
), with MMC and
cis
-Pt also causing crosslinks (
27
,
28
). Actinomycin D (ActD), camptothecin (Cpt), daunomycin (Dauno) and etoposide
(Etop) are topoisomerase poisons. Topoisomerases are enzymes that control the
degree of supercoiling of DNA (
29
). Type I topoisomerase creates transient single-strand breaks in DNA, is crucial for transcription and is present in the
cell throughout the cell cycle, while type II topoisomerases create transient
double-strand breaks and are necessary for DNA replication and cell division (
29
-
31
). Topoisomerase poisons react with complexes formed between the enzyme and DNA
and prevent religation of the strand breaks (
30
,
32
). Cpt traps type I topoisomerase, while Dauno and Etop are topoisomerase II
poisons (
30
,
31
,
33
). ActD is mainly known as a transcription inhibitor, but has been shown to
inhibit both topoisomerases I and II (
34
). In ACH-2 and CEM T lymphocytic cell lines we detected activation of NF-[kappa]B by the topoisomerase poisons but not by drugs generating
adducts or crosslinks. We determined the ability of topoisomerase poisons to
trigger HIV-1 reactivation from latently infected ACH-2 cells. We also studied the effects of cycloheximide, antioxidants
and inhibitors of DNA polymerase and of poly(ADP-ribose) polymerase (PARP) on NF-[kappa]B activation by topoisomerase poisons.
All chemicals were from Sigma (Bornem, Belgium) unless otherwise specified. Cpt,
Dauno, Etop, PMA, araC and 4-NQO were dissolved in dimethylsulfoxide (DMSO) and stored at -20oC.
cis
-Pt and
trans
-Pt were dissolved in dimethylformamide and stored at 4oC. Aphidicolin and ActD were dissolved in methanol and ethanol
respectively and stored at -20oC. The concentration of ActD and Dauno stocks was verified by
absorbance, using [epsilon]
440 nm
= 24700 for ActD and [epsilon]
480 nm
= 11500 for Dauno in water, and stock solutions were protected from light. 3-Aminobenzamide (3-AB) and aclarubicin were freshly prepared in DMSO and ethanol respectively. Stock concentrations were chosen
to avoid more than 0.2% solvent in culture medium. Bleomycin, cycloheximide
(CHX), MMC,
N
-acetyl-L-cysteine (NAC; sodium form) and pyrrolidine dithiocarbamate
(PDTC) were freshly prepared in water.
CEM and ACH-2 cells (obtained through the NIH AIDS Research and Reference Reagent
Program, Bethesda, MD) were cultured in RPMI 1640 + Glutamax I + 10% fetal calf
serum (Life Technologies, Gaithersburg, MD) in a 5% CO
2
atmosphere. Routinely, cells were seeded at 5 * 10
5
cells/ml. On the day preceeding the experiment, the cells were seeded at 1-1.5 * 10
6
cells/ml. For drug treatment, the cells were centrifuged and resuspended in
fresh medium at a density of 10
6
cells/ml. Antioxidants or enzymes inhibitors were added if needed to the cells
and incubation was prolonged for 60 (with araC, aphidicolin and 3-AB) or 120 min (with NAC and PDTC). The DNA damaging drugs were then added
for the indicated times (see figure legends), then the cells were either washed
twice with 1 vol serum-free medium and resuspended in complete medium or the nuclear proteins
were extracted and stored at -80oC for use in gel retardation assays according to the published
procedure (
11
). The amounts of proteins in the samples were determined with the micro-BCA kit from Pierce (Rockford, IL).
The procedure was carried out as described previously (
11
). Briefly, the cells were washed in 1 ml cold phosphate-buffered saline (PBS) and centrifuged at 15 000
g
for 15 s, resuspended in 200 [mu]l cold hypotonic buffer [10 mM HEPES-KOH, 2 mM MgCl
2
, 0.1 mM EDTA, 10 mM KCl, 2 mM dithiothreitol (DTT), 0.5 mM
phenylmethylsulfonylfluoride (PMSF), 2 [mu]g/ml aprotinin, 1 [mu]g/ml leupeptin, 1 [mu]g/ml pepstatin (Boehringer, Mannheim, Germany), pH 7.9], left on ice
for 10 min, then vortexed and centrifuged at 15 000
g
for 30 s. The pellets of nuclei were gently resuspended in 15 [mu]l cold saline buffer (50 mM HEPES-KOH, 50 mM KCl, 300 mM NaCl, 0.1 mM EDTA, 10% v/v glycerol, 2 mM DTT,
0.5 mM PMSF, 2 [mu]g/ml aprotinin, 1 [mu]g/ml leupeptin, 1 [mu]g/ml pepstatin, pH 7.9) and left for 20 min on ice. After
centrifugation (15 000
g
for 5 min at 4oC), aliquots of supernatant containing the nuclear proteins were taken and
stored at -80oC. Gel retardation assays were performed with
32
P-labeled [kappa]B probes (Pharmacia Biotech Benelux, Roosendaal, The Netherlands), as described previously (
11
). The sequences of the probes used were:
wild-type NF-[kappa]B probes
5'-GTGTA
TCCCTGAAAGGCGACCCCTGAAAGGTCGTGT-5'
or
5'-GGTTACAA
TGTTCCCTGAAAGGCGACGGTT-5';
mutated NF-[kappa]B probe
5'-GGTTACAACTCACTTTCCGCTG
TGTTGAGTGAAAGGCGACGGTT-5'.
The [kappa]B sites are underlined. Identical results were obtained whether the probe
contained one or two [kappa]B sites. The antibodies against NFKB1 (p50), NFKB2 (p52), RelA (p65) and
c-Rel used for the supershift experiments have been described elsewhere (
35
) and were kindly provided by Dr V. Bours.
Culture medium of control or drug-treated cells was taken 24 h after treatment and cleared by centrifugation
at 15 000
g
for 5 min. The reverse transcriptase (RT) assay was performed on 1 ml
supernatant as described previously (
11
). Virus production was calculated as the amount of radioactivity divided by the
number of living cells in 1 ml culture medium (determined by Trypan blue
exclusion).
We treated ACH-2 cells, which are T leukemic cells latently infected by HIV-1, or CEM cells, which are their uninfected counterpart, with the
topoisomerase poisons ActD, Cpt, Dauno and Etop for different times at
concentrations in the micromolar range. All these drugs caused the appearance
of a [kappa]B binding complex (Fig.
1
A-D). This complex could be detected at concentrations as low as 0.2 [mu]M for Dauno and Cpt and 2-3 [mu]M for ActD and Etop. The activation reached a maximum
after 2 or 3 h treatment and was sustained for several hours with ActD and
Dauno, while it decreased after a peak with Cpt and Etop. In some experiments,
several bands instead of one appeared (see for instance Fig.
1
C). There was no relationship between the number of induced bands and the drug
used or the number of [kappa]B sites in the probe. When only one band was detected, it had the
mobility of the fastest migrating induced complex. The appearance of a [kappa]B binding complex was also observed with 10 [mu]M amsacrine and 50 [mu]M genistein, two other topoisomerase II poisons known to generate
strand breaks in DNA (
33
,
36
).
We then investigated whether topoisomerase poisons were able to trigger
reactivation of HIV-1 in the latently infected ACH-2 cell line, since expression of HIV-1 genes is mainly under the control of NF-[kappa]B. Since continuous treatment with the drugs is
toxic, the cells were treated for a limited time (2 h) and then washed twice
before being reseeded in culture medium for 24 h to allow cell recovery and
virus production. These treatments caused limited toxicity and allowed the
cells to proliferate again after 1 or 2 days (except for treatments with ActD,
which killed all the cells in <2 days even when administered at 2 [mu]M for 30 min only). Virus production in a sample was calculated as the RT
activity measured in the culture medium divided by the number of surviving
cells in that sample (so as to take drug toxicity into account) and virus
production in drug-treated samples was compared with virus production in control (untreated)
samples. Figure
4
shows that Cpt, Dauno and Etop caused an increase in virus production in drug-treated ACH-2 cells compared with control cells. Levels of reactivation observed
with Cpt and Etop were comparable with those reached with H
2
O
2
(
40
). The highest virus production was observed with Dauno, probably because
activation of NF-[kappa]B by this drug was sustained, while it was transient for Cpt and
Etop (see Fig.
1
).
To study the mechanism by which topoisomerase poisons induce NF-[kappa]B, we first used CHX, a protein synthesis inhibitor. As shown in
Figure
5
, CHX had no inhibitory effect on activation by Cpt, Dauno and Etop after 3 h
and even slightly enhanced NF-[kappa]B activation by these agents. Activation by ActD is presumed to
require little or no protein synthesis, since this drug is a well-known transcription inhibitor. Thus, NF-[kappa]B activation by topoisomerase poisons appears independent of
protein synthesis.
We have shown that topoisomerase poisons, which have in common the property of
generating DNA strand breaks, are able to cause NF-[kappa]B (NFKB1/RelA dimer) activation in the T lymphocytic cell lines ACH-2 and CEM. We propose that DNA strand breaks, which are
produced during oxidative stress, can lead to activation of NF-[kappa]B. DNA strand breaks are produced during exposure to H
2
O
2
or ionizing radiation (
13
,
14
), two agents that activate NF-[kappa]B (
8
,
10
,
22
). We had previously shown that oxidative stress localized in the nucleus and
consisting mainly of DNA strand breaks and guanine oxidation activated NF-[kappa]B in ACH-2 cells (
16
). We also detected NF-[kappa]B activation following treatment of cells with bleomycin, a DNA
strand break-generating drug, but not with aclarubicin, a topoisomerase inhibitor which
does not stabilize cleavable complexes.
Topoisomerase poisons activate NF-[kappa]B in the T lymphocytic cell lines ACH-2 and CEM and in the Jurkat subclone JR (data not shown),
while H
2
O
2
activates NF-[kappa]B at low or moderate concentrations (100-250 [mu]M) in these three cell lines (
9
,
40
). In Jurkat (lymphocytic) or U1 (promonocytic) cells, where millimolar
concentrations of H
2
O
2
are required to induce NF-[kappa]B, ActD and Dauno caused activation of NF-[kappa]B, but Cpt and Etop did so only weakly (at the
concentrations tested). This parallels the results of Anderson
et al.
(
9
), who showed that NF-[kappa]B could be activated by oxidative stress in JR but not in Jurkat
cells. A recent report (
46
) showed that Dauno triggers ceramide generation in monocytic cells. DNA damage
might thus not be the only process by which Dauno triggers NF-[kappa]B and this could explain why Dauno activates NF-[kappa]B efficiently in cell lines where Cpt and Etop do not.
MMC,
cis
-Pt, 4-NQO and araC, which had been previously shown to induce NF-[kappa]B or HIV LTR-mediated transcription in some cell types (
17
-
19
,
21
), did not activate NF-[kappa]B in our T cell lines at the concentrations tested (see Figs
1
and
7
). However, MMC,
cis
-Pt and 4-NQO activated NF-[kappa]B in the monocytic cell line U1 (data not shown). Thus,
response to DNA damage seems to differ between cell lines and between cell
types.
We show that latently HIV-infected cells produce more virions when exposed to topoisomerase poisons
for a short time. Cpt was less effective than Etop and Dauno, but some authors
have reported that this drug inhibits Tat-mediated transactivation of transcription from the HIV-1 LTR (
47
). We presume that this reactivation of the virus is mediated by NF-[kappa]B, although we did not investigate the possible involvement of
other transcription factors. For example, p53 is induced by topoisomerase
poisons (
48
,
49
) and has an up-regulating effect on HIV LTR-driven transcription (
50
).
Topoisomerase poisons cause strand breaks in DNA by trapping the complex formed
between cut DNA and the topoisomerase enzyme. Cpt stabilizes topoisomerase I-DNA complexes once DNA is nicked, with one end (3') linked to the enzyme through a phosphotyrosine and the other end
(5'-OH) free (
30
). This type of damage is repaired rapidly and efficiently when Cpt is removed (
51
,
52
). When topoisomerase II-DNA complexes are trapped by Etop or Dauno, there is a double-strand break where the 5'-ends are linked to the enzyme through phosphotyrosines
and the 3'(-OH) ends are free. For both types of topoisomerases, the DNA ends
are kept together by the enzyme until the complex is encountered by a DNA
replication fork or transcription machinery. This latter event causes a double-strand break with release of the DNA ends (
30
-
32
,
53
), which often leads to cell death because it is very difficult to repair (
54
). Activation of NF-[kappa]B by bleomycin and lack of activation by aclarubicin shows that DNA
strand breaks are necessary and sufficient to trigger the activation process.
In addition to double-strand breaks, the topoisomerase II poisons used here (amsacrine, ActD,
Dauno and Etop) also cause single-strand breaks, probably because they generate free radicals (
33
,
36
,
37
). Diminution by DNA polymerase inhibitors of NF-[kappa]B activation by topoisomerase poisons and lack of sensitivity to the PARP
inhibitor 3-AB suggest, however, that the double-strand breaks mainly activate NF-[kappa]B. Moreover, Cpt and Etop at 10 [mu]M, which lead to similar levels of NF-[kappa]B activation, have been reported to
generate similar amounts of double-strand breaks, while Cpt is much more effective than Etop at creating
single-strand breaks (
36
).
Dauno and Etop are reduced to semiquinone radicals by various cellular enzymes (
33
,
37
,
55
). Dauno can subsequently generate H
2
O
2
, OH
-
and O
2
--
and oxidize membranes. Reduced Etop can oxidize intracellular thiols. However,
we do not think that topoisomerase poisons induce NF-[kappa]B by causing oxidative stress (elsewhere than in the nucleus)
because: (i) Cpt has never been reported (to our knowledge) to generate free
radicals; (ii) NAC did not abolish activation by topoisomerase poisons, while
it did with H
2
O
2
; (iii) MMC, also a prooxidant drug (
9
,
37
), did not activate NF-[kappa]B at moderate concentrations.
NF-[kappa]B activation by Cpt, Dauno and Etop, as with activation by most
other agents studied, is not inhibited by cycloheximide. The hypothesis that NF-[kappa]B activation by topoisomerase poisons does not involve protein
synthesis is reinforced by the fact that these drugs, especially ActD, inhibit
transcription (
30
,
33
,
52
). The possibility remains that transcription inhibition is the actual trigger
for NF-[kappa]B activation. Disappearance of I[kappa]B following inhibition of protein synthesis is unlikely to
explain activation by topoisomerase poisons, since cycloheximide did not induce
NF-[kappa]B strongly under our conditions (see Fig.
4
).
We tested the ability of antioxidants to inhibit NF-[kappa]B activation following treatment with topoisomerase poisons. Under
our conditions, NAC inhibited H
2
O
2
-mediated NF-[kappa]B activation and protected against H
2
O
2
toxicity, but was not effective against topoisomerase poisons or PMA. In
contrast, PDTC, which is effective against all agents reported to activate NF-[kappa]B (
56
), is effective against topoisomerase poisons and against H
2
O
2
, even when added at a much lower concentration (30 [mu]M PDTC versus 250 [mu]M H
2
O
2
). Such discrepancies between the effects of NAC and PDTC on NF-[kappa]B have already been reported (
41
) and a possible explanation is that, in our system, ROS are not involved in NF-[kappa]B activation by PMA or topoisomerase poisons. Thus, PDTC would not
abolish NF-[kappa]B activation by scavenging ROS but, for example, by specifically
inhibiting an enzyme of the activation cascade, possibly because of its metal
chelating properties. We obtained identical results with bleomycin (which
causes single- and double-strand breaks) regarding the effects of CHX, NAC and PDTC (data not
shown). We thus propose that DNA strand breaks trigger a signaling cascade
which leads to activation of a PDTC-inhibitable enzyme (still undetermined), from which the cascade converges
with other signaling pathways leading to NF-[kappa]B activation. One cannot exclude the possibility that PDTC partly
prevents DNA damage by topoisomerase poisons, since Dauno and Etop (as well as
amsacrine and bleomycin, but not Cpt to our knowledge) are reported to generate
free radicals as a part of or in addition to their properties as DNA strand-breaking agents (
33
,
37
). For example, PDTC prevents the formation of DNA double-strand breaks associated with the apoptotic process triggered by Etop (
57
).
The best known response pathway to DNA damage is the UV response, which
activates the transcription factors AP-1, NF-[kappa]B and ATF-2 (
25
,
58
and references therein). DNA strand breaks also trigger activation of the tumor
suppressor p53 (
48
,
49
). Topoisomerase poisons (as well as other DNA damaging agents) and oxidative
stress induce apoptosis (
51
,
59
,
60
), probably because they cause DNA damage impossible to repair without mutation.
The response to oxidative stress might be linked to the control of cell
cycle/cell division that can trigger apoptosis and is directed by p53. Further
studies are required to determine whether this link exists, as well as the role
in NF-[kappa]B activation of oxidative stress, of proteins such as the DNA
damage-dependent protein kinase and other kinases activated by DNA lesions, e.g.
the product of the protooncogene c-Abl (
61
).
In conclusion, our results contribute to the understanding of both the
mechanisms of NF-[kappa]B activation following DNA damage and the consequences of the
administration of some anticancer drugs, considering the crucial role of NF-[kappa]B in immune functions.
We thank Dr V. Bours for the antibodies used in supershift experiments and
helpful discussions and Prof. A. Albert for help with data analysis. This work
was supported by grants from the Belgian National Fund for Scientific Research
(NFSR, Brussels, Belgium), the AIDS Research Program of the Belgian NFSR and
the SIDACTION Program (Paris, France). B.P. is supported by the Belgian Fonds
pour la Formation à la Recherche dans l'Industrie et l'Agriculture (FRIA, Brussels,
Belgium). J.P. is Research Director at the Belgian NFSR.
REFERENCES
Return