Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

A large number of rhythmic variables are influenced by environmental factors such as light, temperature, and social communication that vary cyclically in nature and serve to synchronize biological rhythms to the daily rotation of the earth (Aschoff, 1963). Responses to a variety of drugs show diurnal rhythmicity (Ohdo et al., 1988, 1991, 1996, 1997, 1998; Watanabe et al., 1992). Use of a chronopharmacological strategy can improve tumor response to treatment and overall survival rates and reduce drug toxicities in humans (Hrushesky, 1985; Levi et al., 1997). The mechanisms involved in the diurnal rhythm of drug susceptibility have been examined from the viewpoints of the sensitivity of living organisms to drugs and/or the pharmacokinetics of drugs.

Interferons (IFNs) are multifunctional cytokines that have not only antiproliferative and immunological effects but also potent antiviral effects (Baron et al., 1991). IFNs have been widely used to treat patients with various cancers and hepatitis. However, adverse effects such as fever, headache, leukopenia, and thrombocytopenia are frequently observed in patients treated with IFNs (Baron et al., 1991). One approach to increasing the efficiency of treatment with IFNs is to administer the drugs at a time when they are most effective and/or tolerated. Certainly, the fever (Koyanagi et al., 1997; Ohdo et al., 1997) or leukopenia (Koren and Fleischmann, 1993) induced by IFN- is significantly affected by dosing time. Also, the antitumor activity of IFN- and - varies depending on dosing time in a mouse model (Koren et al., 1993). Such dosing time-dependent differences could occur at many levels, including dosing time-dependent variation in pharmacokinetics, tumor responsiveness, and host immune responsiveness. However, the exact mechanisms have not been clarified yet.

IFNs inhibit cell growth through the up-regulation of the cyclin-dependent kinase (cdk) inhibitor p21 wild-type p53-activated fragment 1 (p21WAF1) (Sangfelt et al., 1997; Mandal et al., 1998). IFNs elicit the transcription of various genes through activation of signal transducers and activators of transcription (STAT) protein, via binding to specific receptors (Darnell et al., 1994). Furthermore, cell cycle dependency has been demonstrated for the specific binding of IFN- to surface receptor (Tamura et al., 1997) and the transcription induced by IFN- (Kumar et al., 1994). However, the relationship between the diurnal rhythm of cell cycle distribution and the antitumor effect of IFNs has not been investigated yet. Although a significant dosing time-dependent pharmacokinetics has been demonstrated for IFNs concentration in plasma, it has not been systematically investigated in tissue. This is because it is often difficult to obtain the time course of drug concentrations in tissue from individual subjects. NONMEM (nonlinear mixed effect model) is a computer program designed to analyze pharmacokinetics in study populations by pooling data (Beal and Sheiner, 1992). In this study, NONMEM was applied to the pharmacokinetic analysis of IFN- concentrations in tumor mass.

The purpose of this study was to investigate the influence of dosing time on tumor growth after the intratumoral administration of IFN- in tumor-bearing mice. The mechanism underlying the dosing time-dependent difference was elucidated based on IFN- pharmacodynamics or pharmacokinetics.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals and Cells. Male C57BL/6 mice (5 weeks old) were purchased from the Laboratory Animal Center, Faculty of Medicine, Kyushu University (Fukuoka, Japan). They were housed 8 to 10 per cage under standardized light-dark cycle conditions (lights on at 7:00 AM, off at 7:00 PM) at 24 ± 1°C and 60 ± 10% humidity with food and water available ad libitum. Their activity increases during the dark phase. Murine B16 melanoma cells (clone F1) (Dainippon Pharmaceutical Co. Ltd., Osaka, Japan) were maintained in vitro in Dulbecco's modified Eagle's medium supplemented with 10% heated-inactivated fetal bovine serum, 0.5% kanamycin, 0.5% penicillin, and 0.5% streptomycin at 37°C in a humidified atmosphere with 5% CO2. Mice on day 7 after tumor cell implantation were used as tumor-bearing hosts. A 50-µl volume of 1.5 × 10⁶ viable tumor cells was inoculated into the left hind footpads 7 days before drug treatment.

Experimental Design. To determine the dose-response of the antitumor effect of natural human IFN- (Feron; Toray Industries Inc., Tokyo, Japan), groups of six to seven tumor-bearing mice were injected intratumorally on days 0 to 6 (for 7 days) with 0.005, 0.05, 0.5, and 5 MI.U./kg of IFN- or saline at 9:00 AM. The mice were monitored for day of death. Tumor volume was measured on day 12. The lyophilized powder of IFN- was dissolved in saline containing 0.1% BSA to yield an appropriate concentration of 0.003, 0.033, 0.333, and 3.33 MI.U./ml. The volume of injection was 15 µl/10 g. To investigate the influence of dosing time on antitumor effect, groups of 6 to 12 tumor-bearing mice were injected intratumorally on days 0 to 6 with IFN- (0.5 MI.U./kg) or saline at 9:00 AM or 9:00 PM. The mice were monitored for day of death. Tumor volume was measured every 2 days. To examine the diurnal rhythm of the cell cycle, tumor masses were removed from groups of five tumor-bearing mice at 9:00 AM, 1:00 PM, 5:00 PM, 9:00 PM, 1:00 AM, or 5:00 AM. To investigate the specific binding of IFN- to receptor on tumor cells, tumor masses were removed from groups of 8 to 10 tumor-bearing mice at 9:00 AM or 9:00 PM. To study the influence of IFN- dosing time on STAT1 protein or p21WAF1 protein expression in implanted tumor cells, groups of five to eight tumor-bearing mice were given an intratumoral injection of IFN- (0.5 MI.U./kg) or saline at 9:00 AM or 9:00 PM. Their tumor masses were removed at 4 h (for STAT1 protein) or 12 h (for p21WAF1 protein) after IFN- or saline injection. To study the influence of dosing time on IFN- concentrations in tumor, groups of 44 to 51 tumor-bearing mice were given an intratumoral injection of IFN- (0.5 MI.U./kg) at 9:00 AM or 9:00 PM. Tumor mass was removed at 0.05, 0.25, 0.5, 1, 2, 3, or 4 h after IFN- injection.

Determination of Antitumor Effect. Tumor volume was estimated using the formula: tumor volume (mg) = 4xyz/3, where 2x, 2y, and 2z are the three perpendicular diameters of tumor. Relative tumor growth rate was expressed as the tumor volume change from the initiation of IFN- or saline treatment. Survival time was estimated as the period from the initiation of treatment to death. Survival rate was calculated as the percentage change for each group of 6 to 12 mice.

Cell Cycle Analysis. To obtain the tumor cell suspension, the tumor mass was minced with scissors in ice-cold PBS and the mixture was filtered to remove unnecessary tissue blocks. The cell suspension was centrifuged at 800 rpm for 10 min at 4°C (model RL-101; Tomy Seiko Co. Ltd., Tokyo, Japan). The pellets were washed twice in 10 ml of ice-cold PBS and resuspended in 2 ml of the same buffer. The cells were then fixed dropwise in ice-cold 96% ethanol and stored at 4°C overnight. Ethanol-fixed tumor cells were washed twice in 10 ml of ice-cold PBS. Thereafter, 1 ml of ribonuclease A (1 mg/ml of PBS) per 1 × 10⁶ cells was added, and the mixture was incubated for 60 min at 37°C. For specific staining of DNA, 1 ml of propidium iodide (0.05 mg/ml in 0.1% sodium citrate solution) per 2 × 10⁶ cells was added, and the cells were analyzed on the EPICS Elite flow cytometer (Coulter Co., Hialeah, FL) (488 nm). The total number of cells analyzed from each sample was 10,000.

Specific IFN--Binding Assay. The iodination of IFN- reduces its biological potency by <30% (Kushnaryov et al., 1985). Both IFN- and - cross-compete for the same receptor (Branca and Baglioni, 1981). Therefore, recombinant human IFN- (Pepro Tech EC Ltd., London, England) was used as ligand to specific receptor of IFN-. IFN- was iodinated using a solid-phase lactoperoxidase kit (ICN Pharmaceuticals, Inc., Irvine, CA). The tumor cell suspension was prepared as described above and resuspended in ice-cold culture medium containing 0.25% BSA, 0.1% sodium azide, 10 µg/ml protamine sulfate, and 2.5 mM CaCl₂. The binding assay was performed at 4°C for 2 h with a reaction mixture (total volume, 200 µl) containing 1 ng/ml ¹²⁵I-IFN- and 3 × 10⁵ viable cells. After the incubation, 200 µl of heated-inactivated fetal bovine serum was added, and the mixture was centrifuged at 10,000 rpm for 1 min. The supernatant was removed. Thereafter, the tube tip containing bound ligand was amputated, and the radioactivity was measured using a gammer counter (ARC-360; Aloka Co., Mitaka, Tokyo, Japan). Nonspecific binding was evaluated in the presence of a 1500-fold excess of unlabeled IFN-. Specific binding was calculated by subtracting nonspecific binding from total binding as follows: specific binding (%) = [(total binding  nonspecific binding)/total binding] × 100.

Western Blot Analysis. The removed tumor mass was placed into polypropylene tubes containing ice-cold hemolysis buffer (Tris-buffered ammonium chloride) to remove erythrocytes. The isolated tumor mass was minced with scissors and centrifuged at 12,000g for 5 min. The pellet was washed with ice-cold PBS and resuspended in ice-cold lysis buffer (120 mM NaCl, 100 mM NaF, 200 µM Na₂V₂O₅, 1 mM PMSF, 0.5% Nonidet P-40, 0.001% leupeptin, 50 mM Tris-HCl, pH 7.4). The pellet was homogenized with ice-cold lysis buffer and centrifuged at 12,000g for 5 min. The tumor mass lysate containing 20 µg (for STAT1 protein) or 40 µg (for p21WAF1 protein) of total protein was mixed with an equal volume of 2× sample buffer (0.125 M Tris-HCl, pH 6.8, 10% 2-mercaptoethanol, 4% SDS, 10% sucrose, 0.004% bromophenol blue) and boiled at 95°C for 5 min. The protein concentrations in tumor mass lysates were determined by Lowry's method (DC Protein Assay; Bio-Rad, Hercules, CA). The lysate sample was resolved by 8% (for STAT1 protein) or 12% (for p21WAF1) SDS polyacrylamide gel electrophoresis, transferred onto nitrocellulose membrane (Clear Blot Membrane-p; Atto Co., Tokyo, Japan), and immunoblotted with the anti-STAT1 (STAT1  or STAT1 ) monoclonal antibody (Transduction Lab., Lexington, KY) or the anti-p21WAF1 monoclonal antibody (Oncogene Research Products, Cambridge, MA). Thereafter, the membrane was incubated with horseradish peroxidase-conjugated secondary antibody (mouse IgG). The blot was visualized with hydrogen peroxide in 3,3',5,5'-tetramethylbenzidine and scanned with Deskscan II (Hewlett Packard Japan Co. Ltd., Tokyo, Japan). The band intensity was quantified by using NIH Image. Plots of band intensity set the mean value of control at 9:00 AM at 1.

Determination of IFN- Concentrations in Tumor. The removed tumor mass was placed into ice-cold PBS containing 0.1% BSA, and 10 µg/ml leupeptin. The tumor mass in the buffer was homogenized and centrifuged at 12,000g for 5 min (Kubota Hematocrit KH-120A; Kubota, Tokyo, Japan). The supernatant was isolated and stored at 20°C until assayed. The IFN- concentrations in tumor were determined by an enzyme-linked-immunosorbent assay method (Human IFN- ELISA kit; Toray Industries Inc., Tokyo, Japan). The coefficient of variation is less than 4%, and assay range is between 3 and 200 I.U./ml. The recovery of IFN- from tumor mass is more than 90%. The protein concentrations in homogenate sample were determined by Lowry's method as described above.

NONMEM Analysis. The population pharmacokinetic parameters were calculated on an HP-9000 series 700 (Yokogawa-Hewlett Packard Ltd., Tokyo, Japan) with the NONMEM program (version IV, level 1.1) following the two-compartment model (the PREDPP program, subroutines ADVAN3 and TRANS1). Bayesian estimates of individual pharmacokinetic parameters were obtained with the post hoc method of the NONMEM program. The statistical moment parameters such as area under the curve (AUC) and mean residence time (MRT) were calculated by using the estimated individual pharmacokinetic parameters.

Statistical Analysis. The percentage of cells in each cell cycle phase (G₁, S, G₂/M) was calculated according to Multicycle, a cell cycle analytical software package (Coulter Co., Hialeah, FL). The values were validated for each phase among six different sampling times by ANOVA. ANOVA and Tukey's test were applied for the multiple comparison. Student's t test was used for two independent groups. Survival curves were compared with the log-rank test. The 5% level of probability was considered to be significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Dose-Response Effects of IFN- on Tumor Growth or Survival. The effects of various dosages (0.005, 0.05, 0.5, and 5 MIU/kg, intratumorally) of IFN- on tumor growth or survival time in tumor-bearing mice injected with the drug at the same time (9:00 AM) are shown in Table 1. Tumor growth on day 12 after initiation of IFN- (0.5 or 5 MI.U./kg) treatment was significantly suppressed compared with that in the control group given saline (P < .05). Also, the survival time after initiation of IFN- (0.5 or 5 MI.U./kg) treatment was significantly prolonged compared with that in the control group (P < .05). However, the tumor growth and survival time did not differ significantly between tumor-bearing mice injected with IFN- (0.005 or 0.05 MI.U./kg) and with saline.

View this table: [in this window] [in a new window]   TABLE 1 Dose-response effects of IFN- on tumor growth or survival time Drug injection is performed at 9:00 AM. Each value is the mean with S.E. of six to seven mice.

Influence of Dosing Time on Tumor Growth or Survival. All tumor-bearing mice injected with saline at 9:00 AM or 9:00 PM died between day 14 and day 22. No significant effect of dosing time was observed for survival after saline injection (data not shown). Also, no dosing time-dependent change in the rate of tumor growth was observed on day 12 after initiation of saline injection (data not shown). Therefore, a mean value between 9:00 AM and 9:00 PM is shown as the control in Fig. 1. The tumor growth in tumor-bearing mice on day 9 or day 12 after initiation of IFN- (0.5 MI.U./kg) treatment at 9:00 AM or 9:00 PM was significantly suppressed when compared with that in control mice given saline (P < .05, respectively; Fig. 1). The tumor growth in tumor-bearing mice on day 12 after initiation of IFN- injection at 9:00 AM was significantly reduced relative to that in mice injected with IFN- at 9:00 PM (P < .05). Also, the survival time after IFN- injection at 9:00 AM was significantly longer than that after saline injection (P < .05; Fig. 1). However, the survival time after IFN- injection at 9:00 PM was not significantly different from that after saline injection or IFN- injection at 9:00 AM.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Influence of dosing time on tumor growth or survival rate (%) after IFN- (0.5 MI.U./kg, intratumorally (, 9:00 AM; , 9:00 PM) or saline (, 9:00 AM or 9:00 PM) injection on days 0 to 6 (for 7 days). Each value is the mean with S.E. of 6 to 12 mice. *P < .05 when compared with the corresponding saline group, §P < .05 when compared between the two dosing times using Tukey's test. Survival rate after IFN- treatment at 9:00 AM was significantly greater than that after saline treatment (P < .05 using log-rank test).

Diurnal Rhythm of the Cell Cycle in Tumor Cells. A significant diurnal rhythm dependence was demonstrated for G₁, S, and G₂/M phases in tumor cells from tumor-bearing mice (ANOVA, G₁ and S phase, P < .01; G₂/M phase, P < .05; Fig. 2). The proportion of tumor cells in the G₁ phase showed a peak at 9:00 PM and a trough at 9:00 AM. The proportion of tumor cells in the S phase was high at 5:00 AM and 9:00 AM and low at 5:00 PM and 9:00 PM. The proportion of tumor cells in the G₂/M phase showed a peak at 1:00 PM and a trough at 9:00 PM. These results were interrelated in that the tumor cells in the G₁ phase enter the S phase and later the G₂/M phase.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Diurnal rhythm of the cell cycle in tumor cells prepared at six different times. Each cell cycle phase represents G1 (), S (), G2/M (). Each value is the mean with S.E. of five mice. A significant diurnal rhythm dependence was demonstrated for G1, S, and G2/M phase (ANOVA, G1, S, P < .01; G2/M, P < .05).

Time-Dependent Change of Specific IFN--Binding in Isolated Tumor Cells. The specific binding of IFN- to receptor was significantly greater in tumor cells prepared at 9:00 AM than in tumor cells prepared at 9:00 PM (P < .05, Fig. 3).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Time-dependent change in specific IFN binding (percentage of total binding) to tumor cells prepared at 9:00 AM or 9:00 PM. Each value is the mean with S.E. of 8 to 10 mice. *P < .05 when compared between the two times using Student's t test.

Influence of IFN- Dosing Time on STAT1 Protein Level in Tumor Masses. As shown in Fig. 4, the STAT1 or STAT1 protein level at 4 h after a single injection of IFN- (0.5 MI.U./kg) at 9:00 AM was significantly higher when compared with that after saline injection at 9:00 AM (STAT1, P < .01; STAT1, P < .05). However, the protein level at 4 h after IFN- injection at 9:00 PM was not significantly different from that after saline injection at 9:00 PM.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Influence of dosing time on STAT1 (STAT1 or STAT1) protein expression (A) or relative STAT1 protein expression (B) in tumor masses at 4 h after IFN- (0.5 MI.U./kg, intratumorally.) () or saline () injection at 9:00 AM or 9:00 PM. Plots of band intensity set the mean value of control at 9:00 AM at 1. Each value is the mean with S.E. of five to six mice. *P < .05, **P < .01 when compared with the corresponding saline group using Tukey's test.

Influence of IFN- Dosing Time on p21WAF1 Protein Level in Tumor Masses. As shown in Fig. 5, the p21WAF1 protein level at 12 h after a single injection of IFN- (0.5 MI.U./kg) at 9:00 AM or 9:00 PM was significantly higher when compared with that after saline injection at the corresponding dosing time (P < .01, respectively). Furthermore, it was significantly higher in tumor-bearing mice injected with IFN- at 9:00 AM than at 9:00 PM (P < .05).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Influence of dosing time on p21WAF1 protein expression (A) or relative p21WAF1 protein expression (B) in tumor masses at 12 h after IFN- (0.5 MI.U./kg, intratumorally.) () or saline () injection at 9:00 AM or 9:00 PM. Plots of band intensity set the mean value of control at 9:00 AM at 1. Each value is the mean with S.E. of seven to eight mice. *P < .05, **P < .01 when compared with the corresponding saline group or between the two dosing times using Tukey's test.

Influence of Dosing Time on IFN- Pharmacokinetics. The time course of the change in IFN- concentration in tumor after a single injection of IFN- (0.5 MI.U./kg) decreased in a biexponential fashion. The concentrations in tumor at 3 or 4 h after IFN- injection at 9:00 AM were significantly higher than those after the drug injection at 9:00 PM (P < .01, Fig. 6). Table 2 shows the pharmacokinetic parameters after IFN- injection. The analysis was conducted by using 95 tumor concentrations obtained from 95 mice. The final model equations estimated for all data were as follows: CL (mg of protein/h) = 24.9 × 1.22^DT, V_c (mg of protein) = 12.3, k₁₂ (1/h) = 2.04, k₂₁ (1/h) = 2.57, where CL, V_c, k₁₂, and k₂₁ are total clearance, central volume of distribution, distribution rate constant from central to peripheral compartment, and distribution rate constant from peripheral to central compartment, respectively. DT represents dosing time: DT = 0 if injection was at 9:00 AM; DT = 1 if injection was at 9:00 PM. Using the population parameters, individual pharmacokinetic parameters were calculated based on Bayesian estimate, and then AUC and MRT were derived from them. CL was significantly larger in mice injected with IFN- at 9:00 PM than at 9:00 AM (P < .01). AUC, MRT, t_1/2 and t_1/2 were significantly larger in mice injected with IFN- at 9:00 AM than at 9:00 PM (P < .01, respectively).

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Influence of dosing time on IFN- concentrations in tumor after IFN- (0.5 MI.U./kg, intratumorally) injection at 9:00 AM () or 9:00 PM (). Each value is the mean with S.E. of 5 to 11 mice. **P < .01 when compared between the two dosing times using Student's t test.

View this table: [in this window] [in a new window]   TABLE 2 Influence of dosing time on pharmacokinetic parameters after IFN- (0.5 MI.U./kg, intratumorally) injection at 9:00 AM or 9:00 PM Each value is the mean with S.E. of 44 to 51 mice.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

IFNs have a capacity to inhibit the proliferation of various tumor cell lines (Ida et al., 1982; Gomi et al., 1983, 1986). And IFN- is more potent in melanoma cell lines than in other tumor cell lines (Ida et al., 1982; Gomi et al., 1983). In this study, the growth of B16 melanoma implanted in mice was inhibited by IFN- in a dose-dependent manner. Furthermore, the antitumor effect after IFN- injection was significantly more potent in tumor-bearing mice injected with the drug at 9:00 AM than at 9:00 PM. This result confirms a previous chronopharmacological finding on the antitumor effect of IFN- (Koren et al., 1993). The intratumoral administration of IFN- did not induce any toxic symptom such as fever or leukopenia. In general, the rhythmic change of drug response is caused by that of pharmacokinetic factors such as drug concentration at the site of action and/or pharmacodynamic factors such as receptor sensitivity to drug.

Whereas both IFN- and - elicit antitumor and antiviral activity by binding to the same specific receptor on the cell surface, IFN- binds to a distinct receptor (Branca and Baglioni, 1981; Zoon and Arnheiter, 1984). A decrease of receptor expression together with an increase of binding affinity is observed in lymphocytes from patients receiving IFN- therapy (Lau et al., 1991). However, the specific binding of the IFN- receptor in chronic myelogenous leukemia cells is up-regulated by synchronizing the cells mainly in early S phase by hydroxyurea treatment (Tamura et al., 1997). The increase of binding is caused by an increase in the number of binding sites with a constant receptor affinity. In this study, the proportion of tumor cells in S phase showed a significant circadian rhythm with higher levels in the late dark phase and the early light phase and lower levels in the late light phase. Moreover, the specific binding of IFN- to cellular receptors was significantly enhanced in the cells prepared at 9:00 AM compared with the cells prepared at 9:00 PM. Namely, more specific binding of IFN- was observed when the proportion of tumor cells in S phase increased and less binding was observed when it decreased. These results suggest that the time-dependent change of IFN- antitumor effect is related to that of the sensitivity, particularly at the receptor level, of tumor cells in S phase to IFN-.

IFNs mediate biological effects through activation of the Janus kinase (JAK)-STAT signaling pathway (Darnell et al., 1994). IFN-inducible STAT1 (STAT1  or STAT1 ) and STAT2 associate with a 48-kDa protein to form the transcription factor, IFN-stimulated gene factor-3 (ISGF3) complex (Qureshi et al., 1995). This complex binds to the IFN-stimulated response element (ISRE) and modulates various genes (Levy et al., 1989). Also, the STAT1 protein-activating effect of IFN- is differentially influenced by the stage of the cell cycle (Kumar et al., 1994). In this study, the STAT1 protein level in tumor cells was significantly higher after injection of IFN- than saline at 9:00 AM. This result is consistent with the time-dependent change in the specific binding of IFN-. Thus, the dosing time-dependent change in the STAT1 protein-increasing effect of IFN- may be caused by that in the specific binding of IFN- to receptor on tumor cells.

The p21WAF1 protein level in tumor cells after IFN- injection at 9:00 AM was significantly higher than that after saline injection at 9:00 AM or the drug injection at 9:00 PM. This result corresponded to the dosing time-dependent change in the STAT1 protein-enhancing effect of IFN-. The progression of the cell cycle is regulated by a number of essential proteins that stimulate or inhibit transition between the different phases of the cycle. The cdks facilitate the restriction point transition in cell cycle progression (Hunter and Pines, 1994). In particular, cdk2 and cdk4 regulate entry from the G₁ phase into the S phase. IFN- inhibits cell proliferation through the up-regulation of cdk-inhibitor p21WAF1, which inhibits the cdk2 and cdk4 activity (Sangfelt et al., 1997; Mandal et al., 1998). However, the activated STAT1 protein specifically recognizes the conserved STAT-responsive elements in the promoter of the gene encoding p21WAF1 and regulates the induction of p21WAF1 messenger RNA (Chin et al., 1996). IFN- or - does not inhibit the proliferation of tumor cells lacking STAT1 expression (Thornton et al., 1996; Sun et al., 1998). Thus, the dosing time-dependent change in the p21WAF1 seems to be caused by that in the STAT1 and influences the antitumor effect of IFN- in a dosing time-dependent manner.

IFN- concentrations in tumor were significantly higher after IFN- injection at 9:00 AM than at 9:00 PM. A significant dosing time-dependent difference was also demonstrated for the pharmacokinetic parameters of IFN-, which showed higher CL for injection at 9:00 PM than at 9:00 AM. The rhythmicity of CL seems to be closely related to that of IFN- concentrations in tumor. The drug clearance is determined by intrinsic CL or blood flow in metabolic or excretive organs. In this study, the predominant pathway of IFN- elimination is via the tumor cells because the drug was directly administered into tumor tissue. IFN- is internalized via receptor-mediated endocytosis and catabolized intracellulary by lysosomal proteinases in metabolic tissue (Bocci et al., 1983). Moreover, the receptor-mediated uptake contributes to the body CL of cytokines such as granulocyte colony-stimulating factor (Kuwabara et al., 1994) and erythropoietin (Kato et al., 1997). However, the higher CL of IFN- was observed when the specific binding of IFN- to receptor decreased. Also, the diurnal rhythm of blood flow in eliminative organs partially influences the time-dependent change of drug pharmacokinetics (Labrecque et al., 1988). The blood flow rate in the tumor tissue is significantly higher during the active phase than during the rest phase in rats (Hori et al., 1995). Therefore, the dosing time-dependent change in IFN- concentration in tumor may be partially explained by the diurnal rhythm of blood flow in tumor mass. The dosing time-dependent difference in antitumor effect of IFN- was consistent with that in MRT as well as AUC. A longer exposure to an effective concentration may be important to obtain an efficient antitumor effect of IFN- because IFN- inhibits the proliferation of various tumor cell lines in not only a concentration-dependent but a time-dependent manner in vitro (Yamada and Shimoyama, 1983; Wong et al., 1989). Thus, the dosing time-dependent difference of IFN- pharmacokinetics in tumor seems to contribute to, at least in part, that of the antitumor effect induced by IFN-.

This study suggests that the dosing time-dependent change in the antitumor activity of IFN- is caused by that in the sensitivity of tumor cells to IFN- and the pharmacokinetics of the drug. Furthermore, the time-dependent change in the sensitivity of tumor cells was related to that in the proportion of tumor cells in S phase. Therefore, the choice of dosing time based on the diurnal rhythm in the cell cycle distribution of tumor cells and the chronopharmacokinetics of IFN- may help us to establish a rational chronotherapeutic strategy, increasing the antitumor activity of the drug in certain clinical situations.