Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The importance of oncogenes in the etiology of cancer has been shown through studies of the molecular mechanisms underlying cancer development (Stein and Cohen, 1988). For example, bcl-2 abnormalities have been shown in a number of tumor types, among which are a large number of lymphoid tumors. In follicular lymphoma, it has been established that >85% of human B cells show the t(14;18) chromosome translocation, which juxtaposes the bcl-2 gene next to the immunoglobulin heavy chain joining region (Bakhshi et al., 1985; Cleary et al., 1986; Weiss et al., 1987; Cotter, 1993). This translocation results in overexpression of bcl-2 protein, which induces an inhibition of the apoptotic pathway (Hockenbery et al., 1990). Antisense oligodeoxynucleotides with sequence complementarity to a specific target mRNA have been developed to inhibit gene expression (Paoletti, 1988; Campbell et al., 1990; Tidd, 1990; Wagner,1994). Chemical analog oligonucleotides, phosphorothioates and methylphosphonates, have been generated to avoid the rapid nuclease cleavage observed with phosphodiesters (Zon, 1988; Sands et al., 1994). Phosphorothioates have been shown to retain the ability to activate RNase H cleavage of the target, which improves their antitumor effect, although high concentrations inhibit RNase H (Gao et al., 1992). Recent studies question the occurrence of an antisense effect, because nonspecific phosphorothioate sequences have been shown to have antitumor activity (Krieg and Stein, 1995; Chrisey et al., 1995). Cellular studies have shown that bcl-2 expression can be inhibited by antisense oligonucleotides and that phosphodiesters are 10-fold less potent than phosphorothioates (Reed et al., 1990). G3139 is an 18-mer full phosphorothioate antisense oligodeoxynucleotide targeted to the first six codons of the open reading frame of bcl-2. Antisense oligonucleotides to this sequence have been shown to have antitumor efficacy against the human lymphoma DoHH2 treated in vitro and then implanted into severe combined immunodeficient mice (Pocock et al., 1993; Cotter et al., 1994). Further experiments showed that an in vivo s.c. infusion of 5 mg/kg/day G3139 over a 3-week period was the optimal schedule for the eradication of lymphoma associated with bcl-2 down-regulation (Cotter et al., 1996). The present study evaluates the pharmacokinetic behavior of G3139 administered by continuous s.c. infusion. In addition, a single i.v. dose has been used to evaluate the pharmacological parameters and compare these results with other reports on phosphorothioate oligonucleotides.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Oligonucleotide Synthesis

Oligonucleotide phosphorothioate G3139, with the sequence 5-TCTCCCAGCGTGCGCCAT-3, was prepared on a 15-µmol scale on a controlled-pore glass solid support, using a Biosearch 8900 Expedite DNA synthesizer, by standard protocols (Stec et al., 1984; Iyer et al., 1990). The oligonucleotide was deprotected by ammonium hydroxide at 60°C for 4 hr. The product was redissolved in water, purified by HPLC on a Source-Q strong ion-exchange column (Pharmacia Biotech, Piscataway, NJ) and desalted. The product was analyzed by HPLC, mass spectroscopy and polyacrylamide gel electrophoresis, indicating 96% purity. The starting material for the ³⁵S-labeling of G3139 was a 16-mer oligonucleotide (n2) made by a similar procedure; the last two bases at the 5-end were attached by manual coupling. After the coupling of the 17th base, it was sulfurized with ³⁵S using elemental sulfur in toluene. The final coupling was sulfurized by nonradioactive Beaucage reagent. The oligonucleotide was then deprotected and purified by a procedure similar to that for the nonlabeled oligonucleotide. The ³⁵S-labeled G3139 was analyzed by HPLC and polyacrylamide gel electrophoresis and was 96% pure. The specific activity of the radiolabel was 1 Ci/mmol. Phosphorothioate oligonucleotides of the sequences 5-TCTCCCAGCGTGCGCCA-3 and 5-TCTCCCAGCGTGCGCC-3 were obtained from Genosys Inc. (Cambridge, UK).

Preparation of Solutions

G3139 preparations, spiked with radiolabeled G3139, were prepared in sterile PBS at 500 µg/9.6 µCi/ml for i.v. injections and 8.8 mg/166.7 µCi/ml for s.c. infusions. Oligonucleotide concentrations were quantitated by optical density measurements at 260 nm (1 optical density unit = 33 µg) of diluted preparations. Alzet micro-osmotic pumps (model 107D; mean ± S.D. pumping rate, 0.52 ± 0.02 µl/hr; mean ± S.D. fill volume, 96 ± 2.4 µl) were filled and primed with PBS for 4 hr at 37°C before implantation.

Animals and Treatment Schedules

Female BALB/c mice (6 weeks of age) were obtained from the Medical Research Council (UK) and were acclimatized to the laboratory conditions for 2 weeks before the experiment. They were allowed food (SDS expanded rodent diet from Special Diet Services, UK) and water ad libitum. The animals weighed 20 ± 1.2 g at the time of treatment. Doses of approximately 5 mg/kg were selected for this study to reflect the antitumor data (Cotter et al., 1996). Because of the limited availability of ³⁵S -labeled G3139, it was not possible to carry out a dose-response study, and continuous s.c. infusions were terminated after 7 days.

For group 1, animals were injected in the tail vein, after transient hyperthermia to induce vasodilation, with a single dose of 100 µg (approximately 5 mg/kg) of oligonucleotide in 0.2 ml of PBS (1.92 µCi of ³⁵S/animal). Blood and tissues were collected 5, 15 and 30 min and 1, 2, 4, 6, 24, 48 and 72 hr after administration (n = 5 animals/time point).

Anesthesia was induced and maintained with 6% halothane in oxygen at a flow rate of 3 liters/min. Blood was collected in heparinized syringes after severing of the axillic vessels. Blood was centrifuged for 10 min at 1500 × g, and the plasma was decanted and frozen at 70°C until analysis. To evaluate the plasma protein binding, an aliquot of the plasma was ultrafiltered through Amicon molecular weight 10,000 exclusion membranes, by centrifugation at 600 × g for 45 min.

Tissues (liver, kidney, spleen, heart, lung, brain and bone marrow) were collected as quickly as possible after cervical dislocation of the animals and were snap-frozen in liquid nitrogen. For harvesting bone marrow, both femurs were removed and flushed with 500 µl of PBS before snap-freezing of the preparation.

For group 2, primed micro-osmotic pumps were implanted dorsally in anesthetized animals to deliver 106 µg/day G3139 (approximately 5 mg/kg/day, 2 µCi of ³⁵S-G3139/day) for 7 days, after which the pumps were removed. Anesthesia was induced and maintained during these procedures as described above. Blood and tissues were collected 24, 48 and 72 hr and 7, 9, 14 and 21 days after administration (n = 3 animals/time point). Samples were processed as described above.

Three animals from each of the above groups were placed in metabolic cages for 3 days. Feces, urine and washes were collected and frozen at 70°C.

Analytical Methods

Direct counting. Plasma (50 µl), plasma ultrafiltrate (100 µl) or urine (500 µl) was added to 10 ml of Ultima Gold scintillation fluid (Packard, Berkshire, UK) and counted for 5 min in a Packard 2000 scintillation counter. Tissues and bone marrow preparations were digested with 1 ml of Soluene/100 mg of tissue (Sigma Chemical Co., Dorset, UK) for 24 hr at 37°C, and 10 ml of Hionic Fluor (Packard) were added to 1 ml of digest before counting. All of these procedures were validated by spiking nonradiolabeled tissues with 2,000, 5,000 and 20,000 dpm in triplicate, which gave recoveries of >95%. Feces were weighed, mixed with 1 ml of water/100 mg of feces and digested in an equal volume of Soluene at 60°C for 24 hr. One hundred microliters were decolorized by dropwise addition of hydrogen peroxide before addition of scintillant and radioactive counting. Recoveries were evaluated with control feces from untreated mice.

Protein measurements. The protein content of the bone marrow preparations was measured by the method of Lowry et al. (1951).

Oligonucleotide extraction. Plasma (0.5 ml) was added to 2.45 ml of 0.4% sodium dodecyl sulfate, 50 mM NaCl, 10 mM EDTA, 10 mM Tris, pH 7.4, and vortex-mixed for 2 min. Tissues were homogenized in PBS (10 ml/g) using a Potter-Elvehjem homogenizer. Homogenate (0.5 ml) was added to 0.6 ml of buffer (0.8% sodium dodecyl sulfate, 40 mM Tris, pH 7.4, 85 mM NaCl, 8 mM EDTA). Then 50 µl of 20 mg/ml proteinase K were added to plasma and tissue homogenates, and the mixtures were vortex-mixed for 2 min and incubated for 2.5 hr at 65°C. Two milliliters of water were added to the tissue homogenates, both plasma and tissue preparations were extracted three times with 0.6 ml of phenol reagent (Kirby, 1965) and the organic phases were back-extracted with 0.4 ml of PBS. The resultant aqueous phases were then extracted with 0.8 ml of isobutanol, followed by 0.5 ml of diethyl ether.

HPLC conditions. The HPLC system consisted of two Kontron pumps, a gradient-former 460 and a Kontron autosampler. UV detection was performed at 254 nm with a Unicam diode-array detector. The HPLC column was a Waters Gen Pak Fax column (4.6 × 100 mm), buffer A was 20% acetonitrile/10 mM LiOH and buffer B was 20% acetonitrile/10 mM LiOH/2 M LiCl. A linear gradient was run from 10 to 100% buffer B over 30 min, with a flow rate of 0.5 ml/min. Eighty microliters of sample were injected into the autosampler. Fractions (0.5 ml) were collected with a Packard 1122 fraction collector, and 5 ml of Hionic Fluor scintillant were added to the samples, which were counted for 5 min. The resolution of the method was two bases.

Calculations

Radioactivity calculations. Concentrations are presented in rad equivalents, which represent the amount of parent compound at the specific activity administered that would result in the observed dpm values. Both the efficiency of the counting and the decay of the radiolabel have been accounted for.

Pharmacokinetic calculations. Pharmacokinetic parameters were calculated with PCNONLIN software (Lexington, KY), with compartmental analysis. Functions consisting of the sum of one, two or three exponential components were fitted to data by a least-squares method. Each set of data was analyzed with one, two or three compartments and the best fit was adopted. For example, it was shown that the best fit for the plasma i.v. bolus concentration vs. time curve was a three-compartment model (model 18), whereas the s.c. study was best fit to a one-compartment infusion model (model 2). The tissue to plasma ratios were calculated using the relative area under the curve calculated to the last point with the trapezoidal method.

Statistical tests. Results were expressed as means ± S.E. The differences between groups were assessed by analysis of variance.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Chromatography. Figure 1 shows a chromatographic profile of G3139, which eluted at 17.6 min. The 17-mer phosphorothioate oligodeoxynucleotide was not resolved from G3139, whereas the 16-mer eluted at 16.7 min, distinct from G3139 (data not shown).

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Chromatography of G3139 (80 µl of 16.2 µg/ml). The separation was achieved on a Gen Pak Fax 15-cm column using a gradient of LiCl in LiOH. UV detection was performed at 260 nm, with 1-V full-scale output.

Bolus i.v. administration. The radioactivity measurements in plasma and various tissues are shown in figure 2. For clarity, error bars have been omitted from figure 2 but full results (mean ± S.E.) are shown in table 1.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Time course of radioactivity in plasma () (in rad equivalents/ml), liver (), spleen (), kidney (), heart plus lung () and gut () (in rad equivalent/g) and bone marrow () (in rad equivalents/g protein) after 100-µg 35S-G3139 i.v. bolus administration to BALB/c mice.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Time course of G3139 in plasma (), kidney () and liver () after i.v. bolus administration of 100 µg of G3139.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Plasma radiochromatograms after i.v. bolus administration of 100 µg of 35S-G3139, at 15 min, 6 hr and 24 hr after administration. Samples were separated by ion-exchange chromatography on a Waters Gen Pak Fax column and eluted with a gradient of LiCl in LiOH with 20% acetonitrile. Fractions were collected every 0.5 min and counted.

Continuous infusion. The radioactivity measurements in plasma and tissues after continuous s.c. infusion for 7 days are shown in figure 5. Again, for clarity, error bars have been omitted from figure 5 but full results (mean ± S.E.) are shown in table 5. Plasma levels increased steadily and reached steady state by day 3 (table 5). By day 14 most of the radioactivity had disappeared from the plasma. The tissue to plasma ratios were 99 for kidney, 30 for liver, 6.5 for spleen, 1 for heart, 1.7 for lung and 8.5 for gut. The radioactivity was shown to accumulate in the organs, with a significant increase between day 3 and day 7 (P < .01). At steady state, half of the plasma radioactivity was protein bound and 66 to 86% was G3139 (table 5). The levels of G3139 at steady state were 0.16 µM (fig. 6), and the half-life of elimination after s.c. administration was 22 hr (table 3). G3139 could not be detected in the plasma after 14 days (table 5). In the tissues, G3139 represented an average of 67% of the total radioactivity in the liver and 77% of the total radioactivity in the kidney; 7.2% of the administered radioactivity was excreted on day 1 and 12.9% after 3 days, 3% was present in the 24 hr feces and 8% after 3 days (table 2).

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Time course of radioactivity in plasma () (in rad equivalents/ml), liver (), spleen (), kidney (), heart plus lung () and gut () (in rad equivalent/g) and bone marrow () (in rad equivalents/g protein) after s.c. infusion of 106 µg of 35S-G3139 daily for 7 days to BALB/c mice.

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Time course of G3139 in plasma (), kidney () and liver () after s.c. infusion of 106 µg of G3139 to BALB/c mice for 7 days.

View larger version (15K): [in this window] [in a new window]   Fig. 7.   Plasma radiochromatograms obtained 3, 9 and 14 days after s.c. infusion of 106 µg of 35S-G3139 daily for 7 days.

View larger version (22K): [in this window] [in a new window]   Fig. 8.   Liver, kidney and urine (from top to bottom) radiochromatograms after G3139 administration. Left, liver at 6 hr, kidney at 6 hr and urine at 24 hr after G3139 i.v. Right, liver at 7 days, kidney at 9 days and urine at 24 hr after G3139 s.c.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

A prerequisite for antisense molecules to become efficacious drugs is their ability to reach the intracellular target. The first step in this process is to resist degrading enzymes in the blood and the interstitium. Phosphorothioate antisense oligonucleotides have been shown to be more resistant than their phosphodiester counterparts to nuclease cleavage and are therefore better potential therapeutic agents (Zon, 1988). Large pharmacokinetic variations have been recorded in various studies, reflecting the size, structure, dose of oligonucleotide used and species studied (mouse, rat or monkey) (Crooke, 1991; Srinivasan and Iversen, 1995). Although 3-exonucleases have been shown to be implicated in the degradation of a 20-mer phosphorothioate antisense to the human immunodeficiency virus tat splice acceptor site (Temsamani et al., 1992), it is not clear whether all phosphorothioates will be metabolized the same way, irrespective of their size and structure. Therefore, further studies are required to gain more information regarding the biodistribution and metabolism of these compounds.

Our study provides real pharmacokinetic parameters for G3139 and not merely the pharmacokinetics of radioactivity. After i.v. administration to BALB/c mice, G3139 was widely and rapidly distributed and slowly eliminated from the plasma, with a terminal half-life of 11 hr. The G3139 plasma pharmacokinetics fitted a three-compartment model. Similar plasma triexponential decay was observed in rats after i.v. administration of a 3.6 mg/kg level of a 20-mer phosphorothioate antisense to papilloma viruses (Cossum et al., 1993). The protein binding that we observed was extensive (initially 98%), consistent with the high affinity of these molecules for albumin and ₂-macroglobulin that has previously been reported (McCormack et al., 1990; Zhang et al., 1995b). After slow infusion, steady-state plasma levels were reached by day 3, which is consistent with rapid absorption and a terminal half-life of elimination of 22 hr. For both routes, the protein binding paralleled the percentage of parent drug, suggesting that G3139 plasma degradation products do not bind to plasma proteins or bind to a lesser extent than does the parent drug.

After G3139 administration by both routes, the radioactivity was shown to concentrate in liver and kidneys, which has also been observed with other S-oligonucleotides (Agrawal et al., 1991; Bigelow et al., 1992; Goodarzi et al., 1992; Cossum et al., 1993; Zhang et al., 1995a). Although it is true that the total radioactivity was higher in the liver than in the kidneys, due to the greater weight of the liver, the relative radioactivity (in micrograms of rad equivalents per gram of tissue) was significantly higher in the kidney; the kidney to plasma ratio was 99, compared with the liver to plasma ratio of 30. A clear cumulative effect was observed in the tissues after slow infusion, with radioactivity levels as high as 8.2 µg rad equivalents/g protein in the bone marrow.

A variety of degradation products were formed mainly by the liver, and metabolism was more extensive with the i.v. schedule, compared with the s.c. schedule. However, we did not examine the metabolic profiles in the s.c. schedule before 24 hr after implantation of the minipumps. It is possible that the oligonucleotide is sequestrated to particular sites within the liver, and this is currently under investigation. The radioactivity extracted from the liver and kidney is close to the total radioactivity, suggesting that all metabolites have been extracted. The observation that the percentage of radioactivity bound to plasma proteins was similar to the percentage of parent drug suggests that the degradation products of G3139 do not bind proteins. Our study also showed that it is these metabolites that are excreted and not the parent drug, in agreement with other studies (Temsamani et al., 1992). Our extraction of urinary radioactivity showed low recoveries (<40%), suggesting that even more degradation occurred. In the i.v. schedule, our data agree with previous reports showing that 30% of the radioactivity was found in 24-hr urine (Agrawal et al., 1991; Temsamani et al., 1992; Zhang et al., 1995a). However, others found that some intact oligonucleotide was excreted (McCormack et al., 1990). Twenty-four hours after administration, when equivalent doses would have been received by the two routes, only 7.2% of radioactivity was excreted after s.c. infusion, as opposed to 33% after i.v. administration. The clearance of this type of compound is relatively rapid and significantly less radioactivity is eliminated by both renal and fecal routes after continuous infusion; this suggests that more oligonucleotide may have been available. The fact that less degradation was observed in the liver suggests that there could be a lower percentage of metabolites to be eliminated and therefore more parent drug present. The nature of the degradation products is as yet unknown, but other studies have revealed that shorter oligonucleotides can be formed (Saijo et al., 1994; Zhang et al., 1995a). A relatively low percentage of the plasma and liver radioactivity was found to elute after the parent drug, which was also observed in previous studies in mice (Agrawal et al., 1991). In that study, the authors concluded that this metabolite could originate from the reaction of the oligonucleotide with a small endogenous compound. The broadening of the G3139 peak in tissues is consistent with the formation of metabolites with structures close to that of G3139. Somewhat surprisingly, although we saw very little sign of kidney degradation, the G3139 half-life of elimination in the kidney was higher than that in the liver.

In conclusion, our data show that G3139 is widely distributed and slowly eliminated, mainly in the urine and feces. Continuous s.c. infusion resulted in significantly more parent drug reaching the tissues and bone marrow, probably reflecting the larger dose received. In addition, there was a reduction in metabolism and elimination, compared with a single i.v. bolus.