We have shown previously that the camptothecin analogue topotecan (TPT), a topoisomerase I (Top 1) poison, inhibits hypoxia-inducible factor 1 (HIF-1) transcriptional activity and HIF-1α protein accumulation in hypoxia-treated U251 human glioma cells. In this article, we demonstrate that TPT does not affect HIF-1α protein half-life or mRNA accumulation but inhibits its translation. In addition, we demonstrate that Top 1 is required for the inhibition of HIF-1α protein accumulation by TPT as shown by experiments performed using camptothecin-resistant cell lines with known Top 1 alterations. Experiments performed with aphidicolin indicated that TPT inhibited HIF-1 protein accumulation in the absence of DNA replication. DNA-damaging agents, such as ionizing radiation and doxorubicin, did not affect HIF-1α protein accumulation. Ongoing transcription was essential for the inhibition of HIF-1α protein accumulation by TPT. Our results demonstrate the existence of a novel pathway connecting Top 1-dependent signaling events and the regulation of HIF-1α protein expression and function. In addition, our findings dissociate the cytotoxic activity of TPT from the inhibition of the HIF-1 pathway and raise the possibility of novel clinical applications of TPT aimed at targeting HIF-1-dependent responses.
Hypoxia inducible factor 1 (HIF-1) is a basic helix-loop-helix PAS (PER-ARNT-SIM) transcription factor composed of two subunits, HIF-1α and HIF-1β. The expression and activity of the HIF-α subunit are regulated uniquely by the intracellular oxygen concentration (1 , 2) . Under oxic conditions, HIF-1α protein is degraded rapidly and continuously by ubiquitination and proteasomal degradation on hydroxylation of Pro-402 and Pro-564, via an enzymatic process that requires O2 and iron, and binding with the von Hippel-Lindau protein (3, 4, 5, 6, 7) . The interaction between HIF-1α and von Hippel-Lindau protein is accelerated additionally by acetylation of Lys 532 through an N-acetyltransferase (8) . Under hypoxic conditions, HIF-1α protein accumulates and translocates to the nucleus where it forms an active complex with HIF-1β, which activates transcription of >60 target genes important for the adaptation and survival under hypoxia (9) . A variety of growth factors, such as epidermal growth factor (10) , heregulin (11) , insulin-like growth factor 1 and 2, and insulin (12, 13, 14) , induce the expression of HIF-1α protein in nonhypoxic conditions (2) . Binding of these growth factors to their receptors activates the phosphatidylinositol 3′-kinase (PI3k)-Akt-mammalian target of rapamycin (mTOR) pathway, leading to an increase in HIF-1α protein translation and stability (10 , 15) .
Overexpression of HIF-1α protein has been demonstrated in many human cancers and their metastases, and it is associated with increased vascularity and tumor progression (16) . Furthermore, HIF-1α overexpression is associated with treatment failure and patient mortality in oropharyngeal squamous cell cancer (17) , early-stage cervical cancer (18) , p53-mutant ovarian carcinoma (19) , oligodendroglioma (20) , and BCL2-positive esophageal cancer (21) . Recent evidence indicates that HIF-1α overexpression may trigger cancer cell invasion and migration. In particular, it has been shown that HIF-1α is involved in the up-regulation of cathepsin D, urokinase plasminogen activator receptor, matrix metalloproteinase 2, autocrine motility factor, and Met (22 , 23) . Therefore, the development of cancer therapeutics targeting HIF-1 activity is appealing (9) .
We have reported previously the results of a cell-based high-throughput screen for the identification of small molecule inhibitors of the HIF-1 pathway using the National Cancer Institute (NCI) Diversity Set, a collection of ∼2,000 compounds representing the greater chemical diversity of the NCI repository (24) . We found four compounds that specifically inhibited HIF-1-dependent induction of luciferase and vascular endothelial growth factor (VEGF) mRNA expression in U251 human glioma cells. Three of the compounds are closely related camptothecin analogues and topoisomerase I (Top 1) inhibitors. In particular, topotecan (TPT; National Service Center 609699) inhibited HIF-1-dependent luciferase expression and VEGF production and blocked HIF-1α protein accumulation. A larger screen of a library of ∼140,000 compounds available to NCI confirmed the camptothecins as a class of compounds that strongly inhibits HIF-1 transcriptional activity in vitro.
In this article, we have explored the mechanism by which TPT inhibits HIF-1α protein accumulation. At concentrations that did not cause cytotoxicity, TPT inhibited HIF-1α protein accumulation by an oxygen- and proteasome-independent pathway. Top 1 was required for the down-regulation of HIF-1α by TPT, which was independent of replication-mediated DNA damage. These results differentiate the mechanism of action of HIF-1 inhibition from the classic cytotoxic effects of camptothecins and raise the interesting possibility of a clinical schedule of administration of TPT aimed at targeting HIF-1-dependent responses.
MATERIALS AND METHODS.
Cell Lines and Reagents.
U251 human glioma cells were maintained as described previously (24) .
Human leukemia cell lines either sensitive (CEM) or resistant (CEM-C2) to camptothecin (because of a mutation on Top 1) were described previously (25) . These cell lines were maintained routinely in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum, penicillin (50 IU/ml), streptomycin (50 μg/ml), and 2 mm glutamine.
Cells were maintained at 37°C in a humidified incubator containing 21% O2 and 5% CO2 in air (referred to as normoxic conditions).
Hypoxia treatment was performed as described previously (24) . For the ionizing radiation treatment, U251 cells were seeded at a concentration of 1.5 × 106 cells per 100-mm dish and allowed to attach. Cells were exposed to 0, 5, or 10 Gy of radiation (dose rate = 7 Gy/min). Rapamycin, LY294002, cycloheximide (CHX), MG132, ALLNL, and desferoxamine (DFX) were purchased from Sigma (St. Louis, MO). Actinomycin D was purchased from Calbiochem (La Jolla, CA). Additionally, 5,6-dichloro-1-β-d-ribofuranosylbenzimidazole (DRB), 9-amino-CPT-20R, MJ-III-65, TPT, PS341, and doxorubicin were obtained from the Drug Synthesis and Chemistry Branch, Developmental Therapeutics Program at NCI. Results presented are representative of at least three independent experiments performed. Statistical analysis was performed using ANOVA (two-factor with replication) test (P < 0.01).
Plasmids, Transient Transfection, and Engineered Cell Lines.
The pGL2-TK-HRE plasmid, containing three copies of the hypoxia-responsive element from the inducible nitric oxide synthase promoter upstream of the firefly luciferase reporter gene, was described previously (26) .
DNA plasmids were prepared using a commercially available kit (Endofree Maxi-Prep; Qiagen, Inc., Valencia, CA). Transfections were performed using Effectene Transfection Reagents (Qiagen, Inc.) according to the manufacturer’s instructions. Efficiency of transfection was assessed by cotransfection with a plasmid harboring the Renilla luciferase gene under control of a constitutive promoter (Promega, Madison, WI).
Experiments using stably transfected cells (U251-HRE) were performed as described previously (24) .
Luciferase reporter assays were performed in 96-well optiplates (PerkinElmer, Boston, MA) using Steady Glo or Bright Glo luciferase assay reagents and dual luciferase reporter assay reagents (Promega).
Cells were washed twice with ice-cold Dulbecco’s PBS 1× and lysed with radioimmunoprecipitation assay buffer [150 mm Tris-HCl (pH 7.5), 150 mm NaCl, 1% NP40, 0.25% Na-deoxycholate, 1 mm EDTA, 2 mm DTT, 1 mm Pefabloc, 1 mm NaVanadate, 1 mm NaF, 4 μg/ml pepstatin, 4 μg/ml leupeptin, and 4 μg/ml aprotinin]. Using mechanical scraping, cell lysates then were collected. The cell suspension was rotated at 4°C for 15 min, and cellular debris was pelleted by centrifugation at 15,000 × g for 15 min at 4°C.
Whole cell lysate was separated on a 4–20% Tris-Glycine gel (Invitrogen, Carlsbad, CA) electroblotted on a polyvinylidene difluoride membrane (Invitrogen) and subjected to immunoblot analysis. Monoclonal anti-HIF-1α antibody was purchased from BD-Transduction Laboratories (Lexington, KY) and used at a 1:300 dilution. Monoclonal anti-HIF-1β antibody was purchased from Novus Biologicals (Littleton, CO) and used at a 1:1,500 dilution. Antibodies specific for phosphorylated (Ser-473) or total Akt, phosphorylated (Thr-389) or total p70S6K, and phosphorylated (Ser-65) or total 4E-BP1 were used at a 1:1,000 dilution (Cell Signaling Technology Inc., Beverly, MA). Horseradish peroxidase-conjugated mouse and rabbit IgG (1:10,000 dilution) and enhanced chemiluminescence reagents were from Amersham Pharmacia Biotech (Piscataway, NJ).
Total RNA was obtained using RNA Mini Kit (Qiagen, Inc.). Reverse transcription-PCR was performed using a reverse transcription-PCR kit (PE Biosystems, Foster City, CA) as described previously (24) . To measure human VEGF and HIF-1α expression, real-time PCR was performed using an ABI-Prism 7700 Sequence Detector (Applied Biosystems, Foster City, CA) as described previously (24) . Primers and specific probes were obtained from Applied Biosystems. The following primers and probes were used: human HIF-1α forward 5′-CCAGTTACGTTCCTTCGATCAGT-3′ and reverse 5′-TTTGAGGACTTGCGCTTTCA-3′. Human VEGF primers and probes were described previously (24) . Used as an internal control, 18S rRNA was assessed using premixed reagents from Applied Biosystems. Detection of VEGF and 18S rRNA was performed using TaqMan Universal PCR Master Mix (Applied Biosystems), and HIF-1α detection was performed using Sybr Green PCR Master Mix (Applied Biosystems).
HIF-1α Protein Translation Assay.
U251 cells were exposed to methionine- and cysteine-free RPMI 1640 for 2 h in the presence or absence of TPT (500 nm). Cells then were labeled by incubation with methionine- and cysteine-free medium containing 35S-methionine and 35S-cysteine (Trans Label; ICN Biomedicals, Irvine, CA) at a final concentration of 150 μCi/ml at 37°C for the indicated time (Fig. 3C) ⇓ . Total cell lysates were subjected to immunoprecipitation using anti-HIF-1α antibody (Novus Biologicals) in constant rotation at 4°C overnight. Protein G-agarose beads (Pierce, Rockford, IL) then were added for 1 h. Cell lysates subsequently were centrifuged at 4°C for 5 min, and G-agarose beads were washed with radioimmunoprecipitation assay lysis buffer three times before the addition of 2× sample buffer. Immunoprecipitation products then were separated on a 4–20% Tris-Glycine gel. The gel was dried using the HydroTech gel drying system (Bio-Rad Laboratories, Hercules, CA) before autoradiographed using Kodak XAR-5 film (Rochester, NY) and intensifying screens at −80°C.
TPT Specifically Inhibits the Accumulation of HIF-1α Protein in a Time- and Dose-Dependent Fashion.
We first performed time course experiments to address the kinetics of TPT-mediated inhibition of HIF-1α protein accumulation. TPT decreased the aerobic basal levels of HIF-1α at all time points (data not shown) and decreased the hypoxic-dependent accumulation of HIF-1α protein by ∼50% at 2 h, with complete abrogation by 6–8 h (Fig. 1A) ⇓ . In contrast, HIF-1β levels were not changed by either hypoxic incubation or addition of TPT. Similarly, TPT completely abrogated induction of HIF-1α accumulation in U251 cells by the iron chelator DFX (data not shown), which is a well-characterized inducer of HIF-1 (27) .
We then examined the effect of different doses of TPT on the hypoxic induction of HIF-1α protein accumulation. TPT decreased hypoxic induction of HIF-1α protein in a dose-dependent fashion with an EC50 of 54.4 nm (Fig. 1B) ⇓ . In contrast, HIF-1β expression was unchanged even in the presence of high concentrations of TPT. A similar dose-dependent inhibition of HIF-1α accumulation was observed in cells treated with DFX with an EC50 of 87.3 nm (data not shown).
These results demonstrate that TPT exerts a dose-dependent inhibition of hypoxic- and nonhypoxic-induced accumulation of HIF-1α protein, which is manifested as early as 2 h at pharmacologic concentrations.
Inhibition of HIF-1α Protein Accumulation by TPT Is Independent of Proteasomal Degradation.
HIF-1α protein is degraded mainly through the proteasome pathway. Blockade of proteasome function by proteasome inhibitors increases the accumulation of HIF-1α protein in normoxic conditions. To address whether TPT inhibited HIF-1α protein accumulation by promoting its degradation, we performed several experiments using different proteasome inhibitors (such as MG132, ALLNL, and PS341) added before TPT. As expected, addition of proteasome inhibitors caused accumulation of HIF-1α protein under nonhypoxic conditions (Fig. 2A) ⇓ . TPT inhibited the accumulation of HIF-1α protein despite the proteasome inhibition, indicating that TPT decreases HIF-1α protein accumulation through a pathway independent of proteasome degradation.
These results were compatible with the possibility that TPT affected HIF-1α protein half-life and promoted degradation through a different pathway. To address this issue, we performed experiments using CHX to block protein synthesis. HIF-1α protein levels were measured before (time 0) and at various times after CHX was added. As shown in Fig. 2B ⇓ , HIF-1α protein half-life in the presence of DFX was ∼93 min. In the presence of TPT, the HIF-1α protein half-life was ∼114 min, which demonstrates that TPT does not affect HIF-1α protein half-life and thus HIF-1α degradation.
Taken together, these experiments demonstrate that TPT does not influence HIF-1α protein degradation, raising the possibility that it may affect its production.
TPT Does Not Affect HIF-1α mRNA Accumulation But Decreases HIF-1α Protein Translation.
To determine whether Top 1 inhibition by camptothecins might affect the modulation of HIF-1α mRNA synthesis, we analyzed whether TPT affected HIF-1α mRNA expression under our experimental conditions. As shown in Fig. 3A ⇓ , addition of TPT did not affect HIF-1α mRNA accumulation under normoxia, hypoxia, or in the presence of DFX. Similar results were obtained at earlier time points or in Northern blot experiments (data not shown). In contrast, TPT decreased DFX- and hypoxic-dependent induction of VEGF mRNA expression by ∼70%, which was 12- and 14-fold higher than under normoxic conditions, respectively (Fig. 3B) ⇓ .
Induction of HIF-1α protein accumulation is controlled, at least in part, at the translational level (2) . Therefore, we performed metabolic labeling to measure the rate of HIF-1α protein translation in the presence or absence of TPT. Because TPT equally inhibits HIF-1α protein accumulation under normoxic and hypoxic conditions and because the role of hypoxia in HIF-1α protein translation is not defined clearly, we performed metabolic labeling experiments under normoxic conditions. As shown in Fig. 3C ⇓ , 35S-labeled HIF-1α progressively accumulated in the absence of TPT, whereas the addition of TPT greatly decreased HIF-1α accumulation. Under the same experimental conditions, TPT did not affect significantly total 35S incorporation, suggesting that it does not inhibit overall protein synthesis (data not shown). In contrast, CHX (40 μg/ml), a known protein synthesis inhibitor, decreased 35S incorporation by 90% compared with untreated cells (data not shown).
Overall, these experiments indicate that TPT does not affect HIF-1α mRNA accumulation but does decrease HIF-1α translation.
Inhibition of HIF-1α Accumulation by TPT Is Independent of the PI3k-Akt-mTOR Pathway.
The PI3k-Akt-mTOR pathway has been implicated in the regulation of HIF-1α protein primarily at the translational level (10 , 15) , although recent evidence also suggests a possible role in the degradation of HIF-1α (28) . Interestingly, it also has been shown that TPT inhibits Akt activity in A549 cells (29) . To address whether TPT inhibited HIF-1α protein accumulation by down-regulating the PI3k-Akt-mTOR pathway, we performed experiments in U251 using inhibitors of the PI3k (LY294002) or mTOR (rapamycin) pathways under normoxic or hypoxic conditions. U251 cells express constitutive phosphorylation of Akt, 4EBP1, and p70S6-kinase under aerobic conditions. As shown in Fig. 4 ⇓ , addition of LY294002 inhibited hypoxia-induced HIF-1α protein accumulation and the phosphorylation of Akt, 4EBP-1, and p70S6-kinase under the same conditions. Addition of rapamycin inhibited hypoxia-induced HIF-1α protein accumulation along with p70S6-kinase and 4EBP-1 phosphorylation and, as expected, did not inhibit Akt phosphorylation. In contrast, the addition of TPT inhibited HIF-1α accumulation but did not affect Akt, p70S6-kinase and 4EBP-1 phosphorylation under normoxic or hypoxic conditions.
These results indicate that the PI3k-Akt-mTOR pathway is involved in the hypoxic induction of HIF-1α protein in U251. They also demonstrate that the inhibition of HIF-1α protein translation by TPT is independent of the PI3k-Akt-mTOR pathway.
Top 1 Is Required for the Inhibition of HIF-1α Protein Accumulation by TPT.
The only known target of TPT is Top 1. To address whether Top 1 is required for the inhibition of HIF-1α by TPT, we tested HIF-1α transcriptional activity and protein accumulation in human leukemia cell lines either sensitive (CEM) or resistant (CEM-C2) to camptothecins (25) . We transiently transfected CEM and CEM-C2 cells with the pGL2TK-HRE vector, which is responsive to hypoxia and DFX stimulation (30) . As shown in Fig. 5A ⇓ , CEM cells expressed fivefold higher levels of luciferase under hypoxia relative to normoxia, and TPT decreased the normoxic and hypoxic levels of luciferase expression by 50% and 70%, respectively. In contrast, addition of TPT did not affect hypoxia-driven luciferase accumulation in CEM-C2 cells, providing evidence that Top 1 is required for the inhibition of HIF-1 activity by TPT.
Consistent with these results, TPT inhibited hypoxic induction of VEGF mRNA expression in CEM cells by 80% but did not affect VEGF mRNA accumulation in CEM-C2 cells (Fig. 5B) ⇓ . Similar results were obtained using other cell lines (P388, P388/CPT45, DU145, and DU145/RC1; data not shown) sensitive or resistant to camptothecins (31) .
To further explore the role of Top 1 in the inhibition of HIF-1 activity by TPT, we used a camptothecin analogue devoid of activity on Top 1 (9-amino-CPT-20R) and a novel Top 1 inhibitor with a noncamptothecin chemical structure (MJ-III-65; Refs. 32 , 33 ). As shown in Fig. 5C ⇓ , U251-HRE cells showed an increase in luciferase accumulation up to 8.8-fold in hypoxic conditions compared with normoxic conditions. Addition of TPT and MJ-III-65 decreased the hypoxia-driven luciferase accumulation by 67.3% and 61.2%, respectively. Contrarily, treatment with the Top 1-inactive derivative 9-amino-CPT-20R did not affect the accumulation of luciferase in response to hypoxia.
Consistent with these findings, TPT and MJ-III-65 inhibited hypoxic induction of VEGF mRNA expression by ∼65%, whereas 9-amino-CPT-20R did not significantly affect VEGF mRNA levels (Fig. 5D) ⇓ . Finally, MJ-III-65 also inhibited the hypoxic induction of HIF-1α protein accumulation by ∼60%, whereas 9-amino-CPT-20R had no effect (Fig. 5E) ⇓ .
These results demonstrate that Top 1 is required for the inhibition of HIF-1α protein accumulation and transcriptional activity. In addition, they indicate that inhibition of HIF-1 may be a property shared by noncamptothecin Top 1 inhibitors.
RNA Transcription, But Not DNA Replication-Dependent DNA Damage, Is Required for the Inhibition of HIF-1α Protein Accumulation by TPT.
Top 1-DNA cleavage complex trapped by camptothecins generates replication-mediated DNA double-strand breaks. Accordingly, blockade of DNA synthesis can prevent the cytotoxicity of camptothecins and the activation of the classical DNA-damaging responses (34) .
To investigate whether inhibition of HIF-1α protein accumulation by TPT was caused by DNA replication-mediated DNA damage, we performed experiments using the DNA replication inhibitor aphidicolin, which was added before TPT in hypoxic or normoxic conditions. As shown in Fig. 6A ⇓ , TPT completely abrogated the hypoxic induction of HIF-1α protein in the absence or presence of aphidicolin. HIF-1β levels were not changed by any of the treatments applied. Experiments performed at 24 h confirmed that aphidicolin had no influence on the ability of TPT to inhibit HIF-1α protein accumulation, suggesting that DNA replication-dependent DNA damage is not required for the inhibition of HIF-1α protein accumulation by TPT (data not shown).
To further explore the potential role of DNA damage in the inhibition of HIF-1α by TPT, we performed experiments using agents that cause DNA double-strand breaks (doxorubicin, a Top 2 inhibitor, and ionizing radiation) to assess whether DNA damage per se was sufficient to inhibit HIF-1α protein accumulation. As shown in Fig. 6, B and C ⇓ , addition of increasing concentrations of doxorubicin and increasing doses of ionizing radiation did not modify HIF-1α protein accumulation under either normoxic or hypoxic conditions in U251 cells. These data demonstrate that DNA double-strand breaks are not sufficient for the inhibition of the hypoxic induction of HIF-1α protein accumulation, suggesting that a replication-independent signaling pathway activated by Top 1 inhibition is involved in the regulation of HIF-1α protein.
The abundance of Top 1-DNA cleavage complexes induced by camptothecins in the actively transcribed regions has led to the hypothesis that RNA transcription also may be involved in the generation of DNA damage or activation of other signaling pathways (35) . It has been shown recently that Top 1-DNA cleavage complexes arrest transcription and trigger transcription-dependent DNA repair (36) . To address the role that transcription-mediated events play in HIF-1α inhibition by TPT, we performed experiments by pretreating U251 cells with known RNA transcription inhibitors (actinomycin D and DRB) before the addition of TPT in normoxic or hypoxic conditions. As shown in Fig. 7A ⇓ , addition of actinomycin D did not affect the oxic or hypoxic accumulation of HIF-1α but totally reversed the effect of TPT. Similar results were obtained in the presence of DRB, which almost completely reversed the inhibition by TPT under oxic (not shown) or hypoxic conditions (Fig. 7B) ⇓ .
Taken together, these results indicate that RNA transcription, but not DNA replication, is required for the inhibition of HIF-1α protein accumulation by TPT.
Camptothecin analogues are a novel class of inhibitors of HIF-1α protein accumulation and transcriptional activity (24) . In this article, we show that (a) TPT does not affect HIF-1α degradation but rather its translation; (b) Top 1 inhibition is required for the inhibition of HIF-1α protein accumulation; and (c) RNA transcription, but not DNA replication, also is required for the inhibition of HIF-1α protein accumulation by TPT. These results reveal the existence of a novel pathway connecting Top 1 and HIF-1α regulation. They also raise the possibility of translating these findings in clinical protocols using TPT to target HIF-1-dependent responses.
HIF-1α is regulated mainly through a proteasome-dependent degradation pathway. We found that TPT inhibited HIF-1α protein accumulation even in the presence of proteasome inhibitors and did not affect HIF-1α protein half-life, indicating that TPT must act on a different step of HIF-1α regulation. Camptothecin analogues can interfere with RNA transcription and elongation and modulate gene expression (37 , 38) . However, in our experimental conditions, TPT did not decrease HIF-1α mRNA expression under either oxic or hypoxic conditions. In contrast, we found that TPT decreases the rate of HIF-1α protein translation. Up-regulation of translation through the PI3k-Akt-mTOR pathway is an important mechanism of HIF-1α protein accumulation induced by growth factors under aerobic conditions (10 , 15) . Interestingly, it also has been reported that TPT inhibits Akt activity in A549 cells, although this effect was observed after prolonged treatment (29) . Under our experimental conditions, the inhibition of HIF-1α protein accumulation by TPT occurred in the absence of inhibition of Akt phosphorylation and the downstream targets of mTOR, suggesting that TPT targets a different pathway involved in HIF-1α translation in these cells.
Although the precise pathway affected by TPT remains elusive, it is conceivable that its activity is not limited to HIF-1α. However, we found that TPT did not inhibit accumulation of other proteins, such as HIF-1β, Akt, 4EBP1, p70S6-kinase, and c-myc, and did not significantly affect total protein synthesis as measured by 35S incorporation (data not shown), suggesting that TPT is not a general inhibitor of protein translation. Recent evidence indicates that HIF-1α mRNA contains an internal ribosome entry site that allows efficient translation during hypoxia (39) . An attractive possibility is that the inhibition of HIF-1α translation by TPT may be linked to the modulation of its internal ribosome entry site region. Additional studies are required to elucidate which, if any, other proteins may be affected by TPT.
Top 1 inhibitors (camptothecins) target Top 1 by inducing the formation of stable Top 1-DNA cleavage complexes (40) . Therefore, a critical question was whether Top 1 poisoning was required for the inhibition of HIF-1α protein accumulation by TPT. We provide evidence that shows conclusively that Top 1 is the biochemical target linking the activity of TPT to the inhibition of HIF-1α. First, cells resistant to TPT, because of a genetic mutation of Top 1, are no longer sensitive to the inhibition of HIF-1α protein accumulation by TPT. These cells have been used extensively as a model of resistance to camptothecin analogues, and they lack the chemical responses in cytotoxicity and signaling associated with the activity of camptothecin analogues (25) . Second, a 20R isomer of 9-amino-CPT inactive toward Top 1 also lacks the ability to inhibit HIF-1α protein accumulation. Conversely, the indenoisoquinoline MJ-III-65, a novel Top 1 inhibitor, with a chemical structure distinct from the camptothecin analogues, mimics TPT not only in the inhibition of HIF-1α protein accumulation but also in the regulation of VEGF and cyclooxygenase-2 mRNA (which are inhibited and induced, respectively, by TPT and MJ-III-65; data not shown). Thus, our results suggest that the inhibition of HIF-1α protein accumulation and transcriptional activity may be a common feature of drugs that affect directly or indirectly Top 1 activity. This finding is relevant because many cytotoxic compounds that interact with DNA also may affect Top 1, although they are not recognized or clinically used as Top 1 inhibitors.
Top 1 inhibitors (camptothecins) trap Top 1-DNA cleavage complexes. This stabilized state is characterized by the presence of a single-strand break in the DNA that reverses within minutes after drug removal (41) . Studies using these agents have demonstrated that the DNA replication process can convert the Top 1-DNA cleavage complexes into replication-mediated DNA double-strand breaks that arrest cell division and cause cell death, and this is believed to be the major mechanism responsible for the anticancer activity of camptothecins (41) . The collision between the DNA replication fork and the Top 1-DNA cleavage complex is responsible for S-phase-specific cytotoxicity, G2 cell cycle arrest, activation of nuclear factor κB and ataxia telangiectasia mutated and Rad3 related protein (ATR), and stabilization of p53, which are all typical DNA damage responses (42) . Blockade of DNA synthesis by inhibitors, such as aphidicolin, prevents the cytotoxicity of camptothecins (40) . In this report, we show that preventing replication-mediated DNA damage by aphidicolin did not prevent the inhibition of HIF-1α protein accumulation by TPT, suggesting that DNA replication is not essential for this effect. The lack of an effect of DNA replication blockade on HIF-1α inhibition by TPT could be explained by at least two possibilities: one is that the DNA replication-dependent DNA damage does not trigger the inhibition of HIF-1α protein accumulation, suggesting that DNA damage is not involved in this phenomenon; the second is that the extent of DNA damage caused by TPT under our experimental conditions was not sufficient to trigger HIF-1α inhibition, raising the possibility that classical DNA-damaging agents might be able to activate this response. However, we found that classical DNA-damaging agents, such as the Top 2 inhibitor doxorubicin and ionizing radiation, did not inhibit HIF-1α protein accumulation, strongly suggesting that DNA double-strand breaks are not involved in the down-regulation of HIF-1α. These results demonstrate clearly that there is a unique pathway connecting Top 1 to HIF-1α protein accumulation, which is independent of DNA replication-dependent DNA damage.
The abundance of Top 1-DNA cleavage complexes induced by camptothecins in the actively transcribed regions has led to the hypothesis that the Top 1-mediated DNA cleavage complexes can undergo collision with the RNA transcription machinery and generate specific responses (35) . It has been shown recently that Top 1-DNA cleavage complexes arrest transcription and trigger transcription-dependent DNA repair (36) . We hypothesized that RNA transcription might be necessary for HIF-1α inhibition by TPT. Accordingly, we found that blockade of RNA transcription using actinomycin D, DRB, or flavopiridol (a CDK inhibitor that also blocks RNA transcription; Ref. 43 ; data not shown) blocked completely the ability of TPT to inhibit HIF-1α protein accumulation. Interestingly, inhibitors of RNA synthesis did not block the hypoxic induction of HIF-1α protein, consistent with data published by others (44) . These findings indicate clearly that there is a requirement for ongoing RNA transcription linking Top 1 poisoning and HIF-1α down-regulation. There are at least two possible explanations for this result. The first hypothesis is that Top 1 poisoning induces the transcription of a mediator that is required to block HIF-1α protein translation. Although we cannot formally exclude this possibility, we found that TPT inhibits HIF-1α protein accumulation at as early as 1 h (data not shown), making the possibility that transcriptional activation of a mediator is responsible for HIF-1α inhibition unlikely. The second hypothesis is that the presence of Top 1-DNA cleavage complexes on the transcribed strand of active genes arrests the RNA transcription machinery and generates a signal that ultimately inhibits HIF-1α protein accumulation. This hypothesis is highly attractive because it has been suggested that, following the stalling of the transcription machinery, RNA polymerase II triggers signaling responses, including rapid DNA repair (45) . Interestingly, von Hippel-Lindau protein, which is required for ubiquitination of HIF-1α, also targets hyperphosphorylated RNA polymerase II for ubiquitination and degradation through a similar recognition motif containing hydroxylated proline (46) .
Hypoxia may contribute to resistance to S-phase selective agents, such as camptothecin, that require ongoing DNA replication to exert their cytotoxic activity (47) . In contrast, we provide evidence that the ability of TPT to decrease HIF-1α protein accumulation is a cell cycle-independent effect because it is manifested in hypoxic cells, cells treated with aphidicolin (S-phase), or with the microtubule depolymerizing agent nocodazole (G2-M phase; data not shown). More importantly, our data indicate clearly that the mechanism by which TPT exerts its cytotoxic effect is distinct from the one leading to HIF-1α inhibition. Overall, these results provide a rationale for using TPT in clinical regimens aimed at targeting the hypoxic fraction of solid tumors and HIF-1-dependent responses.
Grant support: Federal funds from the National Cancer Institute, National Institutes of Health, under contract no. N01-CO-12400 (Article H.36 of the Prime Contract).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Notes: The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the United States Government.
Requests for reprints: Giovanni Melillo, DTP-Tumor Hypoxia Laboratory, Building 432, Room 218, National Cancer Institute at Frederick, Frederick, Maryland 21702. Phone: 301-846-5050; Fax: 301-846-6081; E-mail:
- Received October 6, 2003.
- Revision received December 8, 2003.
- Accepted December 12, 2003.
- ©2004 American Association for Cancer Research.