Pharmacokinetics and bioavailability of oral 5′-deoxy-5-fluorouridine in cancer patients

Abstract

Aims

Oral administration of 5-fluorouracil (FUra), an important cytotoxic agent, is limited by a wide variation in bioavailability. 5′-deoxy-5-fluorouridine (dFUrd), a masked form of FUra, has shown promise clinically when given intravenously or orally as a solution or tablet. This study investigates the efficacy of an oral capsule formulation of dFUrd in generating continuous systemic levels of this compound in cancer patients.

Methods

Six patients with advanced intestinal or ovarian malignancies were given three cycles of dFUrd, days 1–5, at intervals of 4 weeks. The doses of dFUrd were 600 mg m⁻² three times daily, 800 mg m⁻² three times daily, and 1000 mg m⁻² three times daily, on cycles one, two and three, respectively (total dose 36 g m⁻²). The initial dose in each cycle was given as a slow intravenous injection over 10 min, and the remainder orally. Plasma and urine levels of dFUrd and two of its metabolites, FUra and 5,6-dihydro-5-fluorouracil (FUraH₂), were monitored in six patients at each dose level.

Results

All six patients completed the study, receiving three different doses over a 3 month period, following which one had achieved a partial response, one had stable disease, and four had developed progressive disease. Side-effects were negligible, and only two instances of transient diarrhoea WHO grade 1 were seen. Total body clearance (CLtot) of intravenous dFUrd decreased with increasing dose; 2.7, 2.0 and 1.3 l min⁻¹ m⁻², following doses of 600, 800 and 1000 mg m⁻², respectively. The mean elimination half-life of intravenous dFUrd increased with the dose from 15 to 22 min. Oral dFUrd was rapidly absorbed with a lag time of less than 20 min. The mean elimination half-life (t_{1/2, z}) of oral dFUrd was 32–45 min in the dose range 600–1000 mg m⁻². The AUC of FUra and FUraH₂ increased overproportionally with increasing intravenous doses of dFUrd. The mean systemic bioavailability of oral dFUrd was 34–47%.

Conclusions

dFUrd, which selectively releases the antimetabolite FUra in tumour cells, can be given orally at doses of 600–1000 mg m⁻² three times daily for 5 days. The systemic levels achieved are equivalent to those seen following continuous infusions of dFUrd or FUra. Toxicity is tolerable, and further clinical investigation of oral dFUrd is warranted.

Introduction

5-fluorouracil (FUra) is an established anticancer agent, particularly in gastrointestinal malignancies. It is given intravenously as there is a wide variation in oral bioavailability of 0–80% [1–3]. Two common intravenous bolus regimens in colorectal cancer are daily injections for 5 days every 4 weeks, or weekly injections. As such treatment often continues for 6 months, frequent hospital visits are required. Continuous intravenous infusion of FUra through an indwelling venous catheter is becoming increasingly popular [4], permitting a higher dose intensity and overcoming the disadvantage of its short half-life.

5′-deoxy-5-fluorouridine (dFUrd) is a FUra prodrug, and has proven activity when given intravenously [5]; a response rate of 26% [6] is similar to that seen with FUra. More recently dFUrd has been administered orally in the form of a solution or tablet. Effective oral delivery of FUra has advantages for the patient, including less frequent hospital visits and the absence of an indwelling venous catheter. There is a need for further investigation and development of oral formulations of dFUrd. The aim of the study was to evaluate the pharmacokinetics, bioavailability, and if possible, the efficacy of an oral capsule formulation of dFUrd using a dose escalation scheme of 9, 12 and 15 g m⁻².

Methods

The study was approved by the local ethics committees of participating institutions, and each patient gave written informed consent. Seven patients (age 37–67 years) were included in the study: one had carcinoma of the jejunum with peritoneal involvement; two patients ovarian carcinoma, one with peritoneal and one with liver metastases; and four had rectal carcinoma with liver metastases. All patients had received FUra previously. Patients were ambulatory with minimum disability (Karnofsky performance score ≥80), and had adequate liver function (total bilirubin <20 μmol l⁻¹, albumin >30 g l⁻¹, and prothrombin ratio <1.3). Full blood count, haematocrit, serum urea and electrolytes, liver function tests and urinalysis were determined in each patient before the study.

Intravenous dFUrd, in sterile vials containing 1 g, was kindly supplied by Hoffmann-La Roche Basle, and oral capsules containing 200 mg by Hoffmann-La Roche Tokyo. Injection solutions were prepared immediately before drug administration. dFUrd was given three times daily (at intervals of 4 h during pharmacokinetic sampling), days 1–5, every 4 weeks for three cycles. Nine g ⁻² (600 mg m⁻² three times daily for 5 days) was given in the first cycle, increasing to 12 g m⁻² (800 mg m⁻² three times daily for 5 days) in cycle two, and 15 g m⁻² (1000 mg m⁻² three times daily for 5 days) in the third cycle. The initial dose in each cycle was given as a slow intravenous injection over 10 min, and the remainder orally.

A cannula with a three-way stoplock was placed in the contralateral arm, and a 10 ml aliquot of blood collected in a heparinized tube prior to drug administration, and at regular intervals thereafter for 8 h. The plasma was separated immediately and stored at −20° C to minimize degradation of dFUrd [7, 8], FUra [9] and 5,6-dihydro-5-fluorouracil (FUraH₂) [10, 11]. A preadministration urine sample, and all urine passed during the period of blood sampling, were collected in plastic bottles. The volume of each sample was measured, and an aliquot stored at −20° C.

Drug analysis

dFUrd and FUra were determined by reverse phase h.p.l.c. with u.v. detection [7]. After addition of the internal standard 2-thiouracil to 500 μl plasma, the mixture was adjusted to pH 6 with 5 m^M TRIS buffer and extracted with 7 ml ethyl acetate. Precipitation of proteins was not necessary. The organic phase was evaporated at room temperature under a gentle stream of nitrogen. The residue was dissolved in 200 μl 0.05 m TRIS buffer (pH 8), 20 μl of which was injected on to a 100-mm MOS-Hypersil 5-μm column in an h.p.l.c. system. dFUrd and FUra were detected at 269 nm. Elution was carried out isocratically at 2 ml min⁻¹ with water-0.05 mM TRIS-0.005 mM cetrimide. The retention times of FUra, 2-thiouracil, and dFUrd were 4.1, 6.5 and 24.3 min, respectively. Detection limits, based on a signal to noise ratio of 3:1, were 35 ng ml⁻¹ of dFUrd and 7 ng ml⁻¹ of FUra. The coefficient of variation in three successive determinations within the range of the calibration curve, 35 ng ml⁻¹ to 100 μg ml⁻¹ [7], was 3%. The method was linear over plasma and urine concentration ranges of 50 ng ml⁻¹–10 mg ml⁻¹ for both dFUrd and FUra.

In the urinalysis of dFUrd and FUra, 2-thiouracil was added to 1 ml urine and the mixture diluted 1:10 with water; 200 μl of the dilution was introduced into the h.p.l.c. system, and the analysis performed in the same fashion as plasma analysis. The linearity and precision of the urine assay are comparable with those of the plasma assay [7].

FUraH₂ was determined by a modified assay, based on gas chromatography with home-made OV-275 columns [12] installed in a Hewlett-Packard 5731 gas chromatograph equipped with a ⁶³Ni pulse-modified electron capture detector. Samples were introduced through a solid sample injection system. Special care was taken on account of the instability of FUraH₂ in biological fluids [10, 11]; the internal standard chlorouracil (CUra) was added to 200 μl of plasma, which had been thawed to 8° C, followed by acetate buffer (1.5 m, 100 μl, pH 3.5). The mixture was then extracted twice with 3 ml chloroform to remove interfering compounds [13], followed by 2×3 ml ethyl acetate. The ethyl acetate fractions were evaporated under nitrogen, the residue dissolved in a further 40 μl of ethyl acetate, and the solution introduced into the solid sample injection device of the gas chromatograph [13]. Using this technique FUra was determined concomitantly [12]. The limits of detection of FUra and FUraH₂ in plasma (signal to noise ratio 3:1) were 7 and 5 ng ml⁻¹, respectively, and the coefficient of variation for the assay ranged from 6.9% at a level of 100 ng ml⁻¹–0.9% at 5 μg ml⁻¹ (n=3). The recovery of FUraH₂ extracted from plasma was 58±2%.

Pharmacokinetic analysis

The elimination half-lives of dFUrd, FUra and FUraH₂ were calculated by linear regression analysis of the terminal part of the log concentration-time curve. The area under the plasma concentration time curve (AUC) was calculated by the trapezoidal rule with extrapolation to infinity. For dFUrd and FUra the extrapolated AUC was less than 5% of the total AUC, and in the case of FUraH₂, less than 20%. The oral bioavailability of dFUrd (F) was calculated according to eqn 1:

1

where o and i.v. refer to oral and intravenous administration, respectively.

Percentage urinary recovery of dFUrd (Au) was calculated by eqn 2

2

where Xu,_0–12 represents the amount of dFUrd recovered from urine in the 12 h following the first dose of dFUrd.

Total body clearance (Cltot) of dFUrd was determined from eqn 3

3

Statistical analysis

The results are expressed as mean±s.d., unless otherwise stated. Differences in the mean values of pharmacokinetic parameters were assessed, and biovailability analysed by one-way analysis of variance (anova). Differences were considered significant when P<0.05.

Results

Tolerability of low dose dFUrd

All six patients completed the study, receiving a total dose of 36 g m⁻² in 3 months, without serious side-effects. Two cases of mild transient diarrhoea, WHO grade 1 [14], were observed. No nausea, vomiting, stomatitis, hair loss, peripheral or central neurotoxicity, leucopenia, nor thrombocytopenia was seen. Serum creatinine levels remained normal during the treatment period.

Antitumour efficacy of continuous low dose oral dFUrd

One patient had a partial response (a greater than 50% decrease in the cross-sectional area of measurable lesions), one stable disease (a decrease of less than 50%, or an increase of less than 25%, in the extent of the disease, with the appearance of no new lesions), and four showed disease progression (an increase of greater than 25% in the area of measurable disease or the appearance of new lesions) [14].

Plasma pharmacokinetics and metabolism of low dose dFUrd

Intravenous administration

The plasma concentration-time decays of low dose intravenous dFUrd exhibited monophasic characteristics; a curve obtained after a dose of 800 mg m⁻² is shown in Figure 1.

Figure 1.

Example of plasma concentration-time curves of dFUrd and FUra after intravenous and oral administration of 800 mg m⁻² dFUrd. □ FUra i.v., dFUrd i.v., ○ FUra oral, • dFUrd oral.

Immediately following the administration of intravenous dFUrd (600–1000 mg m⁻²), maximum concentrations (C_max) were between 10 and 200 μg ml⁻¹, rapidly declining monophasically within 4 h to the lower ng ml⁻¹ level with doses of 600 or 800 mg m⁻². One patient behaved biphasically when given 1000 mg m⁻² dFUrd.

FUra was detected within 10 min of commencing intravenous dFUrd (Figure 1). FUraH₂ concentrations reached 100–300 ng ml⁻¹ within 30–45 min of the start of intravenous dFUrd, followed by a decline over the next 2–3 h (data not shown). The FUra and FUraH₂ concentration-time curves have similar shapes. Convexity of dFUrd, FUra or FUraH₂ curves was not observed following intravenous administration.

Significant changes in the AUC, clearance and elimination half-life of intravenous dFUrd were observed. The AUC (AUCivdFUrd) increased nonlinearly with the dose, from 235 to 755 μg min⁻¹ ml⁻¹ (Table 1). The total body clearance (CLtot) of dFUrd decreased from 2.7 to 1.3 l min⁻¹ m⁻², with the elimination half-life increasing from 15 to 22 min. With increasing intravenous doses of dFUrd, there was a nonlinear increase in AUCi.v.FUra (Figure 2). The degree of nonlinearity was similar to that seen with AUCi.v.dFUrd: 21.2 μg ml⁻¹ min (dose 600 mg m⁻²) to 59.1 μg ml⁻¹ min (dose 1000 mg m⁻²). The elimination half-life increased from 16 to 22 min, but this was not significant.

Table 1.

Summary of dFUrd, FUra and FUraH₂ pharmacokinetics after oral and intravenous dFUrd (n=6).

Figure 2.

Nonlinear increase in AUC FUra (μg ml⁻¹ min) with increasing doses of dFUrd (600–1000 mg m⁻²). The increase in AUC FUra following oral dFUrd was linear. • oral, ▪ i.v.

In parallel, the increase in AUCivFUraH₂ was not proportional to an increase in the intravenous dose of dFUrd: a 257% increase in AUCivFUraH₂ with a 170% dose increment (Table 1). The elimination half-life of FUraH₂ remained relatively stable: 59 min at 600 mg m⁻²–65 min at 1000 mg m⁻².

Oral administration

Orally administered dFUrd is rapidly absorbed with a lag time (t_lag) of 10–30 min for a dose of 600 mg m⁻², and less than 25 min following doses between 800 and 1000 mg m⁻². Peak concentrations (C_max) were attained within 1.5 h, with a median t_max of 60 min. Concentrations steadily decreased after t_max, with a mean half-life of 32–45 min (Figures 1 and 3; Table 1). The mean oral bioavailability of dFUrd was 47%, 34% and 37%, after 600, 800 and 1000 mg m⁻², respectively. These bioavailability values were not significantly different (anova; T=2.498; P>0.05).

Figure 3.

An example of plasma concentration-time curves of dFUrd (•), FUra (○) and FUraH₂ (▴) following two oral doses of 800 mg m⁻² dFUrd with a time interval of 4 h.

At the start of oral administration concentrations of dFUrd were below 100 ng ml⁻¹. The profiles of dFUrd and FUra after the first and second oral doses of dFUrd were similar (Figure 3). FUra concentrations reached the low ng ml⁻¹ level within 2 h of dFUrd administration. Mean elimination half-lives of FUra were 34±12 min (dose 600 mg m⁻²); 40±16 min (dose 800 mg m⁻²) and 38±13 min (dose 1000 mg m⁻²) (Table 1). The AUC of FUra increased with the dose of oral dFUrd: from 13.2±4.8 μg ml⁻¹ min at 600 mg m^−2–24.6± 9.2 μg ml⁻¹ min at 1000 mg m⁻².

The t_lag of FUraH₂ was between 30 and 90 min of oral administration of 600–1000 mg m⁻² dFUrd (Figure 3). C_max was 100–325 ng ml⁻¹ and t_max 50–180 min. Mean elimination half-lives were more than 1 h, and not prolonged with increasing doses of dFUrd (Table 1). The AUC of FUraH₂ increased substantially with the dose, from 12.1±3.6 μg ml⁻¹ min at 600 mg m⁻² oral dFUrd to 21.9±8.3 μg ml⁻¹ min at 1000 mg m⁻² (Table 1). Accumulation of FUraH₂ in the plasma following repeated dosing was not observed (Figure 3).

Urinary recovery

The urinary recovery of unchanged dFUrd increased sharply with the dose, especially after intravenous administration: following 600 mg m⁻², the mean urinary recovery was ≥2% (intravenous) and 7% (oral), increasing to 46% and 18%, respectively, with a dose of 1000 mg m⁻² (Table 2). Assuming that the formation of FUra from dFUrd is a one step process, recovery of FUra never exceeded 2% of the dose of dFUrd; the urinary excretion of FUraH₂ was negligible.

Table 2.

Mean urinary recovery (%) of dFUrd and FUra in cancer patients treated with different doses of dFUrd.

Discussion

The bioavailability of oral dFUrd capsules, not previously reported, was 34–47% (Table 1), indicating satisfactory absorption, unlike the case of FUra [15]. The variability in the AUC of oral dFUrd was low, and in this respect dFUrd is comparable with oral cyclophosphamide, a widely used anticancer agent [16]. Substantial protracted absorption of oral dFUrd occurs, since the elimination half-life of oral dFUrd is almost twice that of intravenous dFUrd.

Following oral administration, the concentration-time profiles of dFUrd, FUra and FUraH₂ are not compatible with the saturation of dFUrd clearance, as seen with high dose intravenous dFUrd, where convexity of these profiles frequently occurs [7], possibly as a result of saturation of the dihydropyrimidine dehydrogenase system which degrades FUra to FUraH₂. This is reflected by the nonlinear increase in AUC FUra with increasing intravenous dFUrd doses between 800 and 1000 mg m⁻² (Figure 2). When the metabolic clearance of FUra via FUraH₂ is reduced, there is increased urinary recovery of FUra [17]. The percentage urinary recovery of intravenous dFUrd increased sharply with the dose (Table 2), indicating a similar ‘rescue’ mechanism. These data, together with the positive correlation between the mean residence time of intravenous dFUrd and toxicity reported earlier [7], indicate that renal function is crucial to the use of dFUrd.

Oral dFUrd resulted in protracted levels of both the parent drug and FUra (Figure 3), and the profiles of continuously infused dFUrd can be produced by oral administration at appropriate intervals, preferably every 6–8 h. A total dose of 36 g m⁻² in 3 months was well tolerated. Although this was a phase I study, tumour growth was arrested or reversed in two of six patients. Previous studies indicate that higher doses are required for meaningful response rates in colorectal and ovarian cancers. In one phase II study in colorectal cancer, with mostly chemotherapy naive patients, dFUrd was given as an intravenous bolus injection of 4 g m⁻² day⁻¹ for 5 days, and repeated every 3–4 weeks [6]. Of 27 patients receiving two or more courses, seven (26%) responded. In Japan, a solution of dFUrd given orally at a dose of 800 mg m⁻² day⁻¹ continuously, or 2250 mg day⁻¹ 4 days each week, has been investigated [18]. The intermittent schedule was most tolerable, and some activity was demonstrated in colorectal cancer. The low incidence of toxicity following oral doxifluridine tablets and concomitant leucovorin was confirmed recently in a study of 108 patients [19]. Using doses of 25 mg leucovorin and 1200 mg m⁻² dFUrd days 1 to 5, every 10 days, the only toxicities of WHO grade 3 or 4 were diarrhoea in 32 patients, and nausea and vomiting in one patient.

Tumour growth can be influenced by numerous mechanisms, not only tumour cell kill, but also destruction of the microvascular endothelial cells abundantly present in colorectal cancers [20, 21]; prolonged oral dFUrd with sustained systemic levels should be advantageous, perhaps resulting in substantial inhibition of neo-angiogenesis in colorectal tumours. Moreover, platelet-derived endothelial cell growth factor (PD-ECGF) is a potent angiogenic factor [22], and has also been found to activate dFUrd [23].

The membrane protein P-gp 170, by expelling cytotoxic agents such as anthracyclines from the cell interior, is important in the development of resistance to anticancer drugs [24]. Colorectal cancer commonly overexpresses P-gp 170 and is associated with oncogenes such as ras and myc [24, 25]. dFUrd is more cytotoxic to NIH 3T3 or human cancer cells in which ras mutations exist [26–28]. The activity of dFUrd is not diminished by P-gp 170, nor by mutated ras and myc, all of which frequently occur in colorectal cancer; we have recently established that dFUrd acts as an inhibitor of P-gp 170 at concentrations achievable clinically [29].

In conclusion, the oral capsule formulation of dFUrd used in this study has stable and reproducible pharmacokinetics and bioavailability, and warrants further clinical investigation. In the dose range 600–1000 mg m⁻² there is no evidence of saturation of the metabolic or renal clearance of dFUrd or two of its metabolites, FUra and FUraH₂. When the total monthly dose is less than 36 g m⁻² side-effects are likely to be tolerable. Prolonged treatment with oral dFUrd is appropriate in advanced colorectal cancer.