Prognostic significance of thymidylate synthase, dihydropyrimidine dehydrogenase and thymidine phosphorylase protein expression in colorectal cancer patients treated with or without 5-fluorouracil-based chemotherapy.

BACKGROUND: Low tumour expression levels of thymidylate synthase (TS), dihydropyrimidine dehydrogenase (DPD) and thymidine phosphorylase (TP) have been linked with improved outcome for colorectal cancer (CRC) patients treated with 5-fluorouracil (5-FU). It is unclear whether this occurs because such tumours have better prognosis or they are more sensitive to 5-FU treatment. PATIENTS AND METHODS: Associations between TS, DPD and TP levels, determined by tissue microarrays and immunohistochemistry, and survival was evaluated in 945 CRC patients according to treatment status. RESULTS: Low TS and DPD expression associated with worse prognosis in stage II [hazard ratio (HR) = 1.69, 95% confidence interval (CI) (1.09-2.63) and HR = 1.92 (95% CI 1.23-2.94), respectively] and stage III CRC patients treated by surgery alone [HR = 1.39 (95% CI 0.92-2.13) and HR = 1.49 (95% CI 1.02-2.17), respectively]. Low TS, DPD and TP associated with trends for better outcome in stage III patients treated with 5-FU [HR = 0.81 (95% CI 0.49-1.33), HR = 0.70 (95% CI 0.42-1.15) and HR = 0.66 (95% CI 0.39-1.12), respectively]. CONCLUSION: Low TS and DPD expression are prognostic for worse outcome in CRC patients treated by surgery alone, whereas low TS, DPD and TP expression are prognostic for better outcome in patients treated with 5-FU chemotherapy. These results provide indirect evidence that low TS, DPD and TP protein expression are predictive of good response to 5-FU chemotherapy.

Familial deficiency of dihydropyrimidine dehydrogenase. Biochemical basis for familial pyrimidinemia and severe 5-fluorouracil-induced toxicity.

Severe neurotoxicity due to 5-fluorouracil (FUra) has previously been described in a patient with familial pyrimidinemia. We now report the biochemical basis for both the pyrimidinemia and neurotoxicity in a patient we have recently studied. After administration of a "test" dose of FUra (25 mg/m2, 600 microCi[6-3H]FUra by intravenous bolus) to a patient who had previously developed neurotoxicity after FUra, a markedly prolonged elimination half-life (159 min) was observed with no evidence of FUra catabolites in plasma or cerebrospinal fluid and with 89.7% of the administered dose being excreted into the urine as unchanged FUra. Using a sensitive assay for dihydropyrimidine dehydrogenase in peripheral blood mononuclear cells, we demonstrated complete deficiency of enzyme activity in the patient and partial deficiency of enzyme activity in her father and children consistent with an autosomal recessive pattern of inheritance. Patients who are deficient in this enzyme are likely to develop severe toxicity after FUra administration.

Novel Deleterious Dihydropyrimidine Dehydrogenase Variants May Contribute to 5-Fluorouracil Sensitivity in an East African Population.

Clinical studies have identified specific genetic variants in dihydropyrimidine dehydrogenase (DPD; DPYD gene) as predictors of severe adverse toxicity to the commonly used chemotherapeutic 5-fluorouracil (5-FU); however, these studies have focused on European and European-American populations. Our laboratory recently demonstrated that additional variants in non-European haplotypes are predictive of 5-FU toxicity. The objective of this study was to identify potential risk variants in an understudied East African population relevant to our institution's catchment area. The DPYD protein-coding region was sequenced in 588 individuals of Somali or Kenyan ancestry living in central/southeast Minnesota. Twelve novel nonsynonymous variants were identified, seven of which significantly decreased DPD activity in vitro. The commonly reported toxicity-associated variants, *2A, D949V, and I560S, were not detected in any individuals. Overall, this study demonstrates a critical limitation in our knowledge of pharmacogenetic predictors of 5-FU toxicity, which has been based on clinical studies conducted in populations of limited diversity.

Quantitative Contribution of rs75017182 to Dihydropyrimidine Dehydrogenase mRNA Splicing and Enzyme Activity.

Dihydropyrimidine dehydrogenase (DPD; DPYD gene) variants have emerged as reliable predictors of adverse toxicity to the chemotherapy agent 5-fluorouracil (5-FU). The intronic DPYD variant rs75017182 has been recently suggested to promote alternative splicing of DPYD. However, both the extent of alternative splicing and the true contribution of rs75017182 to DPD function remain unclear. In the present study we quantified alternative splicing and DPD enzyme activity in rs75017182 carriers utilizing healthy volunteer specimens from the Mayo Clinic Biobank. Although the alternatively spliced transcript was uniquely detected in rs75017182 carriers, canonically spliced DPYD levels were only reduced by 30% (P = 2.8 × 10-6 ) relative to controls. Similarly, DPD enzyme function was reduced by 35% (P = 0.025). Carriers of the well-studied toxicity-associated variant rs67376798 displayed similar reductions in DPD activity (31% reduction). The modest effects on splicing and function suggest that rs75017182 may have clinical utility as a predictor of 5-FU toxicity similar to rs67376798.

Metabolism and biological activity of 5'-deoxy-5-fluorouridine, a novel fluoropyrimidine.

5'-Deoxy-5-fluorouridine (5'-dFUrd; Roche 21-9738) is a recently synthesized antineoplastic agent with therapeutic potential. The sensitivity of Ehrlich ascites tumor cells in CF-1 mice to 5'-dFUrd, as well as to 5-fluorouridine, was established. 5'-dFUrd was a more effective antitumor agent and was less toxic over a wide dosage range (50 to 400 mg/kg) than the other agents tested as measured by: (a) the ability to prevent gross development of inoculated tumor; (b) 45-day survival; and (c) weight change over the treatment period. With use of these sensitive tumor cells, the intracellular metabolism of 5'-dFUrd in vitro was investigated. Utilizing liquid chromatographic methodology for separation of acid-soluble metabolites, the only detectable metabolic products of 5'-dFUrd were FUra, 5-fluorouridine 5'-monophosphate, and 5-fluorouridine 5'-triphosphate. Novel metabolites of 5'-dFUrd were not detectable in the acid-soluble fraction or in plasma isolated from mice given [14C]5'-dFUrd. The formation of FUra appears to result from the action of nucleoside phosphorylase. 5'-dFUrd was shown to have a Km of 0.633 mM for this enzyme isolated from Ehrlich ascites tumor cells, an affinity similar to that for 5-fluoro-2'-deoxyuridine (Km, 0.278 mM) but much lower than that for 5-fluorouridine (Km, 0.049 mM). Incorporation of radiolabeled drug into the acid-insoluble fraction (representing greater than 95% incorporation into RNA) was also significant. 5-Fluoro-2'-deoxyuridine 5'-monophosphate (FdUMP) was not detectable as an acid-soluble metabolite. However, significant inhibition of thymidylate synthetase activity was detectable by 20 min in cells incubated with 30 microM 5'-dFUrd, suggesting that FdUMP was produced. The production of both 5-fluorouridine 5'-triphosphate and FdUMP appears dependent on the initial expansion of the FUra pool. This correlates with the inability of 5'-dFUrd to form nucleotide directly due to the absence of a 5'-hydroxyl group. It is concluded that the antineoplastic activity of 5'-dFUrd may be dependent on its enzymatic conversion of FUra. The basis for the possible increase in therapeutic index compared with other fluoropyrimidines may involve the rate and duration of the production of the biologically active nucleotides 5-fluorouridine 5'-triphosphate and FdUMP.

Selective activation of 5'-deoxy-5-fluorouridine by tumor cells as a basis for an improved therapeutic index.

The intracellular metabolism of a new fluoropyrimidine, 5'-deoxy-5-fluorouridine (5'-dFUrd), was compared with the metabolism of 5-fluorouracil (FUra), 5-fluorouridine (FUrd), and 5-fluoro-2'-deoxyuridine (FdUrd) in freshly isolated bone marrow cells and Ehrlich ascites tumor cells. Following exposure to tumor cells, all four fluoropyrimidines were metabolized to identical products (i.e., FUra, 5-fluorouridine 5'-monophosphate, 5-fluorouridine 5'-diphosphate, 5-fluorouridine 5'-triphosphate, and 5-fluoro-2'-deoxyuridine 5'-monophosphate), all produced an incorporation of FUra into RNA (FUd greater than FUra greater than FdUrd greater than 5'-dFUrd), and all completely inhibited thymidylate synthetase activity by 1 hr. However, in bone marrow cells, very different patterns were observed. 5'-dFUrd accumulated in the cells, but there were no measurable metabolism, no incorporation of FUra into RNA, and no inhibition of thymidylate synthetase activity. In contrast, both FUra and FUrd were metabolized and produced an incorporation of FUra into RNA (2.7 pmol FUra per micrograms RNA and 4.8 pmol FUra per micrograms RNA at 2 hr, respectively) in bone marrow. Only a minor inhibition of thymidylate synthetase activity was detected. FdUrd also was metabolized by bone marrow cells, produced a low level of FUra incorporation into RNA (0.23 pmol FUra per micrograms RNA at 2 hr), and produced a complete inhibition of thymidylate synthetase activity. Since 5'-dFUrd is not directly cytotoxic itself, its superior therapeutic index compared to other fluoropyrimidines may largely reflect the selective activation of 5'-dFUrd by sensitive tumor cells as opposed to bone marrow cells, which can activate FUra, FUrd, and FdUrd.

Dihydropyrimidine dehydrogenase activity in human peripheral blood mononuclear cells and liver: population characteristics, newly identified deficient patients, and clinical implication in 5-fluorouracil chemotherapy.

Dihydropyrimidine dehydrogenase (DPD) is the initial and rate-limiting enzyme in the catabolism of 5-fluorouracil (FUra), one of the most widely used anticancer drugs. Previous studies from our laboratory demonstrated the clinical importance of DPD in cancer patients (G. D. Heggie, J-P. Sommadossi, D. S. Cross, W. J. Huster, and R. B. Diasio. Cancer Res., 47: 2203-2206, 1987; B. E. Harris, R. Song, S-j. Soong, and R. B. Diasio. Cancer Res., 50: 197-201, 1990), particularly in those with DPD deficiency who experience severe FUra toxicity (including death) following FUra treatment [R. B. Diasio, T. L. Beavers, and J. T. Carpenter. J. Clin. Invest., 81: 47-51, 1988; B. E. Harris, J. T. Carpenter, and R. B. Diasio. Cancer (Phila.), 68: 499-501, 1991]. We now suggest that measurement of DPD activity may be useful in routine screening of cancer patients prior to FUra treatment. In this paper, we describe the following serial studies: (a) we developed a sensitive, accurate, and precise DPD assay and a storage method to stabilize DPD activity, permitting large scale DPD screening in cancer patients; (b) we demonstrated a normal distribution (Gaussian distribution) of human DPD activity from peripheral blood mononuclear cells (PBM-DPD) in a population study. Baselines for PBM-DPD with fresh and frozen samples were 0.425 +/- 0.124 (SD) and 0.189 +/- 0.064 nmol/min/mg protein, respectively. The 95% and 99% distribution ranges for both fresh and frozen samples were also determined, providing criteria for detection of DPD-deficient patients; (c) we identified nine new patients with profound or partial DPD deficiency; (d) we determined a baseline for human liver DPD activity, which was shown to be 0.360 +/- 0.182 nmol/min/mg protein (frozen samples); (e) we did a preliminary evaluation of liver DPD from deficient patients. Low liver DPD activity in two deficient patients correlated with low PBM-DPD activity. Using a polyclonal antibody raised against human liver DPD in our laboratory (Z. Lu, R. Zhang, and R. B. Diasio. J. Biol. Chem., 267: 17102-17109, 1992), Western blot analysis demonstrated decreased DPD protein in the liver cytosol from DPD-deficient patients compared to normal subjects. These results may be useful in improving the effectiveness and/or lessening the toxicity of FUra chemotherapy.

Formation of conjugates of 2-fluoro-beta-alanine and bile acids during the metabolism of 5-fluorouracil and 5-fluoro-2-deoxyuridine in the isolated perfused rat liver.

We have recently demonstrated that the major biliary metabolite of 5-fluorouracil (FUra) in cancer patients is a conjugate of the FUra catabolite 2-fluoro-beta-alanine (FBAL) and cholic acid (D.J. Sweeny, S. Barnes, G. Heggie, and R.B. Diasio. Proc. Natl. Acad. Sci. USA, 84:5439-5443, 1987). This finding prompted us to further examine the metabolism and biliary excretion of clinically relevant concentrations of the fluoropyrimidines FUra and 5-fluoro-2'-deoxyuridine (FdUrd) using the isolated perfused rat liver. During infusion of fluoropyrimidines, rates of appearance of metabolites in bile were similar with both 1 microM FUra and 1 microM FdUrd but were 9-fold higher with 25 microM FUra. Analysis by high performance liquid chromatography demonstrated that unmetabolized fluoropyrimidines and known catabolites (i.e., FBAL) accounted for less than 15% of the total metabolites in bile and that the majority of the biliary metabolites were eluted as three distinct nonpolar compounds. Fast atom bombardment-mass spectrometry demonstrated that these unique metabolites had molecular weights of 497 (peak 1), 497 (peak 2), and 481 (peak 3). These metabolites were hydrolyzed by cholylglycine hydrolase to FBAL and unconjugated bile acids that were identified by gas chromatography-mass spectrometry to be alpha-muricholic acid (peak 1), cholic acid (peak 2), and chenodeoxycholic acid (peak 3). Thus, the major biliary metabolites of FUra and FdUrd were identified as N-(bile acid)-FBAL conjugates. While the N-(bile acid)-FBAL conjugates were the major metabolites in bile, dihydroFUra was the major (greater than 70%) metabolite eliminated into perfusate. In summary, these results demonstrate that FUra and FdUrd undergo similar metabolism in the isolated perfused rat liver and, as was observed in humans, the major biliary fluoropyrimidine metabolites are conjugates of FBAL and bile acids.

5-Fluorouracil incorporation into DNA of CF-1 mouse bone marrow cells as a possible mechanism of toxicity.

Isolated CF-1 mouse bone marrow cells were exposed for 1 hr to 5-fluorouracil (FUra) at concentrations from 1.8 to 50 microM and then washed and suspended in a soft agar growth medium to assess the effect on toxicity (measured as reduction in colony growth compared to control). These data were used to determine specific toxic concentrations ranging from 25 to 90% lethal doses. Subsequent studies examined in parallel the effect of these toxic concentrations of FUra on the possible sites of toxicity including: (a) inhibition of thymidylate synthetase activity using a modified 3H release assay; (b) incorporation of FUra into RNA (FUra-RNA); and (c) incorporation of FUra into DNA (FUra-DNA). Thymidylate synthetase activity was slightly decreased (75% of control) after 1-hr exposure to a 50% lethal dose and was not significantly further reduced as the FUra concentration was increased to an 85% lethal dose. Furthermore, subsequent exposure of FUra-treated cells to a nontoxic thymidine dose (5 microM) failed to reverse toxicity. FUra-RNA increased during 1-hr exposure to increasing concentrations of FUra (25 to 90% lethal doses). Although initially suggesting a relationship between the level of FUra-RNA and toxicity, subsequent studies in cells exposed to FUra in the presence of uridine demonstrated a significantly decreased toxicity while, at the same time, a minimal decrease of FUra in RNA. In contrast, FUra-DNA was significantly decreased in the presence of uridine and correlated with decreased toxicity. In additional subsequent studies, an apparent decrease in subsequent DNA synthesis was observed (measured by 32P or [3H]thymidine incorporation into DNA) as the level of FUra-DNA increased. In conclusion, FUra is demonstrated to be incorporated into DNA of isolated CF-1 mouse bone marrow cells, and the level of FUra-DNA appears to be closely associated with toxicity and inhibition of further DNA synthesis. The parallel studies of thymidylate synthetase activity and FUra-RNA suggest that FUra-DNA may be an unrecognized mechanism of FUra toxicity in these cells.

Relationship between circadian-dependent toxicity of 5-fluorodeoxyuridine and circadian rhythms of pyrimidine enzymes: possible relevance to fluoropyrimidine chemotherapy.

Previous studies in experimental animals and patients have suggested a circadian variation in host toxicity following administration of 5-fluorodeoxyuridine (FdUrd) although the biochemical mechanisms are not fully understood. Thymidine kinase (TK; EC 2.7.1.21), the initial enzyme in the thymidine-phosphorylation pathway, is the first enzyme in the anabolism of FdUrd. Dihydropyrimidine dehydrogenase (DPD; EC 1.3.1.2), is the rate-limiting enzyme in the pyrimidine catabolic pathway and has been shown to be the key enzyme in FdUrd catabolism. The present study examined the relationship between the suggested circadian variation in FdUrd toxicity and potential circadian variations in the activity of these enzymes. Initial studies in Sprague-Dawley rats confirmed that the time of FdUrd administration affected death rate and other drug-related toxicities including loss of body weight, diarrhea, and bone marrow suppression, with the least toxicity and highest survival rate being observed in rats receiving FdUrd at 12:00 noon and 4:00 p.m. and the greatest toxicity and lowest survival rate at 12:00 midnight and 4:00 a.m. Statistical analysis revealed a circadian pattern in FdUrd toxicity (Cosinor analysis, P < 0.001). In subsequent studies with the same species, we simultaneously measured TK and DPD activities in several tissues at various times over 24 h. Under standardized light conditions (lights on, 6:00 a.m. to 6:00 p.m.; lights off, 6:00 p.m. to 6:00 a.m.), with sampling at 4-h intervals (4:00 and 8:00 a.m.; 12:00 noon; 4:00 and 8:00 p.m., and 12:00 midnight), a circadian variation in TK activity was observed (P < 0.0001, Cosinor analysis) in bone marrow, intestinal mucosa, liver, and spleen. In the same group of animals, a circadian pattern of DPD activity in liver and bone marrow was also observed (Cosinor analysis, P < 0.0001) that was inverse compared to the circadian variation in TK activity (Pearson correlation analysis, P < 0.05). Further statistical analysis indicated that the observed circadian variation in FdUrd toxicity was correlated with the circadian variation of TK activity and inversely correlated with DPD activity (Pearson correlation analysis, P < 0.05). Based on the above data, we conclude that the circadian pattern of TK and DPD activity may explain the observed circadian variation in toxicity as the time of FdUrd administration is varied. These results may be useful in the design of improved chemotherapeutic regimens using time-modified administration of FdUrd.

Alteration of the secondary structure of newly synthesized DNA from murine bone marrow cells by 5-fluorouracil.

The present study evaluates the effects of 5-fluorouracil (FUra) on the structure of newly synthesized DNA purified from bone marrow cells. DNA synthesis was decreased by 30 and 45% of control in the presence of 19 and 100 microM FUra, respectively. Furthermore at these concentrations of FUra, the DNA strand sizes were smaller as determined by alkaline sucrose gradients. Enzymatic digestion of the DNA demonstrated that most of the FUra (greater than 90%) was localized in the internucleotide linkage and not at the chain terminus. As the concentration of FUra was varied, the percentage of FUra at the chain terminus was unchanged, suggesting that the decrease in chain size as well as inhibition of DNA synthesis was not due to chain termination. DNA that had been synthesized in the presence of FUra was shown to fragment after increasing time as demonstrated by alkaline sucrose gradient analysis. This time-dependent fragmentation was associated with an increased number of strand breaks as determined by neutral and alkaline sucrose gradient analysis. A parallel study demonstrated a time-dependent excision of FUra from DNA over this same time period. In summary, these studies demonstrate an association between the excision of FUra from DNA and the changes in secondary structure of newly synthesized DNA.

Decreased immunosuppression associated with antitumor activity of 5-deoxy-5-fluorouridine compared to 5-fluorouracil and 5-fluorouridine.

5-Fluorouracil (5-FUra), 5-deoxy-5-fluorouridine (5'dFUrd), and 5-fluorouridine were compared for their relative antitumor activity, their capacity to inhibit leukocyte exudation and macrophage (macrophage) killing of tumor cells in vivo and in vitro, and their ability to induce leukopenia and monocytopenia. 5'dFUrd was less toxic than 5-FUra and exhibited anti-Ehrlich ascites activity over a wider range of drug doses. Inflammatory exudates induced by thioglycollate or pyran were inhibited up to 91% by prior 5-FUra injection but were inhibited not more than 62% by 5'dFUrd. Pyran-induced macrophage inhibition of Ehrlich ascites proliferation in vivo was diminished up to 5-fold by 5-FUra but was never diminished more than 2-fold by 5'dFUrd, while neither agent suppressed in vitro macrophage cytotoxicity of in vivo pyran-activated macrophage. At high doses, 5-FUra reduced white blood cell counts 73%, in contrast to the 8% reduction caused by 5'dFUrd, while at their optimal anti-Ehrlich ascites doses, 5-FUra and 5'dFUrd both lowered white blood cell counts by only 20%. However, 5-FUra caused a severe monocytopenia not seen in animals given injections of comparable doses of 5'dFUrd. Therefore, 5-FUra appeared to inhibit the inflammatory response and antitumor activity by inhibiting the influx of immature macrophage into the peritoneal cavity, not by inhibiting the function of mature effector cells.

Relationship between dihydropyrimidine dehydrogenase activity and plasma 5-fluorouracil levels with evidence for circadian variation of enzyme activity and plasma drug levels in cancer patients receiving 5-fluorouracil by protracted continuous infusion.

The activity of dihydropyrimidine dehydrogenase (DPD) in peripheral blood mononuclear cells and plasma concentration of 5-fluorouracil (FUra) were simultaneously determined in cancer patients receiving FUra by protracted continuous infusion (300 mg/m2/day). Blood samples were drawn every 3 h over 24-h period and the resulting DPD and FUra values analyzed for circadian periodicity. In the seven patients studied, a circadian rhythm of DPD activity was observed (P less than 0.00001, Cosinor analysis) with the peak of activity at 1 a.m. (0.197 +/- 0.007 nmol/min/mg) and the trough at a 1 p.m. (0.113 +/- 0.007 nmol/min/mg). In addition, a circadian rhythm was observed for the plasma concentrations of FUra obtained over a 24-h period (P less than 0.00001, Cosinor analysis) with peak values (27.4 +/- 1.3 ng/ml) occurring at 11 a.m. and trough values (5.6 +/- 1.3 ng/ml) occurring at 11 p.m. The ratio of the maximum concentration of FUra to the minimum concentration observed was almost 5-fold. This study demonstrates a circadian variation of DPD activity in human peripheral blood mononuclear cells and a circadian variation of FUra plasma levels in patients receiving FUra by protracted continuous infusion. An inverse relationship between the circadian patterns of DPD activity and FUra plasma levels was also noted, suggesting that an association may exist between DPD activity and FUra plasma concentration. Further evidence of an association between DPD activity in peripheral blood mononuclear cells and plasma FUra concentration was demonstrated by a linear relationship between the two parameters in all patients (r = -0.627) and within individual patients (-0.978 less than r less than -0.742). With the recent advent of programmable pumps, information on the circadian pattern of FUra and/or DPD may be useful in planning continuous infusion schedules in order that optimal plasma drug concentration may be maintained over a 24-h cycle, thereby enhancing the therapeutic efficacy of FUra administered by continuous infusion.

Circadian variation of 5-fluorouracil catabolism in isolated perfused rat liver.

The catabolism of 5-fluorouracil (FUra) was measured in isolated perfused rat liver (IPRL) at various times of the day. IPRLs were prepared from rats sacrificed at 3-h intervals and the elimination rate of FUra and FUra catabolites (i.e., rate leaving the IPRL in the effluent perfusate) following infusion of [3H]FUra was analyzed for circadian periodicity. Animals were housed under standardized conditions of light and dark and divided into two groups of 24 animals each. The first group was housed under "normal" light conditions (lights on from 6:00 a.m. to 6:00 p.m.; off from 6:00 p.m. to 6:00 a.m.), while the second group was housed under "reverse" light conditions (lights on from 10:00 p.m. to 10:00 a.m.; off from 10:00 a.m. to 10:00 p.m.). A circadian rhythm was observed in the elimination rate of FUra and FUra catabolites by both groups (P less than 0.0001, Cosinor analysis). Under "normal" light conditions, peak and trough elimination rate of FUra was at 19 h after light onset (HALO; 183.8 +/- 3.4 nmol/min/g liver) and 7 HALO (123.8 +/- 3.4 nmol/min/g liver), respectively. There was a reciprocal relationship between the elimination rates of FUra and FUra catabolites with peak and trough values for FUra catabolites at 7 HALO (70.5 +/- 3.6 nmol/min/g liver) and 19 HALO (17.5 +/- 3.6 nmol/min/g liver), respectively. Animals housed under the "reverse" conditions of light and dark also exhibited a circadian pattern. Under the "reverse" conditions, the peak and trough elimination rate of FUra was at 18.5 HALO (170.0 +/- 1.7 nmol/min/g liver) and 6.5 HALO (130.0 +/- 1.7 nmol/min/g liver), respectively. The peak and trough elimination rate of FUra catabolites under these conditions occurred at 6.5 HALO (64.3 +/- 2.2 nmol/min/g liver) and 18.5 HALO (29.7 +/- 2.2 nmol/min/g liver), respectively. These results demonstrate that the elimination rate of FUra and FUra catabolites by IPRL varies over a 24-h period with a circadian rhythm in association with the light/dark cycle. Such a variation in the hepatic elimination rate of FUra in humans could result in a variation in the systemic level of drug during chemotherapy thus affecting the therapeutic efficacy of FUra. This study suggests that a circadian pattern in the hepatic catabolism of FUra needs to be considered when planning chemotherapeutic regimens with FUra.

Structural organization of the human dihydropyrimidine dehydrogenase gene.

Deficiency of the pyrimidine catabolic enzyme, dihydropyrimidine dehydrogenase (DPD), has been shown to be responsible for a pharmacogenetic syndrome in which administration of 5-fluorouracil is associated with severe and potentially life-threatening toxicity. Following the recent availability of the cDNA for DPD, there were initial reports of several molecular defects (point mutations, deletions due to exon skipping) that were suggested as a potential molecular basis for DPD deficiency, even before the complete physical structure of the DPD gene was known. To understand the mechanism responsible for DPD deficiency, we have determined the genomic structure and organization of the human DPD gene. The gene is approximately 150 kb in length, and it consists of 23 exons, ranging in size from 69 to 1404 bp. The sequences of intronic regions flanking the exon boundaries have been determined. The physical map of the DPD gene should permit development of rapid assays to detect point mutations or small deletions in the DPD gene associated with 5-fluorouracil toxicity.

Dihydrofluorouracil, a fluorouracil catabolite with antitumor activity in murine and human cells.

Dihydrofluorouracil (FUH2), the initial catabolite of 5-fluorouracil (FUra), was examined to determine whether this derivative had antitumor activity or host cell (bone marrow) toxicity. Studies were undertaken with Ehrlich ascites tumor and bone marrow cells isolated from CF-1 mice. Cells were exposed for 1 h either to no drug (control) or to varying concentrations, ranging from 1 to 250 microM, of either FUra, FUH2, or alpha-fluoro-beta-alanine. Cells were then cultured and colony formation was assessed after 10 to 14 days. Ehrlich ascites tumor cells were more sensitive to FUra [50% lethal dose (LD50) = 18 microM] than to FUH2 [LD50 = 50 microM], with no sensitivity to alpha-fluoro-beta-alanine even at 250 microM. Bone marrow cells had a toxicity profile similar to that of FUra (LD50 = 10 microM) but were relatively insensitive to FUH2 (LD50 greater than 250 microM), with no sensitivity to alpha-fluoro-beta-alanine. Subsequent studies examined colony formation of the human breast carcinoma cell line MCF-7 following 1 h exposure to varying concentrations of FUra and FUH2. These cells were less sensitive to both FUra (LD50 approximately 80 microM) and FUH2 (LD50 approximately 350 microM). Initial studies on the mechanism of toxicity of FUH2 demonstrated that this FUra catabolite could produce inhibition of thymidylate synthase activity in Ehrlich ascites tumor cells with a pattern similar to that resulting from exposure to FUra. This is the first study to demonstrate that FUH2 (a quantitatively important catabolite of FUra) is cytotoxic, and it suggests that FUH2 may contribute to the toxicity of FUra in vivo, possibly by being anabolized to FUra.

Clinical pharmacokinetics of 5-fluorouracil and its metabolites in plasma, urine, and bile.

Kinetics of 5-fluorouracil (FUra) and FUra metabolites in plasma and urine were investigated in 10 cancer patients following i.v. bolus administration of 500 mg/m2 FUra with 600 microCi of [6-3H]FUra. Biliary excretion was examined in two patients with external biliary catheters. Quantitation of unchanged drug and metabolites was assessed by a highly specific high-performance liquid chromatographic method. FUra plasma levels declined rapidly with an apparent elimination half-life of 12.9 +/- 7.3 min. Dihydrofluorouracil was detected within 5 min in most patients, demonstrating rapid catabolism and reached maximum peak levels of 23.7 +/- 9.9 microM at approximately 60 min. The apparent elimination half-life of dihydrofluorouracil (61.9 +/- 39.0 min) was consistently greater than that of the unchanged drug. The apparent elimination half-lives of the subsequent metabolites alpha-fluoro-beta-ureidopropionic acid and alpha-fluoro-beta-alanine were prolonged with values of 238.9 +/- 175.4 min and 1976 +/- 358 min, respectively. Approximately 60-90% of the administered dose was excreted in urine within 24 h, primarily as alpha-fluoro-beta-alanine. Biliary excretion accounted for 2-3% of total administered radioactivity. The major fraction of this radioactivity eluted on high-performance liquid chromatography as a previously unrecognized FUra metabolite. Analysis of its structure is currently ongoing in our laboratory. In conclusion, this study provides the first comprehensive analysis of the formation and excretion of FUra metabolites in plasma, urine, and bile following i.v. bolus administration of FUra in humans.

Modulation of 5-fluorouracil catabolism in isolated rat hepatocytes with enhancement of 5-fluorouracil glucuronide formation.

The catabolism of 5-fluorouracil (FUra), which accounts for 90% of the elimination of this antimetabolite in vivo, has recently been characterized in freshly isolated rat hepatocytes in suspension using a highly specific high-performance liquid chromatographic methodology. The present study evaluates the effect of thymine and uracil, which are thought to be catabolized by the same enzymes as FUra, on the metabolism and transmembrane distribution of FUra in isolated rat hepatocytes. Following simulataneous exposure of cells for 5 min to 30 microM [6-3H]FUra and increasing concentrations of either thymine or uracil, dihydrofluorouracil (FUH2) levels decreased in a concentration-dependent manner, and the concentration determined for 50% inhibition of FUra catabolism was 8.0 +/- 0.3 (S.D.) and 67.8 +/- 15.6 microM for thymine and uracil, respectively. Analysis of intracellular and extracellular 3H from 1 min to 2 hr after simultaneous incubation of the hepatocytes with 30 microM FUra and thymine (or uracil) in a 1:7 molar ratio resulted in a decrease of intracellular and extracellular FUH2 and alpha-fluoro-beta-alanine (FBAL), while alpha-fluoro-beta-ureidopropionic acid (FUPA) was enhanced. Unmetabolized FUra (not detected in the absence of thymine or uracil) was detected intracellularly in the presence of thymine or uracil and was accompanied by the appearance of a novel metabolite, preliminarily identified as a glucuronide of the FUra base which reached intracellular levels of 44 +/- 9.76 and 27.45 +/- 1.35 microM in the presence of thymine or uracil, respectively, within 1 hr. This metabolite, which penetrates the cell membrane only slowly, accounted for approximately 60% of the intracellular 3H in the presence of 300 microM FUra and 2 mM thymine, whereas FUra catabolism was inhibited by more than 99% under these conditions. The formation of FUra anabolites was insignificant in the presence of thymine and uracil, and incorporation of FUra into RNA was not enhanced. The lack of anabolism of FUra in isolated hepatocytes exposed to either high initial concentrations of FUra or high intracellular FUra concentrations resulting from modulation (inhibition) of FUra catabolism is consistent with the clinical observation of minimal hepatotoxicity with FUra, despite exposure of the liver to high blood levels. These studies indicate that thymine is a more potent modulator of FUra catabolism in hepatocytes than is uracil. Further studies are needed to clarify the biological importance of the glucuronide of the base FUra which accumulates intracellularly as the concentration of FUra increases within the hepatocytes.

Evidence from rat hepatocytes of an unrecognized pathway of 5-fluorouracil metabolism with the formation of a glucuronide derivative.

Isolated rat hepatocytes in suspension were exposed to [3H]-5-fluorouracil for intervals over 2 h, following which the cells were removed from the media and sonicated, and the cytoplasm was sampled. High-performance liquid chromatography was used to separate 5-fluorouracil (FUra) from its known anabolites and catabolites, with subsequent quantitation of these metabolites by measurement of radioactivity. As the extracellular concentration of FUra was increased above 30 microM, the intracellular levels of FUra increased, with detection of a new peak of radioactivity distinct from any of the known anabolites or catabolites. This new metabolite, "G," increased in concentration as the extracellular concentration of FUra was raised above 1 mM. Inhibition of FUra catabolism by 2 mM thymine resulted in a further increase in intracellular FUra (approaching the extracellular FUra concentration) and was accompanied by a further increase in the intracellular concentration of "G," demonstrating that "G" was not formed via the catabolic pathway. The increase in intracellular FUra and "G" was not accompanied by an increase in intracellular anabolites, suggesting that "G" was formed via a novel metabolic pathway. "G" was retained within the hepatocytes, although it was not bound to intracellular macromolecules. "G" was converted to FUra in the presence of beta-D-glucuronidase; this reaction was inhibited with the addition of saccharo-1,4-beta-lactone, a specific inhibitor of the beta-D-glucuronidase. This data, together with evidence from hepatocyte homogenates in which formation of "G" was shown to be dependent on the concentration of uridine-5'-diphosphoglucuronic acid, demonstrates that "G" is a glucuronide of FUra. The formation of "G" suggests that FUra is metabolized via a previously unrecognized metabolic pathway.

Disposition and metabolism of 2-fluoro-beta-alanine conjugates of bile acids following secretion into bile.

Since 2-fluoro-beta-alanine (FBAL) conjugates of bile acids (BA), the primary biliary metabolites of fluoropyrimidine (FP) drugs, have been suggested to be related to the hepatotoxicity which develops in patients receiving FP chemotherapy by intrahepatic arterial infusion (Proc. Natl. Acad. Sci. USA 84, 5439-5443, 1987), it was important to determine whether they undergo enterohepatic circulation and hence accumulate in the liver and biliary system. In initial studies, sensitivity of FBAL-BA conjugates to hydrolysis by pancreatic enzymes was examined. In subsequent in vivo studies, a model FBAL-BA conjugate, FBAL-chenodeoxycholate (FBAL-CDC), was introduced into the lumen of the small intestine of anesthetized rats with biliary fistulas to quantitate the intestinal absorption, metabolism and tissue distribution of the conjugate. The results indicated that: (1) FBAL-BA conjugates were resistant to hydrolysis by pancreatic enzymes (carboxypeptidase A, carboxypeptidase B and trypsin) and by human pancreatic juice, but were completely hydrolyzed by cholyglycine hydrolase. (2) At least one-half of the administered FBAL-CDC was deconjugated during the process of intestinal absorption, as shown by HPLC analysis of the radioactivity in portal venous blood. (3) Deconjugated FBAL or CDC was reconjugated in liver with other bile acids or amino acids (glycine and taurine), respectively, as shown by radiochromatography of bile. (4) FBAL, formed as a result of hydrolysis of FBAL-CDC, had a wide tissue distribution. In conclusion, FBAL-CDC has a rapid turnover during its enterohepatic circulation due to deconjugation in the intestine and reconjugation in the liver.