Introduction

Top Abstract Introduction Methods Results Discussion References

1--D-Arabinofuranosylcytosine is an effective chemotherapeutic agent for the treatment of acute myelogenous leukemia (Gahrton, 1983; Keating et al., 1982; Plunkett and Gandhi, 1993). However, its usefulness is impaired by its rapid deamination to the biologically inactive metabolite ara-U (Ho and Frei, 1971). To increase the cytotoxic activity of ara-C numerous N⁴-derivatives were synthesized to protect the drug from deamination and to alter its pharmacokinetic properties (Kanai and Ichino, 1974; Rosowsky et al., 1982; Wempen et al., 1968). Whereas short-chain modifications of ara-C at the N⁴-amino group generally resulted in a weak enhancement of cytotoxicity (Aoshima et al., 1976), lipophilic derivatives with long-chain fatty acids had a strong antitumor activity in murine tumor models (Kataoka and Sakurai, 1980; Tsuruo et al., 1980). In a previous study, we reported that N⁴-acyl derivatives of ara-C, incorporated into the membranes of small unilamellar liposomes, were active against murine L1210 leukemia and B16 melanoma at lower concentrations than unmodified ara-C (Rubas et al., 1986). However, the protection against enzymatic deamination to ara-U was only partially achieved with the N⁴-acyl derivatives. Therefore we synthesized the N⁴-alkyl-ara-C derivatives NHAC and NOAC shown in figure 1 (Schwendener et al., 1995a; Schwendener and Schott, 1992). These derivatives were found to be extremely resistant toward deamination in plasma as well as in the liver (Schwendener et al., 1995b). Because of the very low solubility in aqueous media NHAC and NOAC were incorporated into the lipid membranes of small unilamellar liposomes to allow their parenteral application. Liposomal preparations of NHAC and NOAC exerted significantly higher cytotoxic activities in the L1210 leukemia model than ara-C. In contrast to the parent drug and to some other lipophilic ara-C derivatives, NHAC and NOAC were found to exert excellent antitumor effects after oral therapy (Schwendener et al., 1996; Schwendener and Schott, 1996). We conclude from these and the following findings that the mechanisms of action of the N⁴-alkyl-ara-C derivatives are different from ara-C. Their cellular uptake is nucleoside-transporter independent and only 2 to 5% were phosphorylated to ara-C triphosphate in HL-60, K-562 and U-937 cells (Horber et al., 1995b, c). The cytotoxicity of NHAC was found to be less S-phase specific than ara-C, and induction of apoptosis occurred only at very high drug concentrations (Horber et al., 1995d). Presently, we concentrate our studies on NOAC because this derivative has the highest antitumor activity of all N⁴-alkyl analogs (Schwendener et al., 1995a). A phase I/II study of liposomal NOAC is currently underway at the University Hospital Zurich.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Chemical structures of ara-C, NOAC and NHAC.

It is known that single-chain acyl compounds like fatty acids are not tightly anchored within the lipid bilayer of liposomes and that they are readily transferred to plasma proteins and Ec membranes (Kamp et al., 1993; Kleinfeld and Storch, 1993; Richieri et al., 1993). We assumed that the N⁴-alkyl-ara-C derivatives, which have no amphiphilic properties and are not charged at physiological pH, move through lipid membranes of the liposomes and are transferred to blood components at rates that are comparable with long-chain fatty acids.

Therefore, in this report we investigated the interactions of NOAC with human blood in vitro and compared them with the properties of NHAC. Ec binding was measured and protein binding was calculated with the Ec partition coefficient D_Ec. In addition, we studied the distribution of the drugs between the serum lipoproteins which were separated by ultracentrifugation on a KBr density gradient and analyzed on agarose gels. In the second part of this contribution we determined the blood and organ distribution in vivo of liposomal NOAC in ICR mice and calculated the pharmacokinetic parameters.

	Methods

Top Abstract Introduction Methods Results Discussion References

Chemicals. NHAC and NOAC were synthesized as described previously (fig. 1; Schwendener et al., 1995a; Schwendener and Schott, 1992). SPC was obtained from L. Meyer, Hamburg, Germany. Cholesterol (Fluka AG, Buchs, Switzerland) was recrystallized from methanol. DL--Tocopherol and all analytical grade buffer salts and other chemicals used were from Merck, Darmstadt, Germany or Fluka, Buchs, Switzerland. NHAC and NOAC were tritium labeled (0.189 Gbq/mmol [5-³H]NHAC and 0.370 Gbq/mmol [5-³H]NOAC) by Amersham Int., Amersham, UK. Soluene 350 and Ultima Gold scintillation cocktail were from Packard Instruments, Groningen, The Netherlands.

Preparation of liposomes. Small unilamellar liposomes were prepared by sequential filter extrusion of multilamellar liposomal preparations through Nuclepore membranes (Sterico, Dietikon, Switzerland) of 0.4 µm, 0.2 µm and 0.1 µm pore diameter with a Lipex extruder (Lipex Biomembranes Inc., Vancouver, Canada; Hope et al., 1985). For experiments with human blood or intravenous injection into mice, liposomes were either prepared in PB (67 mM, pH 7.4) or in saline, containing 0.01% EDTA (saline/EDTA) for the incubations with serum. Liposome size and homogeneity were determined by laser light scattering (Submicron Particle Sizer Model 370, Nicomp, Santa Barbara, CA). The basic lipid composition of the liposomes used for in vitro incubations with blood and serum was 40 mg/ml SPC, 4 mg/ml cholesterol, 0.2 mg/ml DL--tocopherol and 2.5 mg/ml of the drugs NOAC or NHAC, whereas the liposomes used for the pharmacokinetic experiments were composed of 80 mg/ml SPC, 8 mg/ml cholesterol, 0.4 mg/ml DL--tocopherol and 11 mg/ml NOAC. All preparations were trace labeled with [5-³H]NOAC or [5-³H]NHAC, respectively, sterile filtrated (0.2 µm, Schleicher & Schuell, Dassel, Germany), stored at 4°C and used within 3 days after preparation. For control experiments stock solutions of NHAC and NOAC (2.5 mg/ml, corresponding to their highest stable solubility) in DMSO were prepared.

In vitro distribution of NOAC and NHAC in human blood. Fresh venous blood was collected in EDTA tubes (Vacutainer, Becton Dickinson, Meylon Cedex, France) from a healthy donor after an overnight fast. For the incubation experiments, 1.5 ml blood containing 6 ± 1 × 10⁹ Ec (hematocrit, 0.3-0.4) were spiked either with the drugs in liposomes or in DMSO to yield final drug concentrations of 60 to 1100 µM (liposomes) or 100 and 200 µM (DMSO) in a total volume of 2 ml. After incubation on a blood sample shaker (Speci-mix, Bioblock, Frenkendorf, Switzerland) for 4 h at 37°C the blood samples were centrifuged (10 min, 650 × g, 20°C) and plasma was removed. The blood cells were not separated further, because as shown before with NHAC, binding to leukocytes was negligible (approximately 2%; Horber et al., 1995a). Thus, the whole-blood cell fraction was referred to as Ec. The Ec samples were washed three times with PB, centrifuged (10 min, 650 × g, 20°C) and then solubilized, and the [5-³H]NOAC or [5-³H]NHAC activity was detected by scintillation counting as described previously (Horber et al., 1995a). All experiments were carried out in triplicate. To exclude the possibility of precipitation of NOAC dissolved in DMSO in these incubations, the following control experiments in diluted serum were performed. The ³H-labeled drug dissolved in 50 µl DMSO (0.28 mM final concentration) was incubated in 450 µl serum diluted with 500 µl saline/EDTA for 4 h at 37°C. The serum was centrifuged (650 × g, 10 min) and the [5-³H]NOAC concentration was determined in the supernatants. Corresponding controls were made by use of saline/EDTA instead of serum.

Analysis of binding parameters. To fit the concentration-dependent binding curve of NOAC and NHAC to Ec we used a one-site binding model:

(1)

with B_max, maximal binding capacity, and K_d, concentration of the ligand to determine half-maximal binding. The binding parameters of the two drugs to Ec were calculated from linearized curves of the binding rate r given in equation 2 versus unbound drug c_u.

(2)

with [AEc], drug bound to Ec in millimolar (corresponding to 6.022 × 10²⁰ Ec per liter) and [Ec_tot], total concentration of Ec in millimolar. The Ec partition coefficient D_Ec was calculated according to equation 3:

(3)

with A_blood, the absolute amount of drug in whole blood (nanomoles); A_plasma, drug in the plasma fraction (nanomoles) and H, hematocrit (Derendorf and Garrett, 1983). The plasma protein-binding fraction f_b was determined from D_Ec as described in equation 4:

(4)

with c_plasma, concentration in the plasma fraction of the probe, and V_blood, the volume of whole blood.

In vitro distribution of NOAC and NHAC in human serum. Fresh serum with normal cholesterol and lipoprotein levels was obtained from a healthy donor after overnight fasting. The blood was allowed to coagulate for 2 h and then centrifuged (10 min, 650 × g, 20°C). The drugs (0.1 ml; final concentration, 200 µM) were added either as liposomal preparation or dissolved in DMSO to 0.9 ml serum corresponding to the serum contained in 1.5 ml whole blood. After incubation for 4 h at 37°C on a blood sample shaker the separation of the plasma proteins was performed by gradient ultracentrifugation according to Chapman et al. (1981) and Redgrave et al. (1975) with small modifications. A density gradient was formed in ultracentrifuge tubes (Polyallomer, 5 ml, Beckman, Geneva, Switzerland) by adding 0.325 g KBr to 1 ml of the serum supernatants (d₂₀²⁰ 1.21), which were overlaid sequentially with solutions of 1 ml KBr in saline of the decreasing densities d₂₀²⁰ 1.064, 1.020 and 1.007. Finally, 1 ml of distilled water was added on top. The probes containing the drugs in liposomes or in DMSO and the appropriate controls replacing the serum with saline/EDTA and serum alone (fig. 3) were placed in a SW 50.1 rotor (Beckman, Geneva, Switzerland) and centrifuged for 22 h at 300,000 × g and 15°C in an ultracentrifuge (Centrikon T-1065, Kontron Instruments, Zürich, Switzerland). After careful removal of the tubes from the rotor a fine glass capillary was gently immersed to the bottom of each tube and 50 fractions of 0.17 ml were collected with a fraction collector (Superrac LKB, Uppsala, Sweden). Protein absorption was monitored continuously with a flow detector (Uvicord SII, LKB) at 279 nm. Drug concentration per fraction and initial concentration of the probes were analyzed by scintillation counting. The densities of the KBr solutions and of the individual fractions after centrifugation were determined with a densitometer (DMA 38, Anton Paar KG, Graz, Austria). All incubations were carried out in triplicate. The separation of the lipoproteins was controlled by electrophoresis on 1% agarose gels by a modified method described by Nobel (1968). From each centrifuged serum fraction an aliquot prestained with Sudan black B was applied to the gels. To provide further evidence of the transfer of liposomal NOAC to the lipoproteins, selected HDL- and LDL-rich fractions from the KBr gradient separation (cf. fig. 3) were run on a gel (cf. inset, fig. 4), and the bands corresponding to HDL or LDL were cut and analyzed for [5-³H]NOAC activity (cf. curves in fig. 4).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Distribution of NOAC (250 µM) in liposomes () and DMSO () after incubation with human serum (4 h, 37°C) and gradient ultracentrifugation over a KBr gradient (d2020 1.21, 1.064, 1.020, 1.007, 1.000; centrifuged 22 h, 300,000 × g, 15°C), collection of fractions (0.17 ml) and scintillation counting for 5-3H-labeled drug activity. Control serum (full line) was treated as described above and the proteins were monitored by UV-spectroscopy (279nm). Fractions were pooled for albumin (fractions 4-7), HDL (fractions 11-16), LDL (fractions 19-25) and VLDL (fractions 29-50), and the concentrations of bound drug were calculated. Additionally, the equilibrated density gradient (dashed line) given in d2020 after centrifugation was superimposed. Data are means of triplicates. A comparable distribution was obtained with NHAC (not shown). Alb., albumin.

View larger version (48K): [in this window] [in a new window]   Fig. 4.   Binding of NOAC to HDL () and LDL (). Representative fractions containing the lipoproteins from figure 3 were prestained with Sudan black B and run on a 1% agarose gel (inset). The [5-3H]NOAC concentration was determined by cutting out the corresponding bands (graph). S, diluted serum (1:20, v/v).

NOAC pharmacokinetics and organ distribution in mice. Female mice (ICR, 27-38 g) were injected intravenously in the tail vein with 2.3 mg (4.6 µmol) liposomal NOAC in a volume of 200 µl. After periods ranging from 4 min to 24 h groups of three mice were sacrificed and blood, liver, spleen, kidneys, lung and brain were collected. To determine the concentration of NOAC in blood and the organs the samples were solubilized and further treated as described previously (Horber et al., 1995a). Drug concentrations in organs were corrected for the remaining blood (Allen, 1989). All values for blood and organ distribution (figs. 5 and 6) were standardized with the body weight of a mouse of 20 g and given as percent of disintegrations per min injected per ml blood or as percent of disintegrations per min injected per total organ.

View larger version (11K): [in this window] [in a new window]   Fig. 5.   Blood concentration (A), liver (B) and kidney (C) distribution versus time curves of NOAC (2.3 mg/animal) after i.v. application in female ICR mice. Blood samples and organs were removed and prepared for scintillation counting of [5-3H]NOAC as described under "Methods." All values were standardized to a mouse body weight of 20 g and shown as percent of injected dose. Data of the blood curve (A) were fitted with the i.v. open two-compartment model, for which parameters were determined by the residual method. (Parameters: A, 9.2%/ml; B, 1.2%/ml; , 0.0297/min; , 0.0017/min; r, 0.934). Data for the organ curves (B, C) were fitted for the elimination phase for the liver (B, 16.9%; , 0.0015/min; r, 0.968) and the kidneys (B, 2.0%; , 0.00074/min; r, 0.995), respectively.

View larger version (37K): [in this window] [in a new window]   Fig. 6.   Bar graph of spleen, lung and brain distribution of NOAC in female ICR mice. Data calculated as percent of injected dose were obtained after i.v. application of NOAC (2.3 mg) into the tail vein and the sacrifice of three mice at given time points. The organs were removed and prepared for counting [5-3H]NOAC activity as described under "Methods."

Pharmacokinetic analysis. The blood concentration versus time curve (fig. 5A) was approximated by the residual method which led to equation 5 of a two-compartment open model for intravenous drug application. The calculated parameters A, B,  and  of equation 5 were used to fit the curve:

(5)

where (A + B), peak drug concentration in blood at time t = 0 min. The distribution t_1/2 and elimination t_1/2 half-lives were calculated from the slopes  and . The apparent volume of distribution V_d(area), the apparent volume of the central V_c and peripheral V_p compartments, the systemic clearance Cl_total were calculated by equations described elsewhere (Greenblatt and Koch-Weser, 1975). The area under the curve for time zero to infinity AUC_(tr. 0) was calculated model-independently with the trapezoidal rule to the last measured time point t = 1440 min and extrapolation to infinity (c_{blood(1440 min)} divided by ). To compare the organ load of the different organs, the percentage of the organ AUC_(tr. 0) was compared with the sum of all organ AUC_(tr. 0) values. The organ load is a measure of how much drug ever appears in one organ during the whole period between application and total elimination of the drug. The AUMC was calculated with equation 6.

(6)

with the drug concentration in blood (c_blood). The MRT was calculated by dividing the AUMC with the AUC_(tr. 0).

(7)

(0t₁

(8)

	Results

Top Abstract Introduction Methods Results Discussion References

Concentration-dependent binding of NOAC and NHAC to Ec. The binding curves of liposomal NOAC and NHAC to Ec are shown in figure 2. The data were fitted with a one-site binding model described by equation 1, which results in saturation values of 63 nmol/10⁹ Ec for NOAC and 88 nmol/10⁹ Ec for NHAC, respectively. Saturation is reached by use of at least 0.7 mM NOAC or 0.9 mM NHAC under the experimental conditions. A significant difference in the binding to Ec between NOAC and NHAC was found only above saturation (P = .033, Student's t-test, fig. 2). Linearization of r versus c_u according to Lineweaver-Burk (correlation coefficient r = 0.963 for NOAC and 0.992 for NHAC) or Scatchard plots (r = 0.869 for NOAC and 0.987 for NHAC) resulted in 4 × 10⁷ binding sites per Ec for NOAC and 5 × 10⁷ for NHAC with a low binding affinity of 3 × 10³ liters/mol for both drugs. The Ec partition coefficient D_Ec calculated after equation 3 was determined with the drugs (0.1 and 0.2 mM) dissolved in DMSO to avoid possible interference of the liposomal lipids with the binding of the drugs to Ec membranes. The altered drug formulation did not affect the binding characteristics of the drugs to Ec (fig. 2). For NOAC at 0.1 and 0.2 mM, a D_Ec of 4.2 ± 0.4 resulted, whereas for NHAC the D_Ec was 3.0 ± 0.4 (significant difference, P = .0004). According to equation 4, the fraction of drug bound to plasma proteins f_b, which was calculated with the D_Ec values, was 32 ± 3% for NOAC and 35 ± 5% for NHAC, which were not significantly different (P = .236).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Binding curves of NOAC () and NHAC () in liposomes (0.06-1.1 mM) to Ec determined as described under "Methods." For a comparison, the binding of NOAC () and NHAC () dissolved in DMSO (0.1 and 0.2 mM) was determined under the same conditions. Drug concentration in Ec was determined by scintillation counting. Only the values at 1.1 mM were significantly different (*P = .033; Student's t-test). The fit with a one-site binding model (equation 1) resulted in saturation values of 63 nmol/109 Ec for NOAC (correlation coefficient, r = 0.930) and 88 nmol/109 Ec for NHAC (r = 0.997). Each point represents the mean ± S.D. of three measurements. Invisible error bars are smaller than the symbols.

Binding of NOAC to serum proteins. To further characterize the binding properties of NOAC to serum proteins, the drug was incubated (4 h, 37°C) with fresh human serum, followed by separation of the proteins on a KBr density gradient. To monitor the gradient after ultracentrifugation a control run with saline/EDTA was performed and the densities of the fractions were determined. As shown in figure 3 the discontinuous gradient was smoothed after centrifugation. The separation of the lipoproteins was confirmed by agarose gel electrophoresis of all collected serum fractions (data not shown). Lipoprotein-containing fractions were pooled according to their densities and agarose gel patterns (cf. table 1). The centrifuge tubes showed the typical density bands where the corresponding lipoproteins were expected (Chapman et al., 1981; Redgrave et al., 1975). Additionally, a white band with liposomes was detected in the density range of d₂₀²⁰ 1.063 to 1.019, close above the LDL band. The diameters of the liposomes used ranged from 30 to 80 nm and their density was similar to that of LDL. Therefore, it was not possible to separate them clearly from LDL. Control centrifugations containing NOAC in DMSO and with substitution of serum by saline/EDTA resulted in a different distribution pattern with most of the drug accumulated at d₂₀²⁰ 1.1034, corresponding to the density of the solvent DMSO. This formulation was used to document the binding properties of the drug to LDL without interference of the liposomal carriers because only drug bound to LDL appeared in fractions 19 to 25. This experiment resulted in a similar distribution pattern as obtained with liposomal NOAC (fig. 3). With NHAC comparable patterns of distribution were found for all incubations (data not shown).

Pharmacokinetics of NOAC in mice. The NOAC kinetic values in female ICR mice were determined after intravenous injection of NOAC and the sacrifice of groups of three mice at adequate time points. The blood concentration versus time curve is biphasic, yielding a distribution half-life t_1/2 of 23 min and an elimination half-life t_1/2 of 7 h (fig. 5A). As comparison, the elimination half-life t_1/2 of ara-C was found to be 21 min after i.p. administration in mice (Borsa et al., 1969). Thus, the elimination of NOAC is 20 times slower than that of ara-C. For the N⁴-hexadecyl-ara-C derivative NHAC a similar distribution half-life t_1/2 of 16 min, but a faster elimination half-live t_1/2 of 3.8 h, was determined (Horber et al., 1995a). The total amount of NOAC found in blood AUC_(tr. 0) was 11% dose × h/ml and the total clearance Cl_total was rather low with 6 ml/h. This resulted in an apparent total volume of distribution V_d(area) of 58 ml. The MRT was 6 h. The size of the two compartments was 10 ml for the central compartment, V_c, and 33 ml for the peripheral compartment, V_p. Thus, shortly after intravenous bolus injection NOAC is distributed to other compartments. Distribution into deeper compartments occurs because V_d(area) is 3-fold and V_p is 1.7-fold larger than the volume of total body fluids (19 ml for mice of 25 g; Allen et al., 1992). Relative peak drug concentrations (table 2, figs. 5 and 6) were reached shortly after i.v. injection in blood and lung, whereas in liver and spleen they were reached after 30 min. Correspondingly, the peak concentration in the kidneys was found after 3 h. Most of the drug appears in the liver with a high organ load of 69% which can be expected for lipophilic drugs (table 2). NOAC was eliminated from the liver with a half-life of 8 h. Thus, after a single dose of NOAC more than 99% of the drug is removed from the liver within 56 h (7 times t_1/2 liver). Pharmacokinetic parameters for the brain could not be calculated because of the very low drug concentrations that were found in this organ. Distribution into the brain is quite low with less than 0.1% of the applied dose.

View this table: [in this window] [in a new window]   TABLE 3 Calculation of the organ parameters for NOAC according to Nishikawa et al. (1993) and comparison to ara-C

	Discussion

Top Abstract Introduction Methods Results Discussion References

Binding of NOAC and NHAC to Ec. The high values of the Ec partition coefficient D_Ec of 3.0 and 4.2 determined for NHAC and NOAC reflect the lipophilicity of the drugs. The significantly higher affinity of NOAC for Ec might be caused by the longer alkyl chain. For comparison, Ueda and co-workers (1983) found a similarly high accumulation of N⁴-behenoyl-1--D-arabinofuranosylcytosine in the blood cell membranes, which was significantly higher than binding to plasma proteins. This drug is a comparable lipophilic N⁴-acyl-ara-C derivative. The weak binding affinity of NOAC and NHAC for Ec correlates with findings of single-chain acyl compounds like fatty acids that are not tightly anchored within lipid bilayers (Richieri et al., 1993; Kleinfeld and Storch, 1993; Kamp et al., 1993). In an analytical HPLC study, we observed significantly higher concentrations of the drug bound to Ec, reaching a maximal Ec-to-plasma ratio of 7:1 after 6 to 8 h, after the oral application of NOAC to mice (Rentsch et al., 1995). These results are in concordance with the calculated protein binding value of 32% from the in vitro incubations. By extrapolation of the calculated Ec binding parameters to the average Ec concentration in a healthy human (blood volume of 4.2 liters containing 2 × 10¹³ Ec) and under the simplifying assumption that the Ec are the only binding partners of the drugs, saturation would be reached with a dose of 1.8 g NOAC or 2.4 g NHAC, respectively. With these drug amounts a maximal Ec binding of 0.6 g NOAC and 0.8 g NHAC would be achieved. Thus, we postulate that in vivo, after intravenous application of NOAC, a dynamic equilibrium between the blood components is established, favoring the initial distribution of the drug into the Ec membranes which is followed by a redistribution into the plasma proteins and preferentially into the lipoproteins. The binding of NOAC to other cells is negligible, because as we showed before with the distribution of NHAC in whole blood, only about 2% were bound to leukocytes (Horber et al., 1995a).

Binding of NOAC and NHAC to serum proteins. The binding characteristics of NOAC and NHAC to the lipoproteins and mainly to LDL suggests that LDL might function as a natural carrier for these drugs. One reason for the excellent therapeutic activity of NOAC against solid tumors (Schwendener et al., 1995a), which is not observed with ara-C, might be explained by its high affinity to LDL (fig. 3, table 1) and its postulated carrier effect on tumor cells expressing high numbers of LDL receptors. Other investigators (de Smidt et al., 1993; de Smidt and van Berkel, 1990) exploited this LDL carrier effect for selective drug targeting to tumor cells with high-LDL receptor expression. Vitols et al. (1990) observed an increased uptake with radioactively labeled sucrose-LDL in leukocytes of leukemic patients, which suggested a correlation to the increased LDL receptor activity of leukemic compared with normal leukocytes. Finally, lipophilic photosensitizer dyes were found to be associated with lipoproteins and transported mainly by LDL to malignant cells (Ginevra et al., 1990; Reddi et al., 1990; Rensen et al., 1994; Schmidt-Erfurth et al., 1997). Therefore, LDL might be an efficient carrier for various drugs to treat leukemias and solid tumors with increased LDL turnover, e.g., metastatic cancer of the prostate as described by Vitols et al. (1990). Firestone (1994) reported that there are several types of carcinoma cells with an increased LDL uptake, especially tumor cells that have an exceptionally high metastatic potential, or that are aggressive or undifferentiated. The rationale is that large amounts of LDL are taken up by rapidly dividing cells because an increased amount of cholesterol is required for cell membrane assembly. However, the use of LDL as a carrier for lipophilic drugs is limited by complicated procedures required for its isolation from human serum and its modification as a drug transporter molecule (Firestone, 1994; Gerke et al., 1996). These and other disadvantages, such as the ensuing risk of infection of LDL isolated from human blood, render the clinical use of LDL as a drug carrier for cancer therapies rather unlikely. We found that after incubation in serum 30 to 36% of NOAC molecules were bound to LDL. The serum volume of 1 ml contains approximately 5.5 × 10¹⁴ LDL particles as determined and calculated by quantitative agarose gel electrophoresis (Hydragel, Sebia, Issy-les-Moulineaux, France; data not shown). Thus, we calculated that one LDL particle is able to bind about 100 molecules of NOAC.

Pharmacokinetics of NOAC in mice. Liposomes are generally used to improve the blood pharmacokinetics of encapsulated hydrophilic drugs like ara-C (Allen et al., 1992) or doxorubicin (Vaage et al., 1994) that have short plasma half-lives when administered in their free form. Ara-C encapsulated within long-circulating liposomes remains in the circulation because Allen et al. (1992) determined a distribution volume V_d of 2.2 ml (= blood volume). Compared with free ara-C which is distributed in the total body fluid (V_d = 19 ml), the distribution volume of NOAC V_d(area) of 58 ml is about three times larger, which indicates the distribution into deeper compartments that might be caused by the high lipophilicity of NOAC.

1/2

1/2