Abstract
Rats were orally co-administered sorivudine (SRV: 1-β-d-arabinofuranosyl-(E)-5-(2-bromovinyl)uracil), a new oral antiviral drug for herpes zoster, with the oral anticancer drug tegafur (FT: 1-(2-tetrahydrofuryl)-5-fluorouracil) as a prodrug of 5-flourouracil (5-FU) once daily to investigate a toxicokinetic mechanism of 15 Japanese patients’ deaths recently caused within a brief period by the drug interaction of these drugs. All the rats showed extremely elevated levels of 5-FU in plasma and tissues, including bone marrow and small intestine, and died within 10 days, whereas the animals given the same dose of SRV or FT alone were still alive over 20 days without any appreciable toxic symptom. Before their death, there was marked damage of bone marrow, marked atrophy of intestinal membrane mucosa, marked decreases in white blood cells and platelets, diarrhea with bloody flux, and severe anorexia as reported with the Japanese patients. Data obtained by in vivo andin vitro studies strongly suggested that (E)-5-(2-bromovinyl)uracil generated from SRV by gut flora was reduced in the presense of NADPH to a reactive form by hepatic dihydropyrimidine dehydrogenase (DPD), a key enzyme determining the tissue 5-FU levels, bound covalently to DPD as a suicide inhibitor, and markedly retarded the catabolism of 5-FU.
The Pharmaceutical Affairs Bureau, Japanese Ministry of Health and Welfare, reported that in 1993 fifteen Japanese patients with cancer and herpes zoster, a viral disease, died from interactions of the new antiviral drug, sorivudine (SRV1, 1-β-d-arabinofuranosyl-(E)-5-(2-bromovinyl)uracil), with oral anticancer prodrugs of 5-FU within 40 days after SRV was approved by the Japanese government and began to be used clinically (1). Before death, most of these patients had severe symptoms of toxicity, including diarrhea with bloody flux and marked decreases in white blood cell and platelet counts. The report also demonstrated that eight other Japanese patients who received both drugs during this period had severe symptoms of toxicity. All of these patients received SRV daily while being administered long-term anticancer chemotherapy with one of the 5-FU prodrugs. Most became seriously ill several days after receiving SRV.
5-FU and its oral prodrugs, including FT (1-(2-tetrahydrofuryl)-5-fluorouracil), have been recognized to cause toxic symptoms, such as diarrhea and decreases in white blood cells and platelets, in a certain percentage of patients as a result of the increased level of 5-FU in various tissues which have rapid cell proliferation, especially the intestinal membrane mucosa and bone marrow. FT, the most widely used of the 5-FU prodrugs in Japan, is activated to 5-FU mainly by hepatic cytochrome P450 after being absorbed from intestinal membrane (2). The fifteen deaths could have been avoided if the following previously demonstrated facts had been more carefully considered in the safety/risk assessment of drug interactions during the development of the new antiviral SRV: 1) BVU is generated from orally administered SRV by gut flora in rats and absorbed from the intestines (3), and 2) BVU irreversibly inhibits rat liver DPD in the presence of NADPH in vitro and markedly enhances the plasma concentration of 5-FU when 5-FU and BVU are each administered once successively ip to rats (4).
Hepatic DPD has been recognized as the most important enzyme determining plasma and tissue concentrations of 5-FU administered in the human as well as in the rat (5). 5-FU is dihydrogenated at the 5,6-double bond by DPD and rapidly hydrolyzed to α-fluoro-β-alanine by an enzymatic process (6). These findings provide support that the fifteen deaths may have been caused by elevated tissue 5-FU levels from the prodrugs as a result of the inhibition of hepatic DPD by BVU formed from the co-administered antiviral SRV. However, there has been no direct experimental evidence for the lethal drug interaction between SRV and 5-FU derivatives, nor have further toxicological studies been done to bring these deaths to light. We undertook a study using rats to determine a possible mechanism for the deaths. This communication addresses 1) the mechanism of the irreversible inhibition of purified hepatic DPD using [14C]BVU and an enzyme preparation purified from rats, 2) the extreme elevation of 5-FU concentrations in bone marrow and small intestine as well as in plasma and liver of rats orally administered FT and SRV, and 3) the significance of histological findings in these tissues in addition to hematological and toxicological findings.
Materials and Methods
Materials.
SRV was prepared as previously reported (7). [14C]BVU was prepared from 5-formyluracil with [2-14C]malonic acid (9.25 MBq, DuPont NEN Research Products, Boston, MA) in the same manner as used for the synthesis of unlabeled BVU (8). [14C]BVU synthesized had a specific activity of 2.0 MBq/μmol and a radiochemical purity higher than 99% after purification by HPLC.
Animal Treatment and Quantification of Pyrimidine Levels.
Female Wistar rats (75 animals for each group, 6 weeks of age), obtained from Japan SLC, Inc. (Hamamatsu, Japan), were orally administered FT alone (60 mg/kg, triple the daily dose in clinical use) (Taiho Pharmaceutical Co., Tokushima, Japan), SRV alone (30 mg/kg, 12 times the daily dose in clinical use), BVU alone (3.7 mg/kg) (Sigma Chemical Co., St. Louis, MO), or FT (60 mg/kg) and SRV (30 mg/kg) simultaneously once daily for 6 days. The doses of FT and SRV co-administered to the animals were determined through several preliminary experiments by adjusting the days of appearances of severe toxicity and deaths to those reported with the patients: about 3–4 days and by 10 days for the toxicity and deaths after co-administration, respectively. At these doses, the 6-day time co-administration had to be selected as one third of the animals died on days 6–7 during the treatment, and an increasing number of deaths occurred after these days on prolonging the duration of the treatment. The drugs administered were suspended in 10 ml of 0.5% sodium carboxymethylcellulose as a vehicle that was also used for the control animals. Blood and tissue samples were collected at 1, 2, 4, 8, and 24 hr after administration of the drugs on days 1, 2, 4, and 6. Samples were collected from three animals at each time point after they were sacrificed. FT and 5-FU in the plasma and tissue samples were determined as previously reported (9). SRV and BVU were analyzed by HPLC on an Inertsil ODS-2 column (150 × 4.6 mm, GL Sciences, Tokyo, Japan) eluted with 15% acetonitrile-water containing 0.01% trifluoroacetic acid (1 ml/min) after extraction with ethyl acetate from plasma (1 ml) and liver homogenates (0.3 g tissue/1 ml saline). The amounts of SRV and BVU eluted at retention times of 7.4 and 9.0 min, respectively, from the column were determined by an absolute calibration method with a detection limit of 0.1 μg/ml plasma or g liver.
CFU-GM Assay.
Bone marrow cells were collected at the 24th hr on days 1, 2, 4, and 6 from femurs of rats after repeated oral administration of the vehicle, FT alone (60 mg/kg), SRV alone (30 mg/kg) or FT (60 mg/kg) and SRV (30 mg/kg), seeded at 105 mononuclear cells/35-mm dish, and cultured in α-modified Eagle medium (Flow Laboratories, Scotland, KA) with 30% (v/v) fetal calf serum (Gibco BRL, Gaithersburg, MD) supplemented with 1.2% methylcellulose, 1% bovine serum albumin (Sigma Chemical Co.), 0.1 mM 2-mercaptoethanol, and recombinant mouse granulocyte-macrophage colony stimulating factor (10 ng/ml, Intergen Co., Purchase, NY). After incubation for 7 days at 37°C in humidified atmosphere of 5% CO2, colonies containing 40 cells were counted as CFU-GM colonies with an inverted microscope.
Reaction of Purified DPD with BVU.
A reaction mixture containing purified DPD (1.5 μg), 4 or 25 μM BVU, 200 μM NADPH, 2.5 mM MgCl2, and 30% (v/v) glycerol-35 mM K-phosphate buffer, pH 7.4, in a final volume of 50 μl was incubated at 37°C. Aliquots (5 μl) of the reaction mixture were withdrawn at fixed times and immediately assayed for residual DPD activity. For determination of the radioactivity of [14C]BVU (2.0 MBq/μmol) incorporated into DPD protein, a total volume of the incubation mixture was increased up to 300 μl without changing the concentrations of the constituents. The mixture was incubated at 37°C, and aliquots (50 μl) were withdrawn at fixed times, diluted with a large excess (25 times) of BVU, and then rapidly chilled in an ice bath. The radioactivity incorporated into the enzyme protein was separated from unreacted [14C]BVU by HPLC and determined by liquid scintillation counting as described in the legend of fig. 2.
HPLC profile of rat liver DPD adducted with [14C]BVU in the presence of NADPH.
Rat liver DPD isolated by HPLC after incubation with 4 μM [14C]BVU for 20 min in the presence of NADPH was concentrated and re-chromatographed under the same chromatographic conditions. HPLC was carried out on a TSK G2000 SW × L column (300 × 7.8 mm) eluted with 35 mM K-phosphate buffer, pH 7.4 (1 ml/min). The chromatogram was monitored by absorptiometry at 220 nm and by liquid scintillation counting of the column effluent collected every 30 sec. Panels A and B represent chromatograms of the authentic DPD protein and the radio-labeled enzyme protein.
Enzyme Assay.
DPD was purified from young adult female Wistar rats as previously reported (5). DPD activity of the purified enzyme toward 5-FU was assayed by a previously reported method with modifications (4); enzyme source (5 μl) was incubated with 20 μM [14C]5-FU (2.1 MBq/μmol, Moravek Biochemicals Inc., Brea, CA) at 37°C for 5 min in the presence of 200 μM NADPH, 2.5 mM MgCl2, and 10 mM mercaptoethanol in a final volume of 50 μl 30% (v/v) glycerol-35 mM K-phosphate buffer, pH 7.4. DPD activity of rat liver cytosol was assayed with [14C]5-FU as previously reported (10).
Statistical Analysis.
The significance of differences were determined by Student’st test.
Results and Discussion
Inactivation of DPD by Covalent Binding of BVU in the Presence of NADPH.
DPD was isolated from rat liver cytosol and purified 785-fold to homogeneity. The purified enzyme had a specific activity of 816 nmol/mg protein/min toward 5-FU as a substrate and migrated as a single protein band with an apparent molecular mass of 210 kDa on native polyacrylamide gel electrophoresis. The DPD activity was strongly inhibited by preincubations with BVU in the presence of NADPH prior to the incubation with the substrate [14C]5-FU (fig.1). BVU almost completely inhibited DPD activity within 20 min at 4 μM and within 5 min at 25 μM when preincubated in the presence of NADPH. With cofactor NADPH omitted from the preincubation medium, BVU showed no inhibitory effect on DPD. On the contrary, the antiviral drug SRV had no inhibitory effect even at 500 μM on the 5-FU-reducing activity of DPD when preincubated in the presence or in the absence of NADPH.
Relation between inactivation of purified rat liver DPD by BVU and incorporation of [14C]BVU into the enzyme protein.
Rat liver DPD was preincubated with 4 (circle) and 25 (square) μM BVU or [14C]BVU in the presence of NADPH for determining the DPD activity toward 5-FU or the radioactivity incorporated into the enzyme protein as described in Methods. At time 0 of the preincubations with 4 and 25 μM BVU, DPD showed activities of 3810 and 3440 nmol 5,6-dihydro-5-FU formed/mg protein/5 min, representing the enzyme activities after incubations with [14C]5-FU for 5 min in the presence of 0.4 and 2.5 μM BVU, respectively, and the activities were expressed as 100%.
Preincubations of the purified DPD with [14C]BVU indicated that the radioactivity was incorporated into the enzyme protein in a manner reciprocal to the loss of enzyme activity (fig. 1). The radioactivity incorporated into DPD was determined after complete separation of the enzyme protein from [14C]BVU by HPLC on a gel filtration column. The ratio of the radioactivity, determined by liquid scintillation counting, to absorbance at 220 nm in the eluate from the HPLC column was unchanged after the radioactive enzyme protein eluted from the HPLC column was concentrated and rechromatographed under the same chromatographic conditions (fig. 2). Moreover, the radioactivity of [14C]BVU incorporated into DPD was not eliminated from the chromatographically isolated enzyme protein when it was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. No trace amount of the radioactivity of [14C]BVU was incorporated into DPD in the absence of NADPH.
Using a partially purified preparation of DPD from rat liver cytosol, Desgranges et al. previously demonstrated the NADPH-dependent inactivation of DPD by BVU (4). In the present study using radioactive BVU and homogeneous DPD from rat liver cytosol, the previously described irreversible inhibition of DPD by unlabeled BVU was demonstrated to be attributable to covalent binding of a reduced form of BVU to the enzyme protein. We hypothesize that the mechanism of inactivation of DPD by BVU is as follows: BVU may be dihydrogenated by DPD with migration of the vinyl double bond to 5-(2-bromoethyliden)uracil, a reactive allyl bromide type of electrophile, reacting as a suicide inhibitor with a cysteinyl residue in the potential pyrimidine-binding domain of the enzyme. Actually, various thiol compounds, such as cysteine, glutathione, and dithiothreitol, added to the incubation mixture had no retarding effect on the enzyme inactivation and on the radio-labeling of DPD by [14C]BVU (data not shown). In connection with this, 5-IU has been demonstrated to inactivate bovine liver DPD by covalent binding of its active metabolite, 5,6-dihydro-5-IU, to the sulfhydryl group of a cysteine residue located in the putative pyrimidine-binding domain of the enzyme (11). However, the enzyme inactivation by 5,6-dihydro-5-IU metabolically formed from 5-IU was retarded by dithiothreitol to a considerable extent, suggesting that unlike the dihydro-BVU, a part of the dihydro-5-IU formed is released from DPD without reacting with its cysteine residue in the pyrimidine-binding domain (11). Molecularly cloned human and pig liver DPDs also have a pyrimidine-binding domain bearing a cysteine residue (12). The amino acid sequence of rat DPD has very recently been determined and demonstrated to have a cysteine residue in the potential pyrimidine-binding domain, which is highly conserved across species (13).
Decrease in DPD Activity of Liver Cytosol from Rats Given SRV and Increase in Plasma and Tissue Levels of 5-FU in Rats Given FT and SRV.
The DPD activity of liver cytosol from rats given SRV orally once daily for 6 days was markedly decreased to 34.2, 21.1, and 8.9% that of controls at 24 hr after administration on days 1, 4, and 6, respectively.
A toxicokinetic study was performed on days 1, 2, 4, and 6 to determine concentrations of 5-FU in plasma, liver, small intestine, and bone marrow of rats orally administered FT alone or FT and SRV simultaneously once daily for 6 days (fig. 3). A preliminary study indicated that the plasma and tissue 5-FU levels on days 3 and 5 were very similar to those on days 2 and 4, respectively. In rats given FT alone, AUC0–24 hr representing time courses of 5-FU concentrations at 0–24 hr after administration were small for plasma, liver, and small intestine and negligible for bone marrow. However, in rats given FT and SRV, AUC0–24 hr of 5-FU were extremely increased in plasma and in all tissues examined (fig. 3).
AUC of 5-FU in plasma and various tissues of rats given FT alone or FT and SRV.
Rats were orally administered FT alone (open bars) or FT and SRV simultaneously (closed bars) once daily for 6 days. The daily doses of FT and SRV were 60 and 30 mg/kg, respectively. AUC of 5-FU in bone marrow of rats administered FT alone were lower than an undetectable level (5 nmol/g tissue × hr). AUC were estimated by application of trapezoidal rule to the 5-FU levels obtained at 1, 2, 4, 8, and 24 hr after administration of FT alone or FT and SRV. Data are expressed as mean ± SD. Significantly different from mean values in FT-treated rats, *p < 0.05, **p < 0.01.
Plasma and liver of the rats orally administered SRV alone or FT and SRV also contained BVU, e. g., 12.0 nmol/ml and 23.4 nmol/g tissue in plasma and liver, respectively, at 8 hr, the time of maximum concentration of BVU on day 2 after co-administration. On day 6, the plasma and liver concentrations of BVU were 10.8 nmol/ml and 22.3 nmol/g tissue, respectively, at 8 hr after the administration of SRV alone or FT and SRV.
Rats orally administered 3.7 mg BVU/kg once daily showed a hepatic maximum concentration approximately equal to that from SRV orally administered once daily at a dose of 30 mg/kg. The animals administered BVU at the aforementioned dose had markedly decreased hepatic DPD activity throughout the days examined, e. g., the DPD activity was 14% of the controls.
Enhancement in Toxicity of FT in Rats by Co-administration with SRV.
One third of the rats which were orally administered FT and SRV once daily for 6 days died on days 6 to 7; none of the animals remained alive on day 10. However, animals orally administered the same dosage of FT or SRV alone once daily for 20 days showed no appreciable change in their vital signs compared with control animals.
Before death, the animals showed a marked decrease in dietary intake from day 3, and by day 6 the intake was negligible with a concomitant loss of body weight (61.9% of controls). On days 3 to 4 most animals had diarrhea with bloody flux, and by day 6 there were marked decreases in total white blood cells (17.6% that of controls) and platelets (25.7% that of controls). They showed extreme atrophy of intestinal membrane mucosa from day 4.
The co-administration of FT with SRV resulted in a marked toxicity to bone marrow. As early as day 2, bone marrow cells collected from the animals treated with both drugs showed negligible formation of CFU-GM colonies; i.e., colony-forming activity decreased to 2.4% that of controls on day 2 (table 1); no such activity was observed on day 6. .
Decrease in colony-forming activity of CFU-GM isolated from bone marrow of rats orally given FT and SRV 1-a
No appreciable histological and hematological changes were observed in rats orally administered FT or SRV alone once daily for 6 days.
In summary, the present study strongly suggests that the fifteen Japanese patients given 5-FU prodrugs and SRV died from a marked increase in 5-FU levels in various tissues, especially in the bone marrow and intestines, as a result of the irreversible inactivation of hepatic DPD by BVU formed from SRV. Further details of our current toxicokinetic, histological, and hematological studies will be published elsewhere.
Footnotes
-
Send reprint requests to: Dr. Tadashi Watabe, Department of Drug Metabolism and Molecular Toxicology, School of Pharmacy, Tokyo University of Pharmacy and Life Science, 1432–1 Horinouchi, Hachioji-shi, Tokyo 192–03, Japan.
- Abbreviations used are::
- SRV
- sorivudine
- 5-FU
- 5-fluorouracil
- FT
- tegafur
- BVU
- (E)-5-(2-bromovinyl)uracil
- DPD
- dihydropyrimidine dehydrogenase
- HPLC
- high pressure liquid chromatography
- 5-IU
- 5-iodouracil
- AUC
- areas under the curves
- CFU-GM
- colony-forming unit granulocyte-macrophage
- The American Society for Pharmacology and Experimental Therapeutics
