Because angiogenesis is essential for tumor growth and metastasis, inhibition of angiogenesis has emerged as a new therapy to treat cancers. Hypoxia-induced expression of vascular endothelial growth factor (VEGF) plays a central role in tumor-induced angiogenesis. In this study, we found that expression of VEGF in hypoxic tumor cells was affected by the circadian organization of molecular clockwork. The core circadian oscillator is composed of an autoregulatory transcription–translation feedback loop in which CLOCK and BMAL1 are positive regulators, and Period and Cryptochrome genes act as negative ones. The levels of VEGF mRNA in tumor cells implanted in mice rose substantially in response to hypoxia, but the levels fluctuated rhythmically in a circadian fashion. Luciferase reporter gene analysis revealed that Period2 and Cryptochrome1, whose expression in the implanted tumor cells showed a circadian oscillation, inhibited the hypoxia-induced VEGF promoter activity. These results suggest that the negative limbs of the molecular loop periodically inhibit the hypoxic induction of VEGF transcription, resulting in the circadian fluctuation of its mRNA expression. We also showed that the antitumor efficacy of antiangiogenic agents could be enhanced by administering the drugs at the time when VEGF production increased. These findings support the notion that monitoring of the circadian rhythm in VEGF production is useful for choosing the most appropriate time of day for administration of antiangiogenic agents.

Angiogenesis, the formation of new blood vessels from existing ones, is an indispensable biological event for a variety of physiological and pathological processes (1, 2). Because tumors require a supply of blood to nourish and facilitate metastasis, angiogenesis is essential for the growth of small tumors to large ones that can continue to metastasize (3). Tumor-induced angiogenesis is regulated by cell-produced factors that have mitogenic and chemotactic effects on vascular endothelial cells. VEGF3 plays a central role in tumor-induced angiogenesis (4, 5). VEGF regulates both vascular proliferation and permeability and acts as an antiapoptotic factor for newly formed blood vessels. The biological effects of VEGF are mediated by two distinct receptors, VEGFR-1 and VEGFR-2, whose expression is limited to vascular endothelial cells (6, 7). In malignant tumors, VEGF is expressed at substantially increased levels, and its expression is often associated with poor prognosis in several types of cancers. Although the expression of VEGF occurs in response to various stimuli, hypoxia-induced expression of VEGF plays a central role in tumor-induced angiogenesis (8).

Although the role of angiogenesis in tumor growth and metastasis is well recognized, no angiogenesis inhibitor has been approved for clinical use. Antiangiogenic therapy offers several benefits, including lack of drug resistance, lack of synergistic interaction with other drugs, and low toxicity compared with conventional antitumor agents. However, the results of preclinical and early clinical trials suggest that the onset of inhibitor activity can be delayed. In patients with advanced cancers, this delay may result in stabilization of the disease, but not regression (9).

One approach for increasing the efficacy of pharmacotherapy is to administer the drugs at the time of day when they are most effective and/or best tolerated. A chronopharmacological strategy can enhance the effects of drugs and/or attenuate their toxicity (10, 11, 12). Daily variations in biological functions such as gene expression and protein synthesis are thought to be important factors affecting the efficacy of drugs. In mammals, the master circadian pacemaker resides in the suprachiasmatic nucleus of the anterior hypothalamus. It is responsible for adapting endogenous physiological functions to cyclic environmental cues such as light, temperature, and social communication (13). Recently, several clock genes have been identified that control an array of circadian rhythms in physiology and behavior. According to the currently held model, the core circadian oscillator consists of an autoregulatory transcription–translation feedback loop. The basic helix-loop-helix-PAS domain proteins CLOCK and BMAL1 form a heterodimer and then activate transcription of the Per and Cry genes. Once the PER and CRY proteins have reached a critical concentration, they attenuate CLOCK/BMAL1-mediated activation of their own genes in a negative-feedback loop (14, 15).

In this study, we found that the circadian organization of molecular clockwork affects VEGF gene transcription in hypoxic tumor cells. The levels of VEGF mRNA, as well as the mRNA for a set of clock genes, exhibited a circadian oscillation in tumor cells implanted in mice. The negative limbs of the molecular loop governed the rhythmic change in hypoxia-induced VEGF gene transcription, thereby leading to a circadian fluctuation in its protein production. We therefore investigated how the rhythmic variation of VEGF production influenced the pharmacological efficacy of antiangiogenic agents.

Materials.

Angiogenesis inhibitors SU 1498 and BB2516 were purchased from Calbiochem (San Diego, CA). The compounds were dissolved in sterilized saline containing 5% DMSO for treatment. TNP-470 was provided as lyophilized powder by Takeda Chemical Industrial, Ltd. (Osaka, Japan). For treatment, the compound was suspended in sterile saline containing 3% ethanol and 5% gum arabic.

Animals and Cells.

Male ICR mice (5 weeks of age) were purchased from Charles River Japan, Inc. (Kanagawa, Japan). They were housed under a standardized light/dark cycle at room temperature of 24°C ± 1°C and a humidity of 60 ± 10% with food and water ad libitum. The animals were adapted to the light/dark cycle for 2 weeks before the experiments. Under the light/dark cycle, ZT 0 was designated as lights on and ZT12 as lights off. During the dark period, a dim red light was used to aid treatment of the mice. Two murine tumor cells lines (sarcoma 180 and B16 melanoma) were obtained commercially from the Dainippon Pharmaceutical Co. Ltd. (Osaka, Japan). Lewis lung carcinoma cells were supplied by the Cell Resource Center for Biomedical Research, Tohoku University (Sendai, Japan). These tumor cells were maintained in DMEM supplemented with 10% fetal bovine serum at 37°C in a humidified 5% CO2 atmosphere. A 50-μl volume containing 1.5 × 106 viable tumor cells was injected in the right hind footpads of each mouse. The tumor volume was estimated according to the following formula: tumor volume (mm3) = 4π(XYZ)/3, where 2X, 2Y, and 2Z are the three perpendicular diameters of tumor.

Quantitative RT-PCR Analysis.

After the tumor size reached ∼200 mm3, the total RNA was extracted from the sarcoma 180 tumor masses by use of TRIzol reagent (Invitrogen, Carlsbad, CA) at ZT 2, ZT 6, ZT 10, ZT 14, ZT 18, and ZT 22. The cDNA of mouse Per1 (GenBank accession no. AF022992), Per2 (GenBank accession no. AF035830), Cry1 (GenBank accession no. AF156986), Cry2 (GenBank accession no. AF156987), Clock (GenBank accession no. AF000998), Bmal1 (GenBank accession no. AB014494), VEGF (GenBank accession no. M32977), and GAPDH (GenBank accession no. M88354) was synthesized and amplified with use of a superscript one-step RT-PCR system (Invitrogen). To evaluate the quantitative reliability of RT-PCR, we performed a kinetic analysis of amplified products to ensure that signals were derived only from the exponential phase of amplification. From each sample after the first 25 cycles of amplification, we drew a 5-μl aliquot for electrophoresis, and submitted the tubes to one more cycle of PCR. This procedure was repeated for a total of 30 cycles. The PCR products were run on 3% agarose gels. After staining with ethidium bromide, the gel was photographed with Polaroid-type film. The density of each band was analyzed with use of NIH image software on a Macintosh computer.

The exponential phase of GAPDH amplification in all experimental conditions occurred between the 26th and the 28th cycles, and the exponential phases of all target genes (clock genes and VEGF) occurred between the 27th and the 30th cycles. The amplification efficiencies of the GAPDH and clock or VEGF genes were comparable. The amplification products were therefore collected and quantified at the 27th or 28th cycle. The ratio of the amplified target to the amplified internal control (calculated by dividing the value of each Per, Cry, Clock, Bmal1, or VEGF by that of GAPDH) was compared among groups.

Transcriptional Assay.

One day before transfection, sarcoma 180 cells were seeded (1 × 105/well) in 6-well plates containing DMEM supplemented with 10% fetal bovine serum. Cells were transfected with 10 ng of reporter constructs and 1.0 or 1.2 μg (total) of expression constructs, using LipofectAmine-Plus reagent (Invitrogen) according to the manufacturer’s instructions. To correct for variations in transfection efficiency, 0.1 ng of pRL-TK vector (Promega, San Luis Obispo, CA) was cotransfected in all experiments. The total amount of DNA/well was adjusted by adding pcDNA 3.1 vector (Invitrogen). At 48 h after transfection, cell extracts were prepared with use of 200 μl of passive lysis buffer (Promega), and 20 μl of the extracts were taken for assays of firefly luciferase and Renilla luciferase by luminometry. The ratio of firefly luciferase activity (expressed from reporter construct) to Renilla luciferase activity (expressed from pRL-TK) in each sample served as a measure of normalized luciferase activity.

For transcriptional assay of the Per1 gene, a 1289-bp upstream fragment of mouse Per1 gene was isolated from ICR mouse genomic DNA by PCR amplification and subcloned into the pGL3-basic luciferase vector (Promega). Similarly, VEGF reporter constructs were made by ligating a 1218-bp upstream fragment of mouse VEGF gene to the pGL3-basic luciferase vector (Promega). The HRE of the VEGF promoter was mutated at −917 to −912 (TACGTG to TCATTC) by use of the QuickChange site-directed mutagenesis Kit (Stratagene, La Jolla, CA). Expression constructs were made as follows: the cording regions of mouse Clock, Bmal1, Per2, Cry1, HIF-1α (GenBank accession no. AF003695), and ARNT (GenBank accession no. U10325) were obtained by RT-PCR and used after their sequences were confirmed. All cording regions were ligated into the pcDNA 3.1 vector.

Western Blot Analysis.

Nuclear fractions from the implanted sarcoma 180 tumor cells or healthy tissues (brain, liver, kidney, and skeletal muscle) were prepared as follows. Tumor-bearing mice were deeply anesthetized with ether. Tumor masses and healthy tissues were removed and homogenized with ice-cold lysis buffer [20 mm Tris-HCl (pH 7.8), 150 mm NaCl, 5 mm EDTA, 0.5% NP40, 2 mm phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, and 10 μg/ml antipain]. After removal of the soluble fraction by centrifugation at 1000 × g for 10 min, the pellet was washed three times with PBS. After a final centrifugation, the pellets were resuspended in nuclear extraction buffer at 4°C for 30 min; nuclear fractions were then obtained by centrifugation for 10 min at 12,000 × g. The lysates containing 20 μg of total protein were resolved by SDS-PAGE and transferred to a polyvinylidene difluoride membrane. The membrane was reacted with antibodies against HIF-1α (Novus Biologicals, Litteton, CO), PER2 (Alpha Diagnostic International, San Antonio, TX), or CRY1 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Specific antigen–antibody complexes were visualized by use of horseradish peroxidase-conjugated secondary antibodies and Super Signal Chemiluminescent Substrate (Pierce Biotechnology, Inc., Rockford, IL).

Immunoprecipitation.

Nuclear fractions from the implanted sarcoma 180 tumor cells were diluted with 10 volumes of the lysis buffer as described above. The diluted fractions (100 mg of total proteins) were incubated with 20 μl of protein G-Sepharose beads (Roche Diagnostics, Mannheim, Germany) for 3 h at 4°C, and the suspension was centrifuged. An antibody against HIF-1α (10 μl) plus protein G-Sepharose beads were added to the clarified supernatants. After gentle agitation for 12 h at 4°C, the beads were collected by centrifugation. The immune complexes were washed three times, mixed with sample buffer, boiled, and centrifuged. The immune complexes in the supernatants were analyzed by Western blotting to detect PER2 or CRY1.

Determination of VEGF Concentration.

To explore the temporal variations in VEGF production, we measured the amounts of VEGF protein in plasma, sarcoma 180 tumor masses, and liver at each of the six times outlined above. After the tumor volume reached ∼200 mm3, blood samples drawn by cardiac puncture were placed in polypropylene tubes containing 4% EDTA solution. Plasma samples were obtained by centrifugation at 3000 rpm for 3 min and stored at −20°C until assay. Immediately after blood sample collection, the tumor masses and livers were removed and homogenized with 1 ml of the lysis buffer. After removal of insoluble materials by centrifugation at 12,000 × g for 15 min at 4°C, the resulting supernatants were assayed to determine VEGF concentrations. The VEGF concentrations in plasma and tumor masses were determined by an ELISA (R&D systems, Minneapolis, MN). The VEGF concentration in plasma is expressed in pg/ml, and the amount of VEGF in tumor masses is expressed in ng/mg of protein. The protein concentration was determined with use of a detergent-compatible protein assay kit (Bio-Rad, Hercules, CA). We also examined the temporal variations in VEGF content in three different types of tumor cells (sarcoma 180, Lewis lung carcinoma, and B16 melanoma) and explored whether the time dependence of VEGF accumulation is affected by the gaining of tumor volume.

Determination of Antitumor Effect.

Tumor-bearing mice received s.c. injections containing a single daily dose of SU1498 (25 μg/mouse), TNP-470 (30 mg/kg), or BB2516 (10 mg/kg) once every other day at ZT 2 or ZT 14. The dosages of antiangiogenic agents were selected based on previous reports (16, 17, 18). Control mice were injected with vehicle alone. All three types of antiangiogenic agents were injected starting 3–5 days after inoculation with tumor cells. In all mice, tumor volumes were measured every 4 days throughout the duration of the experiment.

Statistical Analysis.

The significance of the 24-h variation in each parameter was tested by ANOVA. The statistical significance of differences among groups was analyzed by ANOVA and the Tukey multiple comparison test. A 5% level of probability was considered to be significant.

Circadian Expression of Clock and VEGF Genes in Implanted Tumor Cells.

The expression of clock gene in implanted cells is subordinated to the dominance exerted by the central clock of the host animal (19). As shown in Fig. 1 A, mRNA levels for the Per1, Per2, Cry1, and Cry2 genes in implanted tumor cells showed circadian oscillation. For Per1, Per2, and Cry2, mRNA levels peaked around the early dark phase, whereas Cry1 mRNA levels peaked around the late dark phase. The amplitude of the Cry mRNA rhythms was smaller than that of the Per1 and Per2 oscillations. Bmal1 and Clock mRNA oscillations were antiphase to those of the Per and Cry2 mRNA rhythms, with peak levels from the late dark phase to the early light phase. These clock gene expression patterns are similar to those reported previously in liver and skeletal muscle (15, 20, 21, 22).

As the result of alternative splicing, the gene for murine VEGF encodes three spliced isoforms of VEGF proteins of 120, 164, and 188 amino acids (23). The mRNA levels for VEGF120 and VEGF164 in the implanted tumor cells exhibited significant circadian oscillations (P < 0.01). Their mRNA levels peaked during the light phase and decreased around the early dark phase (Fig. 1 B). However, no obvious expression of VEGF188 was observed in the sarcoma 180 tumor cells.

Regulation of VEGF Transcription by Clock Genes.

The clock genes, consisting of core oscillation loops, control downstream events by regulating the expression of clock-controlled output genes. CLOCK/BMAL1 heterodimers act through an E-box enhancer element of the output genes to activate transcription, and the activation is inhibited by PER and CRY proteins (14, 15). We therefore examined the promoter sequence of the mouse VEGF gene. Four E-box elements (CANNTG) are found within 1218-kbp of the mouse VEGF 5′-flanking region (Fig. 2 A). However, none of these E-boxes corresponds to the CACGTGA sequence, the strict consensus binding site for CLOCK/BMAL1 heterodimers (24).

To examine whether circadian clock genes directly regulate the transcription of VEGF, we performed in vitro transient transcription assays, using two reporter constructs. Per1-Luc is a luciferase reporter driven by the 1289-bp mouse Per1 promoter that contains three E-box consensus sequences. VEGF-Luc is a luciferase reporter driven by 1218-bp mouse VEGF promoter. The sarcoma 180 cells were transfected with the two reporters in separate experiments. As shown in Fig. 2 B, cotransfection of the Per1-Luc reporter with either CLOCK or BMAL1 alone had little effect on the promoter activity, whereas cotransfection of the Per1-Luc reporter with both CLOCK and BMAL1 resulted in a 12-fold increase in the transcriptional activity. As reported earlier, PER2 protein partially inhibited the CLOCK/BMAL1-induced transcription, whereas the CRY1 protein inhibited the induction of Per1 transcription at basal level.

Compared with Per1-Luc reporter activity, cotransfection of the VEGF-Luc reporter with both CLOCK and BMAL1 caused a small increase (2.8-fold) in the promoter activity, whereas the activation of VEGF-Luc reporter by CLOCK/BMAL1 was repressed by PER2 and CRY1 (Fig. 2 B). Although these in vitro data may partly account for the circadian oscillation of VEGF gene expression in the implanted tumor cells, other transcriptional factors may also participate in the regulation of circadian oscillation of VEGF.

Regulation of VEGF Transcription by Clock Genes under Hypoxia.

Solid tumors often have hypoxic areas in which circulation is compromised because of structurally disorganized blood vessels (3). Hypoxia-induced expression of VEGF plays a central role in neovascularization, which is essential for tumor growth beyond 1–2 mm3. We therefore also examined the effect of circadian clock genes on VEGF transcription under hypoxic conditions. As shown in Fig. 2,C, incubation of the sarcoma 180 cells under hypoxic conditions (5% O2) resulted in a significant (7-fold) induction of VEGF transcription. However, cotransfection with CLOCK and BMAL1 had no significant effect on this hypoxia-induced activation response. By contrast, cotransfection of PER2 inhibited, in a dose-dependent manner, VEGF promoter activity induced by hypoxia. A similar dose-dependent inhibition was also observed when cells were transfected with CRY1 expression plasmid (Fig. 2 D). These results suggest that PER2 and CRY1 act as negative regulators of VEGF transcription under hypoxic conditions.

Accumulation of HIF-1α Protein in Implanted Tumor Cells.

HIF-1α, one of the basic helix-loop-helix-PAS transcriptional proteins, is thought to be the major regulator of VEGF transcription in response to hypoxia. HIF-1α dimerizes with ARNT and acts through HRE to activate the transcription of VEGF (25, 26). Although HIF-1α protein is undetectable in most cell types because of the rapid degradation by the ubiquitin-proteasome system, hypoxia induces significant accumulation of HIF-1α protein (8). As shown in Fig. 3,A, we observed extensive expression of HIF-1α protein in sarcoma 180 tumor masses. The levels of HIF-1α protein were consistently increased throughout the day (Fig. 3 B). The accumulation of HIF-1α protein seems to reflect a hypoxic response of the tumor cells.

PER2 and CRY1 Inhibit HIF-1α/ARNT-Induced VEGF Promoter Activity.

To examine whether HIF-1α participates in the regulation of circadian oscillation of VEGF gene transcription, we performed in vitro transient transcription assays in cultured sarcoma 180 cells. The assay was executed under hypoxic conditions (5% O2) to avoid the degradation of HIF-1α protein. As shown in Fig. 4 A, cotransfection of VEGF-Luc reporter with HIF-1α resulted in an ∼15-fold increase in VEGF promoter activity. The transcriptional activity was further enhanced by cotransfection with ARNT expression plasmid (30-fold). This HIF-1α/ARNT-induced activation was abolished when HRE was mutated (data not shown).

We next investigated the effect of PER2 and CRY1 on the HIF-1α/ARNT-induced activation of VEGF transcription under hypoxic conditions. Both PER2 and CRY1 repressed, in a dose-dependent manner, the HIF-1α/ARNT-induced VEGF promoter activity. Cotransfection with 1.0 μg of CRY1 plasmid partially inhibited the HIF-1α/ARNT-mediated transcription, whereas cotransfection with the same amount of PER2 plasmid suppressed the induction of VEGF promoter activity at the basal level (Fig. 4 B). These results suggest that the negative limbs of the molecular loop repress HIF-1α/ARNT transactivation, thereby preventing the hypoxic induction of VEGF transcription.

Time-Dependent Interaction of PER2 with HIF-1α.

Both PER2 and CRY1 proteins in tumor nuclei showed a clear circadian accumulation (Fig. 3 C). The levels of PER2 protein peaked around the early dark phase, whereas CRY1 protein peaked from the late dark phase to the early light phase. The cyclic accumulation of PER2 protein was nearly antiphase to that of VEGF mRNA expression. Immunoprecipitation experiments revealed that PER2 protein precipitated together with HIF-1α in tumor nuclei and that the amount of PER2 associated with HIF-1α was much greater at ZT 14 compared with that at ZT 2. In contrast, a small amount of CRY1 protein precipitated together with HIF-1α at both times, but the amount of CRY1 associated with HIF-1α did not differ significantly between the two times. These results indicate that PER2 protein interacts with HIF-1α in a time-dependent manner. The dark-phase increase in PER2–HIF-1α interactions may contribute to the attenuation of hypoxia-induced transcriptional activity of VEGF.

Circadian Variation of VEGF Production in Implanted Tumor Cells.

Because the mRNA levels for the VEGF gene showed a clear circadian oscillation, we explored the time dependence of VEGF protein production in the implanted tumor cells. As shown in Fig. 5,A, VEGF content in sarcoma 180 tumor masses exhibited a significant circadian oscillation, with higher levels during the light phase and lower levels during the dark phase (P < 0.01). The amplitude of the VEGF fluctuation in tumor masses was significantly larger than that in healthy liver. Similarly, a large amplitude in rhythmic variation was also observed in the VEGF concentration in the plasma of tumor-bearing mice (P < 0.01; Fig. 5 B). The rhythmic pattern resembled the overall increases and decreases in VEGF content in sarcoma 180 tumor masses.

We next investigated whether the circadian oscillation of VEGF was affected by the gaining of tumor volume. This experiment was performed with three different types of tumor cells (sarcoma 180, Lewis lung carcinoma, and B16 melanoma). Although the tumor volumes were <200 mm3, significant diurnal variations in VEGF content were observed in all three types of tumor cells (Fig. 5 C). The protein abundances of VEGF in the tumor masses were significantly higher at ZT 2 compared with those at ZT 14 (P < 0.01). The amounts of VEGF protein in all three types of tumor cells substantially increased with the gaining of tumor volume. At a tumor volume >800 mm3, VEGF content in the tumor masses still exhibited significant diurnal variations (P < 0.01 for sarcoma 180 and B16 melanoma; P < 0.05 for Lewis lung carcinoma).

Influence of Dosing Time on Antitumor Effect of Angiogenesis Inhibitors.

Because VEGF production in tumor masses varied diurnally, we examined whether this diurnal variation affects the pharmacological efficacy of antiangiogenic agents. The activation of the VEGF signaling pathway successively induces various downstream events, such as the proliferation of endothelial cells and the activation of matrix metalloproteinase (27, 28). Thus, three different types of antiangiogenic agents, SU1498 (VEGFR-2 tyrosine kinase inhibitor), TNP-470 (endothelial cell growth inhibitor), and BB2516 (matrix metalloproteinase inhibitor), were tested in this study. Fig. 6 shows the influence of dosing time on the ability of these antiangiogenic agents to inhibit tumor growth. Because no significant time-dependent difference was observed in the growth of tumor cells in mice treated with vehicle alone, the mean value of the tumor volume between ZT 2 and ZT 14 was shown as the control. The growth of tumor cells was significantly suppressed by administration of antiangiogenic agents. The antitumor effects of all three types of antiangiogenic agents were more potent in mice in which the drugs were injected at ZT 2 than in mice receiving the drugs at ZT 14. These results suggest that the antitumor effects of antiangiogenic agents could be improved by optimizing the dosing schedule.

Recent molecular dissection of the circadian biological clock system has revealed that oscillation in the transcription of specific clock genes plays a central role in the generation of circadian rhythms. The present study demonstrated that the circadian organization of molecular clockwork governs the circadian rhythm of VEGF production in hypoxic tumor cells. In this mechanism, transcription of Per and Cry genes is activated by CLOCK/BMAL1 heterodimer. The negative limbs of the feedback loop, PER and CRY proteins, repress their own transcription. At the same time, these negative components inhibit the hypoxia-induced activation of VEGF transcription. The inhibitory action appears to be elicited by preventing HIF-1α/ARNT-mediated transactivation. The time-dependent appearance and disappearance of VEGF transcripts seem to be the underlying cause of the circadian fluctuation of its protein production. Because rapid degradation makes HIF-1α protein undetectable in normal cells, this mechanism may be applicable only to hypoxic cells such as malignant tumors. This hypothesis is supported by the present findings that the magnitude of VEGF oscillation in the tumor cells was much higher than that in healthy liver. The potent activation of VEGF transcription induced by HIF-1α/ARNT may contribute to its high oscillation magnitude.

Because the levels of HIF-1α protein in tumor masses consistently increased throughout the day, it is unlikely that the inhibitory effect of PER2 or CRY1 on the hypoxia-induced activation of VEGF transcription is caused by the decrease in the expression of HIF-1α. The clock gene products, except for Cry1 and Cry2, are called PAS proteins, because they all have PAS domains that function as a dimerization surface for protein-protein interaction (29, 30). The importance of this domain in the circadian clock system is demonstrated by the fact that homozygous Per2 mutant mice, in which the Per2 gene encodes the protein lacking the PAS domain, show significant differences in circadian locomotor activity and altered expression of clock genes (31). It has been suggested that there is cross-talk between members of the PAS family of proteins participating in hypoxic and circadian systems. For example, PER1 protein is stabilized through a protein–protein interaction with HIF-1α under hypoxic conditions (32). In this study, we found that PER2 precipitated together with HIF-1α in tumor nuclei and that the amount of PER2 associated with HIF-1α was greater at ZT 14 compared with that at ZT 2. Because PER2 itself has no DNA-binding motif (33), the PER2-caused inhibition of the HIF-1α/ARNT transcriptional activation may be dependent on a protein–protein interaction.

CRY1 protein also inhibited the HIF-1α/ARNT-mediated transactivation, but the inhibitory mechanism is presumed to be distinct from that of PER2. CRY1 protein represses CLOCK/BMAL1-induced transactivation by preventing H3 histone acetylation (34). The CLOCK/BMAL1 heterodimer exerts its transcriptional activity by forming a transcriptional coactivator complex with p300 histone acetyltransferase. CRY1 protein disrupts the transcriptional coactivator complex, p300, thereby reducing histone acetyltransferase activity and altering chromatin structure to decrease CLOCK/BMAL1 transcriptional activation. The p300 histone acetyltransferase also participates in HIF-1α/ARNT-mediated transactivation (35). Therefore, CRY1-caused inhibition of HIF-1α/ARNT transcriptional activation may be attributable to the result of prevention of H3 histone acetylation.

Although both PER2 and CRY1 proteins inhibited HIF-1α/ARNT-induced VEGF transactivation, the results of immunoprecipitation experiments and luciferase reporter gene analysis indicated that PER2 protein acts as a major regulator of circadian expression of VEGF mRNA in hypoxic tumor cells. In fact, circadian accumulation of PER2 protein was nearly antiphase to that of VEGF mRNA expression, and the protein levels peaked around the early dark phase (ZT 14 and 18). The dark-phase increase in PER2–HIF-1α interactions appears to cause the inhibition of HIF-1α/ARNT-induced VEGF transcription, and the inhibitory effect may then be relieved by decreasing the amount of PER2 protein around the late dark phase.

The daily rhythmic variations in biological functions such as gene expression and protein synthesis are considered to be a critical factor influencing the effectiveness of drugs. Thus, it is important to determine whether circadian variations in VEGF production affect the pharmacological efficacy of antiangiogenic agents. The antitumor effects of three different types of antiangiogenic agents (SU1498, a VEGFR-2 tyrosine kinase inhibitor; TNP-470, an endothelial cell growth inhibitor; and BB2516, a matrix metalloproteinase inhibitor) were enhanced when they were administered during early light phase rather than during early dark phase. VEGF is an endothelial cell-specific mitogen that acts through two tyrosine kinase receptors, VEGFR-1 and VEGFR-2. Because VEGFR-2 is the major signal-transducing receptor for VEGF (23), much effort has gone into the development of VEGFR-2 tyrosine kinase inhibitor. The activation of the VEGFR-2 signaling pathway induces migration and proliferation of endothelial cells in the tumor. Matrix metalloproteinases, which are also activated by VEGF stimulation, contribute to the degradation of extracellular matrix and basement membrane (28, 36). Thus, the time-dependent changes in antitumor efficacies of all three different types of antiangiogenic agents are, at least in part, attributable to the circadian variation in VEGF production in tumor cells.

It has recently been reported that the plasma concentrations of VEGF in diurnally active humans with POEMS (polyneuropathy, organomegaly, endocrinopathy, m-protein skin changes) syndrome show circadian fluctuation; the plasma VEGF levels peaked at night and decreased in the daytime (37). The findings correspond to the present observation indicating that the plasma concentrations of VEGF in nocturnally active rodents increase during the rest period. These facts support the notion that monitoring the rhythmic variation in VEGF secretion is useful for choosing the most appropriate time of day for administration of antiangiogenic agents. The effectiveness and toxicity of many drugs vary with the dosing time and are associated with 24-h rhythms of various biological processes under the control of the circadian clock. However, many drugs are still administered without regard to the time of day. Identification of a rhythmic marker for selecting dosing time of antiangiogenic agents may be an aid to achieve rational chronopharmacotherapy for treatment of cancers.

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.

1

This study was supported by a Grant-in Aid for Encouragement of Young Scientists from the Japan Society for the Promotion of Science (13771448 to S. K.), a Grant-in Aid from the Japan Research Foundation for Clinical Pharmacology (to S. K.), a Grant-in Aid from the Takeda Science Foundation (to S. K.), and a Grant-in Aid from the Uehara Memorial Foundation (to S. K.).

3

The abbreviations used are: VEGF, vascular endothelial growth factor; VEGFR, vascular endothelial growth factor receptor; PAS, Per, Arnt, Sim; PER and Per, period protein and gene, respectively; CRY and Cry, cryptochrome protein and gene, respectively; ZT, zeitgeber time; RT-PCR, reverse transcription-PCR; GAPDH, glyceraldehyde-3-phosphate; HRE, hypoxia-responsive element; HIF-1α, hypoxia-induced factor-1α; ARNT, aryl hydrocarbon receptor nuclear translocator.

Fig. 1.

Circadian expression of clock genes and VEGF in implanted tumor cells. A, representative electrophoretic image of RT-PCR products of clock genes in tumor masses. GAPDH mRNA was used as an internal control for transcripts whose expression was constant throughout the day. The horizontal bar at the bottom indicates light and dark cycles. B, temporal profiles of mRNA expression of VEGF164 and VEGF120 in tumor masses. For plots of RNA, the mean peak values for VEGF164 and VEGF120 are set at 100. Each point represents the mean ± SE (bars; n = 4–6). The mRNA levels for both VEGF164 and VEGF120 exhibit significant circadian variations (P < 0.01, respectively; ANOVA). The upper panel shows a representative electrophoretic image of RT-PCR products of the VEGF gene. The horizontal bar at the bottom indicates light and dark cycles.

Fig. 1.

Circadian expression of clock genes and VEGF in implanted tumor cells. A, representative electrophoretic image of RT-PCR products of clock genes in tumor masses. GAPDH mRNA was used as an internal control for transcripts whose expression was constant throughout the day. The horizontal bar at the bottom indicates light and dark cycles. B, temporal profiles of mRNA expression of VEGF164 and VEGF120 in tumor masses. For plots of RNA, the mean peak values for VEGF164 and VEGF120 are set at 100. Each point represents the mean ± SE (bars; n = 4–6). The mRNA levels for both VEGF164 and VEGF120 exhibit significant circadian variations (P < 0.01, respectively; ANOVA). The upper panel shows a representative electrophoretic image of RT-PCR products of the VEGF gene. The horizontal bar at the bottom indicates light and dark cycles.

Close modal
Fig. 2.

Transcriptional regulation of the mouse VEGF gene by clock genes. A, location of E-boxes and HRE within the 5′-flanking region of the mouse VEGF gene. The numbers represent distances, in bases, from the putative transcription start site, marked as +1. B, transcriptional regulation of Per1-Luc (left) and VEGF-Luc (right) reporter activity by clock genes. Presence (+) or absence (−) of expression plasmids (0.25 μg each) is denoted. Each value represents the mean ± SE (bars) of three independent experiments. C, influence of CLOCK and BMAL1 on hypoxia-induced VEGF transcription. Presence (+) or absence (−) of expression plasmids (0.25 μg each) is denoted. Each value represents the mean ± SE (bars) of three independent experiments. D, dose-dependent inhibition of hypoxia-induced VEGF promoter activity by PER2 and CRY1. The amounts of PER2 or CRY1 expression constructs transfected are listed (in μg) at the extremes of the triangles. Each value represents the mean ± SE (bars) of three independent experiments.

Fig. 2.

Transcriptional regulation of the mouse VEGF gene by clock genes. A, location of E-boxes and HRE within the 5′-flanking region of the mouse VEGF gene. The numbers represent distances, in bases, from the putative transcription start site, marked as +1. B, transcriptional regulation of Per1-Luc (left) and VEGF-Luc (right) reporter activity by clock genes. Presence (+) or absence (−) of expression plasmids (0.25 μg each) is denoted. Each value represents the mean ± SE (bars) of three independent experiments. C, influence of CLOCK and BMAL1 on hypoxia-induced VEGF transcription. Presence (+) or absence (−) of expression plasmids (0.25 μg each) is denoted. Each value represents the mean ± SE (bars) of three independent experiments. D, dose-dependent inhibition of hypoxia-induced VEGF promoter activity by PER2 and CRY1. The amounts of PER2 or CRY1 expression constructs transfected are listed (in μg) at the extremes of the triangles. Each value represents the mean ± SE (bars) of three independent experiments.

Close modal
Fig. 3.

Temporal profiles of HIF-1α, PER2, and CRY1 protein abundance in tumor masses. A, comparison of HIF-1α abundance in tumor masses with those in other healthy tissues. B, temporal profile of HIF-1α abundance in tumor masses and liver. C, immunoprecipitation analysis of the interaction of HIF-1α with PER2 or CRY1. The left panel shows the temporal profile of PER2 and CRY1 abundance in sarcoma 180 tumor nuclei. The right panel shows the co-immunoprecipitation of HIF-1α with PER2 or CRY1. Nuclear extracts were immunoprecipitated (IP) with antibodies against HIF-1α and separated by SDS-PAGE. The blot was incubated with antibodies against PER2 or CRY1.

Fig. 3.

Temporal profiles of HIF-1α, PER2, and CRY1 protein abundance in tumor masses. A, comparison of HIF-1α abundance in tumor masses with those in other healthy tissues. B, temporal profile of HIF-1α abundance in tumor masses and liver. C, immunoprecipitation analysis of the interaction of HIF-1α with PER2 or CRY1. The left panel shows the temporal profile of PER2 and CRY1 abundance in sarcoma 180 tumor nuclei. The right panel shows the co-immunoprecipitation of HIF-1α with PER2 or CRY1. Nuclear extracts were immunoprecipitated (IP) with antibodies against HIF-1α and separated by SDS-PAGE. The blot was incubated with antibodies against PER2 or CRY1.

Close modal
Fig. 4.

Repression of the HIF-1α/ARNT-induced VEGF promoter activity by PER2 and CRY1. A, transcriptional activation of the VEGF gene by HIF-1α and ARNT. Presence (+) or absence (−) of expression plasmids (0.08 μg of HIF-1α; 0.12 μg of ARNT) is denoted. Each value represents the mean ± SE (bars) of three independent experiments B, dose-dependent repression of HIF-1α/ARNT-induced VEGF promoter activity by PER2 and CRY1. The amounts of the PER2 or CRY1 expression constructs transfected are listed (in μg) at the extremes of the triangles. Each value represents the mean ± SE (bars) of three independent experiments.

Fig. 4.

Repression of the HIF-1α/ARNT-induced VEGF promoter activity by PER2 and CRY1. A, transcriptional activation of the VEGF gene by HIF-1α and ARNT. Presence (+) or absence (−) of expression plasmids (0.08 μg of HIF-1α; 0.12 μg of ARNT) is denoted. Each value represents the mean ± SE (bars) of three independent experiments B, dose-dependent repression of HIF-1α/ARNT-induced VEGF promoter activity by PER2 and CRY1. The amounts of the PER2 or CRY1 expression constructs transfected are listed (in μg) at the extremes of the triangles. Each value represents the mean ± SE (bars) of three independent experiments.

Close modal
Fig. 5.

Circadian fluctuation in VEGF protein levels in tumor-bearing mice. A, temporal profile of VEGF content in tumor masses and liver. Each point represents the mean ± SE (bars) of four to six mice. There was a significant circadian variation in VEGF content in tumor masses (P < 0.01, ANOVA). B, time dependence of plasma concentration of VEGF in tumor-bearing mice. Each point represents the mean ± SE (bars) of four to six mice. There was a significant circadian variation in plasma concentrations of VEGF in tumor-bearing mice (P < 0.01, ANOVA). C, influence of the gaining of tumor volume on the circadian variation in VEGF content in the implanted tumor cells (sarcoma 180, Lewis lung carcinoma, and B16 melanoma). Each value represents the mean ± SE (bars) of three or four mice. ∗∗, P < 0.01; ∗, P < 0.05 for comparison between the two groups using Tukey’s test.

Fig. 5.

Circadian fluctuation in VEGF protein levels in tumor-bearing mice. A, temporal profile of VEGF content in tumor masses and liver. Each point represents the mean ± SE (bars) of four to six mice. There was a significant circadian variation in VEGF content in tumor masses (P < 0.01, ANOVA). B, time dependence of plasma concentration of VEGF in tumor-bearing mice. Each point represents the mean ± SE (bars) of four to six mice. There was a significant circadian variation in plasma concentrations of VEGF in tumor-bearing mice (P < 0.01, ANOVA). C, influence of the gaining of tumor volume on the circadian variation in VEGF content in the implanted tumor cells (sarcoma 180, Lewis lung carcinoma, and B16 melanoma). Each value represents the mean ± SE (bars) of three or four mice. ∗∗, P < 0.01; ∗, P < 0.05 for comparison between the two groups using Tukey’s test.

Close modal
Fig. 6.

Influence of dosing time on the antitumor effect of angiogenesis inhibitors. Sarcoma 180 tumor cells were injected into ICR male mice. SU1498 (25 μg/mouse), TNP-470 (30 mg/kg), BB2519 (10 mg/kg; ▴, ZT 2; ▪, ZT 14), or vehicle (○) was administered s.c. once every other day. Each point is the mean ± SE (bars) of 8–10 mice. ∗∗, P < 0.01; ∗, P < 0.05 for comparison between the two dosing times using Tukey’s test.

Fig. 6.

Influence of dosing time on the antitumor effect of angiogenesis inhibitors. Sarcoma 180 tumor cells were injected into ICR male mice. SU1498 (25 μg/mouse), TNP-470 (30 mg/kg), BB2519 (10 mg/kg; ▴, ZT 2; ▪, ZT 14), or vehicle (○) was administered s.c. once every other day. Each point is the mean ± SE (bars) of 8–10 mice. ∗∗, P < 0.01; ∗, P < 0.05 for comparison between the two dosing times using Tukey’s test.

Close modal

We are indebted to Takeda Chemical Industrial, Ltd. (Osaka, Japan) for providing the TNP-470 used in this study.

1
Folkman J., Shing Y. Angiogenesis.
J. Biol. Chem.
,
267
:
10931
-10934,  
1992
.
2
Klagsbrun M., D’Amore P. A. Regulators of angiogenesis.
Annu. Rev. Physiol.
,
53
:
217
-239,  
1991
.
3
Folkman J. Tumor angiogenesis: therapeutic implications.
N. Engl. J. Med.
,
285
:
1182
-1186,  
1971
.
4
Dvorak H. F., Sioussat T. M., Brown L. F., Berse B., Nagy J. A., Sotrel A., Manseau E. J., Van de Water L., Senger D. R. Distribution of vascular permeability factor (vascular endothelial growth factor) in tumors: concentration in tumor blood vessels.
J. Exp. Med.
,
174
:
1275
-1278,  
1991
.
5
Asano M., Yukita A., Matsumoto T., Kondo S., Suzuki H. Inhibition of tumor growth and metastasis by an immunoneutralizing monoclonal antibody to human vascular endothelial growth factor/vascular permeability factor 121.
Cancer Res.
,
55
:
5296
-5301,  
1995
.
6
de Vries C., Escobedo J. A., Ueno H., Houck K., Ferrara N., Williams L. T. The fms-like tyrosine kinase, a receptor for vascular endothelial growth factor.
Science (Wash. DC)
,
255
:
989
-991,  
1992
.
7
Terman B. I., Dougher-Vermazen M., Carrion M. E., Dimitrov D., Armellino D. C., Gospodarowicz D., Bohlen P. Identification of the KDR tyrosine kinase as a receptor for vascular endothelial cell growth factor.
Biochem. Biophys. Res. Commun.
,
187
:
1579
-1586,  
1992
.
8
Levy A. P., Levy N. S., Wegner S., Goldberg M. A. Transcriptional regulation of the rat vascular endothelial growth factor gene by hypoxia.
J. Biol. Chem.
,
270
:
13333
-13340,  
1995
.
9
O’Reilly M. S. The combination of angiogenic therapy with other modalities.
Cancer J.
,
volume 8(Suppl. 1)
:
S89
-S99,  
2002
.
10
Koyanagi S., Ohdo S., Yukawa E., Higuchi S. Chronopharmacological study of interferon-α in mice.
J. Pharmacol. Exp. Ther.
,
283
:
259
-264,  
1997
.
11
Koyanagi S., Nakagawa H., Kuramoto Y., Ohdo S., Soeda S., Shimeno H. Optimizing the dosing schedule of TNP-470 [O-(chloroacetyl-carbamoyl)fumagillol] enhances its antitumor and antiangiogenic efficacies.
J. Pharmacol. Exp. Ther.
,
304
:
669
-674,  
2003
.
12
Ohdo S., Koyanagi S., Suyama H., Higuchi S., Aramaki H. Changing the dosing schedule minimizes the disruptive effects of interferon on clock function.
Nat. Med.
,
7
:
356
-360,  
2001
.
13
Moore R. Y., Eichler V. B. Loss of a circadian adrenal corticosterone rhythm following suprachiasmatic lesions in the rat.
Brain Res.
,
42
:
201
-206,  
1972
.
14
Gekakis N., Staknis D., Nguyen H. B., Davis F. C., Wilsbacher L. D., King D. P., Takahashi J. S., Weitz C. J. Role of the CLOCK protein in the mammalian circadian mechanism.
Science (Wash. DC)
,
280
:
1564
-1569,  
1998
.
15
Kume K., Zylka M. J., Sriram S., Shearman L. P., Weaver D. R., Jin X., Maywood E. S., Hastings M. H., Reppert S. M. mCRY1 and mCRY2 are essential components of the negative limb of the circadian clock feedback loop.
Cell
,
98
:
193
-205,  
1999
.
16
Strawn L. M., McMahon G., App H., Schreck R., Kuchler W. R., Longhi M. P., Hui T. H., Tang C., Levitzki A., Gazit A., Chen I., Keri G., Orfi L., Risau W., Flamme I., Ullrich A., Hirth K. P., Shawver L. K. Flk-1 as a target for tumor growth inhibition.
Cancer Res.
,
56
:
3540
-3545,  
1996
.
17
Ingber D., Fujita T., Kishimoto S., Sudo K., Kanamaru T., Brem H., Folkman J. Synthetic analogues of fumagillin that inhibit angiogenesis and suppress tumor growth.
Nature (Lond.)
,
348
:
555
-557,  
1990
.
18
Watson S. A., Morris T. M., Collins H. M., Bawden L. J., Hawkins K., Bone E. A. Inhibition of tumor growth by marimastat in a human xenograft model of gastric cancer: relationship with levels of circulating CEA.
Br. J. Cancer
,
81
:
19
-23,  
1999
.
19
Pando M. P., Morse D., Cermakian N., Sassone-Corsi P. Phenotypic rescue of a peripheral clock genetic defect via SCN hierarchical dominance.
Cell
,
110
:
107
-117,  
2002
.
20
Lee C., Etchegaray J. P., Cagampang F. R., Loudon A. S., Reppert S. M. Posttranslational mechanisms regulate the mammalian circadian clock.
Cell
,
107
:
855
-867,  
2001
.
21
Preitner N., Damiola F., Molina L. L., Zakany J., Duboule D., Albrecht U., Schibler U. The orphan nuclear receptor REV-ERBα controls circadian transcription within the positive limb of the mammalian circadian oscillator.
Cell
,
110
:
251
-260,  
2002
.
22
Zylka M. J., Shearman L. P., Weaver D. R., Reppert S. M. Three period homologs in mammals: differential light responses in the suprachiasmatic circadian clock and oscillating transcripts outside of brain.
Neuron
,
20
:
1103
-1110,  
1998
.
23
Ferrara N., Davis-Smith T. The biology of vascular endothelial growth factor.
Endocr. Rev.
,
18
:
4
-25,  
1997
.
24
Hogenesch J. B., Gu Y. Z., Jain S., Bradfield C. A. The basic-helix-loop-helix-PAS orphan MOP3 forms transcriptionally active complexes with circadian and hypoxia factors.
Proc. Natl. Acad. Sci. USA
,
95
:
5474
-5479,  
1998
.
25
Hogenesch J. B., Gu Y. Z., Moran S. M., Shimomura K., Radcliffe L. A., Takahashi J. S., Bradfield C. A. The basic helix-loop-helix-PAS protein MOP9 is a brain-specific heterodimeric partner of circadian and hypoxia factors.
J. Neurosci.
,
20
:
RC83
2000
.
26
Kimura H., Weisz A., Ogura T., Hitomi Y., Kurashima Y., Hashimoto K., D’Acquisto F., Makuuchi M., Esumi H. Identification of hypoxia-inducible factor 1 ancillary sequence and its function in vascular endothelial growth factor gene induction by hypoxia and nitric oxide.
J. Biol. Chem.
,
276
:
2292
-2298,  
2000
.
27
Leung D. W., Cachianes G., Kuang W. J., Goeddel D. V., Ferrara N. Vascular endothelial growth factor is a secreted angiogenic mitogen.
Science (Wash. DC)
,
246
:
1306
-1309,  
1989
.
28
Wang H., Keiser J. A. Vascular endothelial growth factor upregulates the expression of matrix metalloproteinases in vascular smooth muscle cells: role of flt-1.
Circ. Res.
,
83
:
832
-840,  
1998
.
29
Wang G. L., Jiang B. H., Rue E. A., Semenza G. L. Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension.
Proc. Natl. Acad. Sci. USA
,
92
:
5510
-5514,  
1995
.
30
Reyes H., Reisz-Porszasz S., Hankinson O. Identification of the Ah receptor nuclear translocator protein (Arnt) as a component of the DNA binding form of the Ah receptor.
Science (Wash. DC)
,
256
:
1193
-1195,  
1992
.
31
Zheng B., Larkin D. W., Albrecht U., Sun Z. S., Sage M., Eichele G., Lee C. C., Bradley A. The mPer2 gene encodes a functional component of the mammalian circadian clock.
Nature (Lond.)
,
400
:
169
-173,  
1999
.
32
Chilov D., Hofer T., Bauer C., Wenger R. H., Gassmann M. Hypoxia affects expression of circadian genes PER1 and CLOCK in mouse brain.
FASEB J.
,
15
:
2613
-2622,  
2001
.
33
Shearman L. P., Zylka M. J., Weaver D. R., Kolakowski L. F., Jr., Reppert S. M. Two period homologs: circadian expression and photic regulation in the suprachiasmatic nuclei.
Neuron
,
19
:
1261
-1269,  
1997
.
34
Etchegaray J. P., Lee C., Wade P. A., Reppert S. M. Rhythmic histone acetylation underlies transcription in the mammalian circadian clock.
Nature (Lond.)
,
421
:
177
-182,  
2003
.
35
Kallio J. P., Okamoto K., O’Brien S., Carrero P., Makino Y., Tanaka H., Poellinger L. Signal transduction of hypoxic cells: inducible nuclear translocation and recruitment of the CBP/p300 coactivator by the hypoxia-inducible factor-1α.
EMBO J.
,
17
:
6573
-6586,  
1998
.
36
Pluda J. M. Tumor-associated angiogenesis: mechanisms, clinical implications, and therapeutic strategies.
Semin. Oncol.
,
24
:
203
-218,  
1997
.
37
Endo I., Mitsui T., Nishino M., Oshima Y., Matsumoto T. Diurnal fluctuation of edema synchronized with plasma VEGF concentration in a patient with POEMS syndrome.
Intern. Med.
,
41
:
1196
-1198,  
2002
.