Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Antisense oligonucleotides represent an exciting new class of therapeutic agents that are designed to inhibit the expression of viral and mammalian genes in a selective and sequence specific manner. Because unmodified phosphodiester linkages proved quite labile to serum and cellular nucleases (Sands et al., 1994), replacement of each nonbridging oxygen in the backbone with a sulfur atom, thereby forming a phosphorothioate linkage, provided a dramatic improvement in the drug half-life in vitro and in vivo (Crooke et al., 1995). Phosphorothioate oligodeoxynucleotides have demonstrated broad distribution and acceptable toxicity profiles after systemic delivery in a variety of species (Crooke et al., 1996; Agrawal et al., 1995; Iverson et al., 1995 and Rifai et al., 1996). These studies revealed that the major sites of distribution of phosphorothioate oligonucleotides after i.v. administration included the liver, kidney, spleen and bone marrow. Although distribution data are important, these experiments did not provide information concerning localization of phosphorothioate oligonucleotides within the cells of these organs.

More recent in vivo experiments have begun to address suborgan distribution and intracellular localization of phosphorothioate oligodeoxynucleotides in a variety of organs and tumor isolates after systemic administration (Butler et al., 1997; Plenat et al., 1995; Rifai et al., 1996; Nicklin et al., 1998, in press). Endpoints for these analyses included direct detection of rhodamine and digoxigenin-conjugated phosphodiester and phosphorothioate oligomers, autoradiography using ¹²⁵I, ³ H, ¹⁴ C and ³⁵ S labeled compounds, as well as an indirect immunostaining method using a monoclonal antibody recognizing a 20-mer phosphorothioate oligodeoxynucleotide heterosequence (ISIS 2105). Results of these studies suggested that, within 24 hr, phosphorothioate oligodeoxynucleotides accumulated in renal proximal convoluted tubular cells, skin fibroblasts, dendritic cells and within the hepatocytes, Kupffer and endothelial cells of the liver. In hepatoctyes, lower concentrations of phosphorothioate oligonucleotides were observed relative to other cell types.

Although the aforementioned experiments demonstrated that phosphorothioate oligodeoxynucleotides are grossly localized within cells of the rodent liver, none of these studies provided quantitative pharmacokinetic data. Recently, an attempt to quantify distribution of phosphorothioate oligonucleotides within hepatic cell types using ³H-labeled oligomer has been reported (Bijsterbosch et al, 1997). Results from these experiments indicated that uptake of radiolabeled oligonucleotides is predominantly found in endothelial and parenchymal cell types, with minimal uptake observed in Kupffer cells. In addition, whole liver subcellular fractionation indicated that a majority of the internalized radiolabeled oligonucleotides were sequestered within lysosomal compartments with minimal nuclear accumulation. It should be noted that Bijsterbosch et al. (1997) made these observations without demonstrating the integrity of the oligonucleotide isolated from the liver after systemic administration. These results differed from most reports using autoradiography and antibody staining which indicated that Kupffer cells internalize substantial levels of phosphorothioate oligonucleotides with nuclear-association evident (Butler et al., 1997; Rifai et al., 1996).

To provide even greater detail concerning the cellular and subcellular distribution and metabolic fate of a 21-mer phosphorothioate oligonucleotide in rat liver over 24 hr using a variety of dose levels, without the caveats associated with radiolabeled or conjugated phosphorothioate oligonucleotides, we used cellular isolation, fractionation and analytical techniques (Deschenes et al., 1980; Crooke et al., 1995, 1996; Leeds et al., 1996) which were previously standardized in our laboratories. Our data suggest that, although pharmacokinetic parameters vary as a function of liver cell type, there is significant intracellular delivery within parenchymal and nonparenchymal cells after systemic administration.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Oligonucleotide Synthesis

ISIS 1082 (GCCGAGGTCCATGTCGTACGC), a 21-base phosphorothioate oligodeoxynucleotide, was synthesized at Isis Pharmaceuticals on a Milligen 8800 DNA synthesizer by the phosphoramidite method. The thiation reagent was synthesized as described previously (Iyer et al., 1990). The oligomer was purified using reverse phase HPLC and determined to be greater than 85% full length by CGE.

Preparation of ³⁵S -1082

The ³⁵S radiolabel was incorporated at a single position at the most 5' linkage of the ISIS 1082 sequence by the after method. P(III) solid phase oligonucleotide synthesis was performed utilizing commercial phosphoramidites and the unlabeled sites oxidized to P(V) phosphorothioates via Beaucage reagent as described previously (Iyer et al., 1990). During the synthetic process, site-specific S was incorporated by initially oxidizing P(III) triesters to P(V) ³⁵S -phosphorothiotriesters with ³⁵S₈. Complete oxidation of the site was insured by after the ³⁵S oxidation with standard Beaucage reagent oxidation. After complete synthesis of the oligonucleotide, full length product was purified from shorter synthetic failure sequences, as well as any remaining trapped ³⁵S₈, by trityl-on reverse phase HPLC. The final product was more than 90% full length radiolabeled compound with a specific activity of 1.9 × 10⁹ dpm/µmole.

Formulation of Radiolabeled ISIS 1082

The ³⁵S labeled ISIS 1082 was formulated in 1X PBS (pH 7.4, without calcium or magnesium). Unlabeled and radiolabeled oligonucletides were combined to produce a 600 µM dosing solution and were sterilized using a 0.2 µM Centrix centrifugal filter (Schleicher & Schuell, Keene, NH) before administration. The specific activity for the radiolabeled compound was 2.48 × 10⁸ dpm/µmole.

Animals

Male Sprague-Dawley rats (130-200 g) were obtained from Harlan Sprague-Dawley (Madison, WI). Results from multiple experiments performed within this age and weight range revealed no significant hepatic oligonucleotide uptake or distribution differences. The animals were housed in polycarbonate cages and fed ad libitum. Doses of 5, 10, 25 or 50 mg/kg were administered by i.v. injection (tail vein) and animals killed after 24 hr. Time course experiments were conducted using a 10-mg/kg dose.

For radioactive studies, two animals were injected with approximately 25 µCi of ³⁵S, corresponding to a 10-mg/kg dose. This dose was chosen because it was within a relevant pharmacological and toxicological range (Crooke et al., 1995). After 4 hr, animals were sacrificed and various isolation procedures were performed. For whole organ determinations, the liver was perfused with 200 ml of wash buffer (10 mM HEPES, 142 mM NaCl and 6.7 mM KCl pH 7.4) to remove the blood, excised and weighed, followed by direct homogenization of the entire organ in Proteinase K digestion buffer (0.5% NP-40, 25 mM Tris pH 8.0, 25 mM EDTA and 100 mM NaCl). Quadruplicate (100 mg) aliquots were combined with 10 ml Beckman Ready Safe (Fullerton, CA) scintillation cocktail and counted using a Beckman LS 6000IC Liquid Scintillation Counter using dual source counting to correct for quench differences. S counting efficiency of physiological isolates under these conditions was approximately 93%.

Liver Perfusion

The liver was perfused, with minor modification, as described previously (Berry and Friend, 1972; Deschenes et al., 1980; Seglen, 1976). Briefly, rats were anesthetized with an i.p. injection of sodium pentobarbital. The inferior vena cava was catheterized and perfused in a retrograde flow with the inferior vena cava clamped proximal to the liver with the mesenteric vessels cut for drainage. Perfusion was performed using 250 ml of a wash buffer (described previously), followed by and additional 200 ml of wash buffer supplemented with 1.26 mM CaCl₂ and 0.5 mg/ml collagenase (Sigma Chemical Co., St. Louis, MO) at a flow rate of 30 ml/min. Buffers maintained an organ temperature of approximately 37°C. After collagenase treatment, the liver was removed and placed in 100 ml of ice cold 1X PBS. After gentle mincing, the suspension was poured through sterile 150-µ nylon mesh (Tetko, Buffalo, NY).

Purification of Parenchymal and Nonparenchymal Cells

Hepatocytes, Kupffer and endothelial cell types were isolated from whole liver as described previously (Seljelid and Smedsrod, 1981; Berry and Friend, 1972; Deschenes et al., 1980; Seglen, 1976). Hepatocytes were isolated from the liver perfusate described above by centrifugation at 50 × g for 5 min. in a Beckman tabletop centrifuge (Fullerton, CA). The supernatant containing nonparenchymal cells was removed and retained on ice. Centrifugation was repeated three additional times to remove contaminating cells. After the final pelleting, hepatocytes were resuspended in 1X PBS, counted and viability determined using trypan blue. Viability of hepatocytes was generally more than 90% after this procedure.

The initial supernatant, containing nonparenchymal cells, was further separated into endothelial and Kupffer cell populations by brief adherence to plastic at 37°C followed by trypsinization and cell scraping to take advantage of the differential adherence characteristics of the two cell types. It has been well documented that Kupffer cells interact, in a trypsin insensitive manner, with plastic or glass substrates (Seljelid and Smedsrod, 1981; Garvey and Caperna, 1981). By isolating endothelial cells with trypsin, washing and then removing Kupffer cells by scraping, a rapid and effective nonparenchymal cell type enrichment was accomplished.

More specifically, the washes from the hepatocyte preparations were pooled and centrifuged at 500 × g for 10 min to recover the mixed nonparenchymal cell types. The resulting cell pellet was resuspended in 40 ml of DME (Gibco/BRL, Grand Island, NY) containing 15% heat inactivated FCS (HyClone, Logan, UT), transferred to a T-175 flask (Falcon, Becton/Dickinson; Franklin Lakes, NJ) and allowed to adhere for 1 hr at 37°C in a humidified incubator with 5% CO₂. After adherence, endothelial cells were removed using 7 ml of trypsin, followed by suspension in 7 ml of 15% DME. The flask was then washed with an additional 25 ml of 15% DME before isolation of Kupffer cells. To recover the Kupffer cells, 15 ml of the media containing serum were added to the flask and cells removed using a cell scraper. Subsequently, both cell isolates were recovered by centrifugation at 500 × g for 10 min. Typically, 20 to 30 × 10⁶ endothelial cells and 1 to 3 × 10⁶ Kupffer cells were obtained, with viabilities more than 85% using trypan blue dye exclusion. For distribution and metabolism studies, 4 × 10⁶ hepatocytes and endothelial cells and the total yield of Kupffer cells were used for each sample.

Validation of Parenchymal and Nonparenchymal Cells

Flow cytometry. Because hepatocytes are considerably larger than nonparenchymal cell types, the enrichment of parenchymal cells was confirmed by observing the forward and side scatter display using a flow cytometer (Becton Dickinson FacsScan, San Jose, CA). Comparison of pre- vs. postseparated cells revealed that hepatocytes were 95% free of other contaminating cell types, although the nonparenchymal cell isolates were 92% free of parenchymal cells (data not shown).

Immunohistochemistry

Cytocentrifuge preparations. The isolated liver cell populations were washed in PBS and suspended at a concentration of approximately 5 × 10⁵/ml. A 150-µl cell suspension was added to cytocentrifuge funnels clipped to glass slides. The slides were spun at 520 × g for 5 min and air dried overnight at room temperature before fixation and immunohistochemical staining.

Double label immunohistochemistry. Endothelial cell preparations were stained with monoclonal antibodies to rat PECAM-1 (Serotech MCA1334, Oxford, England) and detected using HRP-donkey anti-mouse IgG F(ab') 2 (Jackson ImmunoResearch laboratories, West Grove, PA). Hepatocyte preparations were stained with rabbit antibodies to mouse albumin (ICN, Costa Mesa, CA) and detected using alkaline phosphatase-donkey anti rabbit IgG F(ab')2. Kupffer cells were stained with mouse anti-rat mononuclear phagocyte antibodies (Pharmingen clone 1C7, San Diego, CA) and were detected with either HRP or alkaline phosphatase conjugated donkey anti-mouse IgG F(ab') 2. All slides were stained sequentially on a DakoAutostainer (Dako Corporation, Carpenteria, CA). The substrate used for HRP conjugated detection was DAB and new fuchsin for the alkaline phosphatase conjugated antibodies (Dako Corp.). All slides were counterstained with hematoxylin and mounted with Gel/Mount (Biomeda Corp., Foster City, CA).

View larger version (86K): [in this window] [in a new window]   Fig. 1.   Double label immunohistochemistry results from hepatocyte (A), Kupffer cell (B) and endothelial cell (C) isolates, respectively. After isolation, cells were applied to slides using a cytocentrifuge and incubated with anti-rat PECAM-1 (brown) and antirat mononuclear phagocyte (red) antibodies to identify endothelial and Kupffer cells, respectively. In the hepatocyte isolate, two Kupffer cells can be observed in the center of the field. Cells were visualized at 380 × magnification.

Subcellular fractionation. The fractionation procedure used to isolate nuclear, membrane and cytosolic constituents was performed as described previously, with minor modification (Crooke et al., 1995). Cell lysis was accomplished using an isotonic buffer consisting of 0.5% NP-40, 10 mM Tris-Cl (pH 7.4), 140 mM KCl, 5 mM MgCl₂ and 1 mM DTT (Berger and Chirgwin, 1989). This buffer preserved the morphology of the isolated nuclei as assessed by phase-contrast microscopy (data not shown). After lysis, nuclei were maintained at 4°C and isolated by centrifugation for 5 min at 1310 × g. The supernatant, which was retained, contained the cytosolic/membrane fraction. The pellet was washed and centrifuged once again under the same conditions, with the final pellet representing the purified nuclear fraction and was confirmed by phase contrast microscopy. Membrane constituents were separated from the cytosol by transfer to thick-walled polycarbonate tubes and centrifugation (200,000 × g) for 30 min at 4°C in a Beckman TLA 100.2 fixed angle rotor using a Beckman TL-100 benchtop ultracentrifuge (Fullerton, CA). After centrifugation, the supernatant represented purified cytosol although the pellet contained plasma membrane and various subcellular organelles including endosomes and lysosomes. Previous control experiments using this technique revealed minimal cross-contamination between the various isolated fractions (Crooke et al, 1995).

Cellular digestion and organic extraction for CGE analysis. Whole cells and cellular fractions were digested using proteinase K extraction solution, as described previously. Proteinase K enzyme (Boehringer Mannheim, GmbH, Germany) was added to each sample to a final concentration of approximately 100 µg/ml. In addition, 30 pmol of an internal standard (polyT 27-mer phosphorothioate oligodeoxynucleotide) was added before enzymatic digestion to permit quantitation of ISIS 1082. Samples were incubated for 2 hr at 55°C to digest proteins. After digestion, 200 µl of 30% ammonium hydroxide was added to each sample before organic extraction using 1 ml of phenol/isoamyl alcohol/chloroform (24:1:24), as described previously (Crooke et al., 1996).

Solid phase extraction. To purify samples sufficiently for CGE, two solid phase extraction columns were required (Crooke et al., 1996). Briefly, removal of residual contaminants was accomplished using a SAX SPE column (J & W Scientific, Folsom, CA) followed by desalting using a reverse phase SPE column [Isolute C18(EC), Mid Glamorgan, U.K.]. A final desalting step was employed prior to CGE analysis to further reduce the amount of competitive anions that would be loaded during electrokinetic injections. Samples were placed on 0.025-µm dialysis membranes (Millipore, Bedford, MA) and floated over 60-mm culture dishes containing 10 ml of 18.3 M-cm dH₂O for 30 min before analysis.

CGE analysis. Samples were placed into microvials and analyzed using a Beckman PA/CE System Gold 5010 capillary electrophoresis system (Fullerton, CA) with a UV- detection at A260 nm. Samples were resolved using a 100 µM ID capillary column (Polymicro Technologies Inc, Phoenix, AZ) filled with 11% polymerized acrylamide (Fluka, Neu-Ulm, Switzerland). The electrophoretic buffer used in the capillary and running buffer contained 200 mM bis [2-Hydroxyethyl] iminotris [hydroxymethyl]-methane (Sigma Chemical Co., St Louis, MO), 200 mM boric acid (Fluka) and 8.3 M urea (Boehringer Mannheim, GmbH, Germany). Samples were electrokinetically applied using 5 to 10 kV for 5 to 10 sec, although separations were achieved operating at 20 kV constant voltage in approximately 5 min at 50°C. Samples were injected and quantified within the linear range of the detector which spanned approximately 0.01 to 0.001 measured absorbance units.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Total Organ Distribution

To delineate the ratio of the systemic dose distributed between whole liver, parenchymal and nonparenchymal cells and the extracellular matrix, in vivo uptake and fractionation experiments were performed using whole livers isolated 4 hr after i.v. bolus administration of a 10-mg/kg dose of ³⁵S labeled ISIS 1082. Results, which were derived from quadruplicate 100-mg samples of whole liver homogenates, indicated that liver-associated radioactivity represented 12.3% (±0.047) of the total dose, corresponding to an organ oligomer concentration of 2.88 µM. Results from CGE analysis performed under the same conditions revealed a concentration of intact drug and metabolites of 3.83 µM, which correlated well with the radioactivity derived determination.

After collagenase treatment, radioactivity in the total unpurified cell isolates was measured and this value compared to the level determined in whole liver. Results from this experiment indicated that 53.6% (±1.32) of the total liver dose was cell-associated, although the remaining 46.4% of the drug was found in the perfusate. The material within the perfusate could arise from drug associated with connective tissue, extracellular spaces and cells damaged during the perfusion.

Kinetics of Whole Liver Association

To assess the pharmacokinetic behavior of ISIS 1082, rats were dosed by i.v. bolus injection and killed at 1, 4, 8 and 24 hr after a 10-mg/kg drug dose. The concentration of intact drug and metabolites as a function of time was determined by CGE analysis using 100 mg of total liver homogenate (fig. 2). After 1 hr, total drug level reached 4.21 µM and remained relatively constant for up to 8 hr. Between 8 and 24 hr, oligomer levels decreased by approximately 50% to 2.72 µM.

View larger version (85K): [in this window] [in a new window]   Fig. 2.   Effect of time on the level of ISIS 1082 and metabolites within 100 mg of homogenized whole rat liver after 10 mg/kg i.v. bolus. Whole liver is expressed in µM amount detected and each point represents the mean and S.E.M. of quadruplicate samples.

Effect of Dose

The effect of dose on the concentration of intact drug and metabolites detected within whole liver after 24 hr was measured after i.v. administration of 10, 25 and 50 mg/kg of ISIS 1082 (fig. 3). As might be expected, levels of liver-associated oligomer increased as dose was escalated. At a dose of 10 mg/kg, measured drug levels were approximately 3 µM, although at 50 mg/kg, the total liver concentration achieved was approximately 12.3 µM.

View larger version (64K): [in this window] [in a new window]   Fig. 3.   Effect of dose on the accumulation of ISIS 1082 and metabolites in 100 mg of homogenized whole rat liver 24 hr after i.v. bolus of 10, 25 and 50 mg/kg, respectively. Whole liver concentrations are expressed in µM detected and each point represents the mean and S.E.M. of quadruplicate samples.

Suborgan Localization

Liver cell distribution. The changes in oligonucleotide abundance in the total liver cell population as a function of time were determined (fig. 4). One hour after i.v. bolus dosing, the amount of ISIS 1082 and metabolites was 2 × 10⁸ molecules/cell. No increase was detected by 4 hr, however, by 8 hr, levels had increased to approximately 5.5 × 10⁸ molecules/cell. At 24 hr levels decreased to roughly 3.5 × 10⁸ molecules/cell. The kinetics of cellular accumulation differed slightly from those observed for whole liver (fig. 2). The higher clearance rate observed after 24 hr in whole liver is probably the result of hepatic interstitial efflux.

View larger version (91K): [in this window] [in a new window]   Fig. 4.   Effect of time on the amount of ISIS 1082 and metabolites found in whole rat liver cells after 10 mg/kg i.v. bolus. Data are expressed as the number of molecules/cell. Values represent the mean and S.E.M. of 12 replicates.

View larger version (72K): [in this window] [in a new window]   Fig. 5.   Effect of dose on the accumulation of ISIS 1082 and metabolites in whole rat liver cells 24 hr after i.v. bolus at 5, 25 and 50 mg/kg, respectively. Data are expressed as the number of molecules/cell. Values represent the mean and S.E.M. of 12 replicates.

Accumulation of ISIS 1082 and Metabolites into Different Liver Cell Types

Kinetics. The effect of time on the proportion of ISIS 1082 and metabolites found within hepatocytes, Kupffer and endothelial cells over a 24-hr period after a 10 mg/kg i.v. bolus dose was evaluated (fig. 6). One hour after dosing, substantial differences in the uptake of drug by hepatocytes and nonparenchynmal cell types were evident. In hepatocytes, the proportion of ISIS 1082 and metabolites was approximately 20% of the total liver cell dose for the first 8 hr, decreasing to approximately 7% after 24 hr. In contrast, both Kupffer and endothelial cells retained roughly 80% of the total cell-associated dose. After 1 hr, endothelial cells contained a greater proportion of ISIS 1082 relative to the Kupffer cells, but thereafter, maintained equivalent proportions of ISIS 1082 and metabolites for the remaining 24 hr.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   The effect of time on the proportion of ISIS 1082 and metabolites in hepatocytes (n = 9), Kupffer cells (n = 3) and endothelial cells (n = 6) 1, 4, 8 and 24 hr after 10 mg/kg i.v. bolus. The mean and S.E.M. for each cell type are plotted. Symbols are as follows: [squlo], hepatocytes; [trio], Kupffer cells; [circo], endothelial cells.

Dose response. When cellular distribution was evaluated after 24 hr, as a function of ISIS 1082 dose, some interesting trends emerged (fig. 7). At the 5- and 10-mg/kg doses, nonparenchymal cells contained the majority of the total liver cell-associated drug, results that were consistent with data presented previously in figure 6. However, as the dose was increased to 25 mg/kg, substantial shifts in relative accumulation amongst the cell types became evident. The percentage of cell-associated ISIS 1082 and metabolites in hepatocytes and endothelial cells increased, but Kupffer levels declined. This suggests that, at doses more than 10 mg/kg, Kupffer cells were saturated. At 50 mg/kg, the percentage of the dose found in the endothelial cells also declined, with saturation of those cell types occurring at approximately 25 mg/kg. In contrast, hepatocytes continued to internalize drug, with no evidence of saturation even at the highest dose evaluated.

View larger version (17K): [in this window] [in a new window]   Fig. 7.   The effect of dose on the proportion of ISIS 1082 and metabolites in hepatocytes (n = 9), Kupffer cells (n = 3) and endothelial cells (n = 6) 24 hr after i.v. bolus dosing at 5, 10, 25 and 50 mg/kg. The mean and S.E.M. for each cell type are plotted. Symbols are as follows: [squlo], hepatocytes; [trio], Kupffer cells; [circo], endothelial cells.

Subcellular Localization

Kinetics. To fully characterize the intracellular localization of ISIS 1082 and metabolites within liver cell types after systemic administration of the compound, fractionation of cell isolates was performed. The intracellular distribution patterns in hepatocytes, Kupffer cells and endothelial cells were evaluated over a 24-hr period (fig. 8). In hepatocytes, at 10 mg/kg, no nuclear-associated drug was detected at any time. However, both cytosolic and membrane drug levels appeared to increase from 1 to 8 hr followed by a significant decrease by 24 hr. In these cells, as well as nonparenchymal cell types, the amount of membrane-associated ISIS 1082 and metabolites was lower than that seen in the cytosolic fractions.

View larger version (43K): [in this window] [in a new window]   Fig. 8.   The subcellular localization of ISIS 1082 and metabolites over a 24-hr period in rat hepatocytes (A). Kupffer cells (B) and endothelial cells (C), respectively. Values for each fraction represent the mean and S.E.M. of intact drug and metabolites expressed in molecules/cell. Replicates are as follows: Hepatocytes (n = 9), Kupffer values (n = 3) and endothelial cells (n = 9). Bars are as follows: [squlo], nuclear;[rhbox], cytosolic and [squlf], membrane.

Dose response. To determine the effect of dose on the subcellular localization of ISIS 1082 and metabolites in subcellular isolates, the same dose regimen as described previously for whole liver and liver cells was performed. The effect of dose on subcellular distribution was most pronounced in hepatocytes (fig. 9A). After 24 hr, at the dose of 5 mg/kg, ISIS 1082 and metabolites were detected only in cytosolic fractions. However, when the dose was increased to 10 mg/kg, membrane-associated drug was detected. Finally, as the dosage was escalated from 25 to 50 mg/kg, nuclear-associated ISIS 1082 and metabolites were evident. Thus, to achieve nuclear drug levels in rat hepatocytes after i.v. administration, a minimal dose of 25 mg/kg was required. At 50 mg/kg, ISIS 1082 and metabolites levels were still increasing with no suggestion of cell saturation.

View larger version (39K): [in this window] [in a new window]   Fig. 9.   The subcellular distribution of ISIS 1082 and metabolites 24 hr after 5, 10, 25 and 50 mg/kg i.v. bolus hepatocytes (A). Kupffer cells (B), and endothelial cells (C), respectively. Note that at doses less than 25 mg/kg, hepatocytes possessed no nuclear-associated oligonucleotide. Values for each fraction represent the mean and S.E.M. of intact drug and metabolites expressed in molecules/cell. Replicates are as follows: Hepatocytes (n = 9), Kupffer values (n = 3)and endothelial cells (n = 9). Bars are as follows:, nuclear; , cytosolic and , membrane.

Metabolism

The metabolic profiles of ISIS 1082 isolated from whole organ, cellular and subcellular fractions were evaluated. Typical electropherograms generated from hepatocyte, Kupffer and endothelial cellular compartments 24 hr after a 10-mg/kg i.v. bolus are shown (fig. 10).

View larger version (16K): [in this window] [in a new window]   Fig. 10.   Electropherograms generated from hepatocyte (A), Kupffer cell (B) and endothelial cell (C) subcellular fractions 24 hr after 10 mg/kg i.v. bolus. Nuclear, cytosolic and membrane isolates are labeled and the positions of full length ISIS 1082 and internal standard T27 are indicated by arrowheads. The migration time (in min) is indicated.

Results from whole liver drug stability studies evaluating the integrity of ISIS 1082 as a function of time are shown in figure 11, with A representing the percent of intact ISIS 1082 and B representing full length drug, as well as the first two nucleotide elimination products (n-1, n-2). Because it has been demonstrated that, in general, loss of the two residues results in metabolic products with some, albeit reduced, pharmacological antisense activity (Crooke, 1995), we believed it worthwhile to include these data. One striking feature of these data is that it appears that a majority of the observed drug metabolism occurred within the first hour after administration to animals. In general, the amount of intraorgan metabolism appears to be rather modest over the rest of the 24-hr period, with percent full length being reduced from approximately 38% after 1 hr to 28% after 24 hr. This would suggest that a majority of the metabolism observed occurred extracellularly. After 24 hr, when full length and the first two metabolic products are included, more than 50% of the total oligomer represented pharmacologically active drug and metabolites. The effects of dose on stability were also evaluated (fig. 12). Once again, after trends observed in the 10-mg/kg time course experiments, only a moderate increase intact ISIS 1082 was observed over the entire dose range.

View larger version (83K): [in this window] [in a new window]   Fig. 11.   Effect of time on the stability of ISIS 1082 in whole rat liver (100 mg) after 10 mg/kg i.v. bolus. A shows the level of intact drug relative to metabolites and B includes intact and n-1 and n-2 metabolites as a measure of total pharmacologically active oligonucleotide present.

View larger version (84K): [in this window] [in a new window]   Fig. 12.   Effect of dose on the integrity of ISIS 1082 in whole rat liver (100 mg) 24 hr after a single 5, 10, 25 or 50 mg/kg i.v. bolus. A shows the level of intact drug relative to metabolites and B includes intact and n-1 and n-2 metabolites as a measure of total pharmacologically active oligonucleotide present.

Subcellular Metabolism

The extent of metabolism as a function of liver cell type and cell compartment was also compared (fig. 13). In general, the fractions prepared from all cell types displayed minimal time-dependent metabolism. As was seen in whole liver, the greatest degradation was apparent as early as 1 hr after 10 mg/kg i.v. bolus. Of the cell types evaluated, hepatocytes appeared to metabolize the drug to the greatest extent, with percent full length averaging 40% over the entire time course. Kupffer and endothelial and cells displayed somewhat less drug metabolism, approximately 50 to 60% over the same period.

View larger version (51K): [in this window] [in a new window]   Fig. 13.   Effect of time on the stability of ISIS 1082 in hepatocyte (A), Kupffer cell (B) and endothelial cell (C) subcellular fractions after a 10-mg/kg i.v. bolus dose. Note that in hepatocytes at this dose, no nuclear-associated oligonucleotide could be detected. Bars are as follows:[squlo], nuclear; [rhbox], cytosolic and [squlf], membrane.

The effect of dose on the stability of ISIS 1082 isolated from liver cell types was also assessed (fig. 14). In hepatocytes, the stability of ISIS 1082 increased as the dose was increased to 10 mg/kg, although at higher doses, the metabolic rate appeared constant. In Kupffer cells, 24 hr after 5 mg/kg dosing, there was difference in compartmental nuclease activity observed, with the nuclear-associated oligonucleotide roughly 50% full length, with cytosolic and membrane-associated oligonucleotide 30 and 35% intact, respectively. At higher doses this difference was not as dramatic, but in general, oligonucleotides isolated from the Kupffer cell nuclei were less metabolized than in either cytosolic or membrane isolates. In contrast to the other cell types evaluated, endothelial cell fractions revealed no dose-dependent trends in stability. Additionally, the subcellular compartments behaved similarly.

View larger version (47K): [in this window] [in a new window]   Fig. 14.   Effect of dose on the integrity of ISIS 1082 in hepatocyte (A), Kupffer cell (B) and endothelial cell (C) subcellular fractions 24 hr after a single 5, 10, 25 or 50 mg/kg i.v. bolus. Note that in hepatocytes, above 10 mg/kg dosing, oligonucleotide was present within the nuclei. Bars are as follows: [squlo]], nuclear; [rhbox], cytosolic and [squlf], membrane.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Previous studies have reported the in vivo distribution and metabolism of phosphorothioate oligodeoxynucleotides after i.v. administration (Nicklin et al, 1998, in press; Iverson et al., 1995; Rifai et al., 1996; Geary et al., 1997b). Although these studies provided pharmacokinetic information, they did not address intraorgan distribution. In contrast, more recent experiments have determined the intracellular localization of oligonucleotides within a variety of organ systems using autoradiography, immunostaining and rhodamine-conjugated phosphorothioate oligodeoxynucleotides (Butler et al., 1997; Rappaport et al., 1995; Rifai et al., 1996; Carome et al., 1997). Although these studies identified cell types within organs that internalized the oligonucleotides, they did not provide quantitative pharmacokinetic information. The most recent hepatic cell fractionation experiments used radiolabeled materials as a means of quantitation (Bijsterbosch et al., 1997) but could not evaluate oligonucleotide metabolism, which is a critical component of pharmacokinetic analysis.

By combining established separation, fractionation and analytical techniques described previously (Seglen, 1976; Crooke et al., 1995; Crooke et al., 1996), the first complete analysis of phosphorothioate oligodeoxynucleotide disposition and metabolism within the whole rat liver, hepatic cell types and subcellular compartments has been achieved using CGE. This analytical technique not only eliminates concerns about unincorporated radiolabel or loss of signal due to nucleolytic cleavage, but also provides a highly sensitive and quantitative method for determining in vivo intracellular oligomer distribution and metabolism with single base elimination resolution.

Studies using radiolabeled compound provided information on whole organ drug levels, as well as the fraction of the dose that was cell-associated. At 4 hr after a 10-mg/kg i.v. dose, the concentration of drug within the liver was 2.88 µM using ³⁵S ISIS 1082 vs. 3.83 µM by CGE analysis. These results are well within the range of variability found between different animals using a single analytical technique. At 4 hr, the amount of drug within the liver (using ³⁵S) represented 12.3% of the administered dose. Of that 12% liver-associated dose, approximately 50% was cell-associated, suggesting that a substantial proportion of the drug existed within the interstitium and extracellular spaces. These data are consistent with other reports showing that phosphorothioates have been localized to connective tissue and can bind to various proteins within matrices such as laminin and fibronectin (Butler et al., 1997; Plenat et al., 1995; Benimetskaya et al., 1995).

After a 10-mg/kg i.v. bolus dose, peak organ concentration was achieved after 8 hr, with slow efflux evident for 24 hr. Over a range of 5 to 50 mg/kg, organ concentrations rose from 2.7 to 12.3 µM, respectively, after 24 hr. These data suggest that a nonlinear relationship exists between dose and hepatic drug levels that was consistent with previous observations from our own and other laboratories (Geary et al., 1997b; Rafai et al., 1996; Nicklin et al., 1998, in press) .

Within 1 hr after a 10-mg/kg i.v. bolus, a majority of the oligonucleotide degradation detected within the liver had already occurred. At this timepoint, the compound was approximately 38% full length, and yet after 24 hr, the drug was still approximately 28% full length. These data suggest one or more of the following are occurring: 1) a majority of ISIS 1082 metabolism occurred from serum nuclease activity, although the drug was being cleared from the blood compartment to the liver, 2) endogenous hepatic nucleases are actually inhibited by the phosphorothioates and/or their metabolic breakdown products or not localized in the same cellular compartments, 3) there is a differential rate of metabolism for the Rp and Sp diastereoisomers present at each phosphorothioate linkage or 4) there is preferential efflux of degradation products although intact drug continues to be distributed from other sites in the body to the liver. It is likely, in fact, that all four processes occur in vivo, as suggested previously by others (Nicklin et al., 1998, in press; Glover et al., 1997). These studies demonstrated that within 1 hr after i.v. administration, metabolism was evident in both rodent and human plasma extracts, using a variety of different phosphorothioate oligonucleotide heterosequences. This degradation process appeared to be very rapid, with a majority occurring within 10 min and minimal metabolism observed thereafter, which would also agree with the diastereoisomeric nuclease selectivity seen by others (Spitzer and Eckstein, 1988). Various experiments, both in vitro and in vivo, have demonstrated the principal serum enzymatic activity results from 3' exonucleolytic cleavage (Crooke, 1993, 1995) resulting in progressive nucleotide deletions and formation of n-1, n-2, etc. degradation products. It has also been suggested that metabolism is the principal means of drug clearance in vivo (Nicklin et al., 1997).

Phosphorothioates have also been reported to inhibit a variety of enzymes, including RNase H, topoisomerases, DNA polymerases and nucleases (Crooke, 1995). Our data are consistent with those previous observations and in vitro experiments from our laboratories (Crooke, 1995). This trend was also noted in whole rat liver, where metabolism of ISIS 1082 appeared to be moderately attenuated as delivery concentration escalated. Specifically, 24 hr after dosing at 5 mg/kg approximately 18.2% of ISIS 1082 was still intact, whereas, after a 50-mg/kg dose, 34.9% represented full length material.

Suborgan localization studies revealed dramatic differences among liver cell types. On a per cell basis, nonparenchymal cells contained approximately 80% of the total liver cell-associated drug, being equally distributed between Kupffer and endothelial cells. Additionally, the kinetics of drug accumulation and elimination in these cellular populations were comparable. In contrast, hepatocytes accumulated the remaining 20% of the liver dose, with maximal levels being observed 4 to 8 hr after i.v. administration. Our data also suggest that ISIS 1082 was not extensively transferred from one cell type to another within the liver, as the pattern of hepatic cell distribution observed after 1 hr remained relatively constant.

Although these results differ from those described in a recent report (Bijsterbosch et al., 1997), suborgan distribution studies performed in our laboratories as well as others, using both autoradioradiography and immunohistochemical detection methods, have demonstrated substantial phosphorothioate oligodeoxynucleotide accumulation within Kupffer cells after i.v. administration (Butler et al., 1997; Plenat et al., 1995; Rifai et al., 1996; Inagaki et al., 1992). It should also be noted that the oligonucleotide doses given by Bijsterbosch et al. (1997) were much lower than ours. This may have influenced the results. Finally, Bijsterbosch et al. (1997) relied on radioactivity measurements and did not determine the integrity of the material in the liver or cellular compartments. In contrast, we have used CGE to extensively characterize the oligonucleotide metabolites in various cellular fractions. This difference in approach would also account for some of the discrepancies.

After 24 hr, at doses above 10 mg, Kupffer cell drug levels appeared to saturate, although in endothelial cells, maximal drug levels were observed at the 25-mg/kg dose. In contrast, above 25-mg/kg doses, cellular levels continued to increase in hepatocytes, with no evidence of saturation. This enhancement in uptake capacity observed in hepatocytes might not be surprising in light of their substantially greater cell volume relative to the nonparenchymal cell types. These results also suggest that, to deliver phosphorothioate oligodeoxynucleotides to hepatocytes at levels necessary for antisense inhibition, higher doses may be required relative to nonparenchymal cell types. This is consistent with studies in which effects on target mRNA levels in whole liver have been measured after systemic administration (Dean and McKay, 1994; Brown-Driver V, unpublished data). Specifically, when a 20-mer rodent specific C-raf kinase antisense deoxyphosphorothioate (ISIS 11061) was administered to rats, a dosing regiment of 100 mg/kg every 24 hr for 4 days was required to demonstrate mRNA reduction in whole liver.

Subcellular fractionation to determine the intracellular localization and metabolism of ISIS 1082 as a function of time and dose also highlighted interesting hepatic cell differences. Although hepatocytes contained both cytosolic and membrane-associated drug, no nuclear accumulation was observed over the entire 24-hr period after 10-mg/kg dosing. In general, 2- to 4-fold greater oligonucleotide levels were observed in the hepatocyte cytosol vs. membrane fraction over this period. The highest drug levels in either fraction were observed 8 hr after dosing. In contrast, both Kupffer and endothelial cells accumulated substantial drug in all three subcellular fractions. Both nonparenchymal cell types maintained readily detectable nuclear-associated drug levels over a 24-hr period, with the highest intracellular drug levels occurring after 8 hr postdosing. Relative to the Kupffer cells, endothelial cells accumulated greater nuclear-associated drug. At the 5- and 10-mg/kg dose levels, no nuclear ISIS 1082 could be detected in hepatocyte nuclei. However, drug could be driven into the hepatocyte nuclei by increasing the input dose to 25 mg/kg and above (fig. 9). At these doses of ISIS 1082, nuclear drug levels equaled those detected in the nonparenchymal nuclear fractions measured after 24 hr. These data clearly demonstrate that to achieve nuclear phosphorothioate oligonucletide accumulation in rat hepatocytes, and perhaps, pharmacological activity (Dean and McKay, 1994; Brown-Driver V, unpublished data), higher dosing levels and longer treatment times may be necessary.

Although the level of cytosolic-associated phosphorothioate oligodeoxynucleotide appears highly elevated relative to that found in the nucleus at any given time point, these data represent the total number of molecules present within a single cell. As such, the nucleus would represent a significantly smaller volume when compared to the cytoplasm and would therefore contain a greater proportion of drug.

Drug stability as a function of time, dose and intracellular location was also assessed. In general, 24 hr after a 10-mg/kg dose, minimal intracellular drug metabolism was observed in hepatocytes. In nonparenchymal cells degradation of ISIS 1082 was observed, with Kupffer and endothelial cells metabolizing drug at roughly equivalent rates. Metabolism of drug in subcellular fractions was also examined as a function of dose. As revealed in whole organ experiments, increasing the input dose only moderately enhanced stability.

In conclusion, a more complete picture of the complex cellular and subcellular distribution of a phosphorothioate oligodeoxynucleotide within rat liver has been achieved. It is clear, as has been described previously from in vitro cellular uptake studies and from early in vivo pharmacokinetic experiments, that certain cell types in vivo have the capability of internalizing oligonucleotide without the need for cationic lipids (Nestle et al., 1994; Wu-Pong et al., 1994). In disease states that require drug delivery to endothelial cells, for example, antisense drugs will likely prove highly effective, as these cells appear to readily take up drug. In other instances, drug delivery concentrations could be modulated to target less compliant cell types. In any event, we now have a means for evaluating the in vivo distributional changes influenced by either chemical modification or alternative formulations of antisense oligonucleotides after systemic delivery. This system also provides a basis for comparison of mechanisms of extravasation and intraorgan cellular uptake in vivo. Although we are interested in antisense oligonucleotides, the system should lend itself to analysis of other types of drugs as well.