Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Antisense oligonucleotides, designed to sequence specifically down-regulate the expression of target genes, have great potential as a research tool and as therapeutic agents (Szymkowski, 1996; Juliano et. al., 1999; Akhtar et al., 2000). A multitude of nuclease-resistant oligonucleotide analogs has been developed, of which the phosphorothioate analog is the most widely applied. Antisense phosphorothioate oligodeoxynucleotides (PS-ODNs) have been tested in a variety of in vitro and in vivo systems and were shown to be potent inhibitors of the expression of their target genes (Szymkowski, 1996; Juliano et al., 1999; Akhtar et al., 2000). Several have entered clinical trials, and a cytomegalovirus-specific PS-ODN (Vitravene) is approved for marketing (Akhtar et al., 2000).

In spite of many successful applications, the issue of effective cellular delivery of PS-ODNs remains to be resolved. Being large polyanions, PS-ODNs do not readily penetrate cell membranes. Various delivery strategies have been developed to increase the cellular uptake of PS-ODNs. Complexation of PS-ODNs with cationic lipids is at present commonly applied to facilitate their uptake by cells in culture, but it is not very suited for use in vivo (Juliano et al., 1999; Akhtar et al., 2000). An approach that was found to be effective, both in vivo and in vitro, is conjugation of PS-ODNs with cholesterol. It has been demonstrated in several cellular and animal models that cholesteryl-conjugated PS-ODNs display a higher antisense activity than their unconjugated counterparts (Desjardins et al., 1995; Alahari et al., 1996; Zhang et al., 1997; Epa et al., 1998; Okamoto and Nakano, 1999). Furthermore, it has recently been claimed that conjugation with cholesterol improves the oral bioavailability of PS-ODNs (Okamoto and Nakano, 1999). The improved efficacy of the cholesteryl-conjugated PS-ODNs may be explained by a higher cellular uptake, because it has been found in several studies that conjugation with cholesterol enhances the association of PS-ODNs with cells in culture (Temsamani et al., 1994; Alahari et al., 1996; Epa et al., 1998). Furthermore, we showed recently that the attachment of a single cholesterol at the 3' end of a PS-ODNs also affects the in vivo fate of the oligonucleotide. The cholesteryl-conjugated PS-ODN was found to display altered plasma protein binding and higher liver uptake (Bijsterbosch et al., 2000). In the present study, we investigate the disposition of a PS-ODN that is conjugated with cholesterol at both the 3' and 5' ends. It is shown that conjugation of a second cholesterol, in addition to a single cholesterol, markedly alters the plasma protein binding and disposition of the PS-ODN.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents. Polyinosinic acid (5'), polyadenylic acid (5'), and rat serum albumin were from Sigma-Aldrich (St. Louis, MO). Emulsifier Safe and Hionic Fluor scintillation cocktails and Soluene-350 were from Packard Instrument Company, Inc. (Downers Grove, IL). Na¹²⁵I (carrier-free) was from Amersham Biosciences UK, Ltd. (Little Chalfont, Buckinghamshire, UK). Proteinase K was purchased from Roche Applied Science (Mannheim, Germany). All other reagents were of analytical grade.

Oligonucleotide Synthesis and Purification. PS-ODNs, specific for murine intercellular adhesion molecule-1 (ICAM-1), were used. ISIS-3082 (sequence 5'-TGC ATC CCC CAG GCC ACC AT-3') and ISIS-9388 (same sequence as ISIS-3082 with a 3'-cholesterol-modified uridine as 3'-end base instead of thymidine) were synthesized and ³H labeled as described previously (Stepkowski et al., 1994; Bijsterbosch et al., 1997, 2000). ISIS-9389 was synthesized as ISIS-9388. The synthesis was carried out using the phosphoramidite that was derived from condensing 5'-DMT-2'-O-(6-aminohexyl)-thymidine with cholesteryl chloroformate and subsequent phosphitylation (Manoharan et al., 1995). To allow monitoring of its biological fate, ISIS-9389 was radiolabeled with ³H by heat-catalyzed exchange at the C8 positions of the purine nucleotides as described previously (Graham et al., 1993). [³H]ISIS-9389 was purified by reversed phase high-performance liquid chromatography, using a C₄ column (5 µm, 300 Å, 300 × 3.9 mm; Waters, Milford, MA), at a flow rate of 1 ml/min using the following mobile phases: A, 50 mM triethyl ammonium acetate, pH 7.0; and B, acetonitrile. After injection of the samples (0.5 ml), the column was eluted for 5 min with 10% B, followed by a gradient of 10 to 90% B (25 min). Subsequently, the column was eluted for 10 min with 90% B. The retention time of ISIS-9389 under these conditions was approximately 28 min (ISIS-3082 and ISIS-9388, 13 and 25 min, respectively). The radiolabeled oligonucleotide was precipitated as sodium salt by adding 10 volumes of 3% (w/v) NaClO₄ in acetone as described previously (Rump et al., 1998). The specific radioactivity of [³H]ISIS-9389 was approximately 50 × 10⁶ dpm/mg, and the radiochemical purity >98%.

Isolation and Radioiodination of Rat Lipoproteins. Rat low-density lipoprotein (LDL; density 1.024-1.063 g/ml) and rat high-density lipoprotein (HDL; density 1.063-1.210 g/ml) were isolated by density gradient centrifugation and radioiodinated with carrier-free ¹²⁵I as described previously (Bijsterbosch et al., 2000). Less than 2% of the radioactivity in the labeled protein preparations was trichloroacetic acid-soluble.

Determination of Stability of ISIS-9389 in Rat Serum and Plasma. [³H]ISIS-9389 was incubated at 37°C at a concentration of 20 µg/ml with rat serum or EDTA-plasma (4 mM EDTA). After 90 min, aliquots of 200 µl of the incubation mixtures were mixed with an equal volume of extraction buffer (25 mM Tris-HCl buffer, pH 8.0, containing 25 mM EDTA, 100 mM NaCl, 0.5% Nonidet P-40, and 1 mg/ml proteinase K), and incubated for a further 2 h at 56°C. Subsequently, the samples were mixed with 400 µl of phenol/isoamyl alcohol/chloroform (25:1:24, by volume). After shaking for 10 min, the phases were separated by centrifugation. The organic phase was washed four times with 400 µl of water. The aqueous phases were combined (total extraction efficiency approximately 40%), and dried in a Speed-Vac concentrator. The residues were dissolved in water, and 30 µg of unlabeled ISIS-9389 was added as marker (final volume 600 µl). An aliquot of 500 µl was subjected to reversed phase high-performance liquid chromatography as described above. Fractions of 1 ml were collected and assayed for radioactivity. It was found that after 90 min of incubation of [³H]ISIS-9389 with rat serum or plasma, >95% of the radioactivity eluted at the position of the unlabeled ISIS-9389 marker. Because the retention of ISIS-9389 depends on the presence of cholesterol, this indicates that the radiolabeled oligonucleotide still contains both cholesterol moieties, implicating that it was fully intact.

Determination of Plasma Clearance and Tissue Distribution. Male Wistar rats, weighing between 200 and 350 g, were used. The animals were anesthetized by intraperitoneal injection of sodium pentobarbital (60 mg/kg of body weight), and the abdomen was opened. Radiolabeled oligonucleotide, dissolved in phosphate-buffered saline (10 mM sodium phosphate buffer, pH 7.4, containing 0.15 M NaCl), was injected via the vena penis (2 ml/kg of body weight). At the indicated times, blood samples of 0.2 to 0.3 ml were taken from the inferior vena cava and collected in heparinized tubes. The samples were centrifuged for 2 min at 16,000g, and the plasma assayed for radioactivity. The total amount of radioactivity in plasma was calculated using the equation plasma volume (ml) = [0.0219 × body weight (g)] + 2.66 (Bijsterbosch et al., 1989). At the indicated times, liver lobules were tied off and excised, and at the end of the experiment the remainder of the liver was removed. The amount of liver tissue tied off successively did not exceed 15% of the total liver mass. The amount of radioactivity in the liver at each time point was calculated from the radioactivities and weights of the liver samples. Uptake by extrahepatic tissues was determined by removing the tissues at the end of the experiment and counting of radioactivity. Radioactivity in tissues was corrected for radioactivity in plasma present in the tissue at the time of sampling (Bijsterbosch et al., 1989).

Pharmacokinetic Analysis. The plasma clearance of intravenously injected radiolabeled oligonucleotide was analyzed by a nonlinear regression program (GraphPad; ISI Software, San Diego, CA). The data were fit by a two-compartment model. The distribution volume (V_dis) was calculated by extrapolation of the elimination curve to time 0. The half-life of elimination was calculated from the elimination rate constant (k_e) using the formula T_1/2= 0.693/k_e. The total plasma clearance (CL) was calculated using the formula CL = V_dis × k_e.

Determination of Distribution over Liver Cell Types. Rats were anesthetized and injected with radiolabeled oligonucleotides as described above. The liver was perfused at 60 min after injection, and parenchymal, Kupffer, and endothelial cells were isolated from the liver as described in detail previously (Nagelkerke et al., 1983). The cell fractions were assayed for radioactivity and protein. Shortly before separation of the cells, a liver lobule was tied off and excised to determine the total liver uptake. The contributions of the various cell types to the total liver uptake was calculated from the uptake per milligram of cell protein and the contribution of each cell type to the total liver protein (Nagelkerke et al., 1983). As found with other ligands (Nagelkerke et al., 1983; Bijsterbosch et al., 1989), no significant amounts of radioactivity were lost from the cells during the isolation procedure. This was checked in each experiment by comparing the calculated liver uptake (i.e., the summation of the contributions of the various cell types) with the value actually measured in the liver lobule. The percentage of the dose taken up by each cell type was calculated from the contribution of the cells to the total liver uptake and the contribution of the liver to the clearance of ISIS-9389. The overall intracellular concentrations were calculated from the molecular weight of the oligonucleotide, liver weight (4.3 ± 0.1% of body weight, mean ± S.E.M. of 10 determinations), liver density (1.07 mg/ml; Blouin et al., 1977), and the volumes of the different cellular compartments in the liver (Blouin et al., 1977). Furthermore, it was assumed that 75% of the cellular volume consists of water.

Determination of Association of ISIS-9389 with Plasma Components. [³H]ISIS-9389 was incubated at 37°C with rat plasma. For comparison, [³H]ISIS-9388 and [³H]ISIS-3082 were also incubated with rat plasma. After 30 min, aliquots of the incubation mixtures were injected onto a Superose 6 Precision column (3.2 × 300 mm), equipped with a 50-µl sample loop (Amersham Biosciences AB, Uppsala, Sweden). The column was eluted with phosphate-buffered saline at a flow rate of 50 µl/min. Fractions of 100 µl were collected and assayed for radioactivity. To determine association with lipoproteins, [³H]ISIS-9389 was incubated at 37°C with rat ¹²⁵I-LDL or rat ¹²⁵I-HDL, dissolved in phosphate-buffered saline. The incubation mixtures were analyzed as described above.

Determination of Proteins. Protein concentrations in cell suspensions and preparations of LDL and HDL were determined by the method of Lowry et al. (1951), with a bovine serum albumin standard.

Determination of Radioactivity. Samples containing ³H were counted in a Packard Tri-Carb 1500 liquid scintillation counter. Liquid samples were counted without further processing by liquid scintillation spectroscopy, using Emulsifier Safe or Hionic Fluor scintillation cocktails. Tissue samples were processed using a Packard 306 sample oxidizer. Some tissues (e.g., bone) were dissolved in 10 M NaOH at 95°C before counting. In samples containing both ¹²⁵I and ³H, the ¹²⁵I radioactivity was counted in a Packard Auto-Gamma 5000 counter. The ³H radioactivity was subsequently measured as described above and corrected for the contribution of ¹²⁵I radioactivity.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Plasma Clearance and Tissue Uptake of ISIS-9389. A 3',5'-bis-cholesteryl-conjugated PS-ODN specific for murine ICAM-I, denoted ISIS-9389, was used as test compound (Table 1). The unconjugated and 3'-cholesteryl-conjugated counterparts (ISIS-3082 and ISIS-9388, respectively) were used for comparison (Bijsterbosch et al., 1997, 2000). The disposition of ISIS-9389 was studied after an intravenous bolus injection of the radiolabeled oligonucleotide into rats. The dose, 1 mg/kg of body weight, was in the range of doses of ICAM-1-directed antisense oligonucleotides that have been found to be effective in preclinical models and in patients (Stepkowski et al., 1994, 1998; Bennett et al., 1997; Yacyshyn et al., 1998). Figure 1 shows the plasma clearance of radioactivity after injection of [³H]ISIS-9389. After an initial rapid distribution phase, radioactivity was cleared from the circulation with a half-life of 23.6 ± 0.3 min. The plasma clearance of ISIS-9389 was followed for 90 min. In vitro incubation studies with rat plasma and serum indicate that ISIS-9389 remains for >95% intact during this time period. In Table 2, the pharmacokinetic parameters of ISIS-9389 are compared with those of ISIS-3082 and ISIS-9388. The distribution volumes of the three oligonucleotides were not significantly different. ISIS-9389 was cleared from the circulation at approximately the same rate as the parent compound ISIS-3082, but more rapidly than the mono-cholesteryl derivative ISIS-9388.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Plasma clearance of intravenously injected [3H]ISIS-9389. Rats were intravenously injected with [3H]ISIS-9389 at a dose of 1 mg/kg of body weight. Blood samples were taken at the indicated times, and the radioactivity in the plasma was determined. Values are means ± S.E.M. of three rats.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Comparison of tissue uptake of intravenously injected [3H]ISIS-3082, [3H]ISIS-9388, and [3H]ISIS-9388. Rats were intravenously injected with [3H]ISIS-3082 (), [3H]ISIS-9388 (), or [3H]ISIS-3082 (), all at a dose of 1 mg/kg of body weight. The distribution of radioactivity over all tissues was determined at 90 min (ISIS-9389 and ISIS-3082) or 180 min (ISIS-9388) after injection. Radioactivity in the tissues is expressed as the percentage of the radioactivity cleared from the circulation at the time of sampling, and constitutes the contribution of each tissue to the clearance. At the time of sampling, 97.9 ± 0.3, 88.9 ± 2.3, and 94.4 ± 0.6% of the injected dose of ISIS-3082, ISIS-9388, and ISIS-9389, respectively, had been cleared. Values are means ± S.E.M. of three rats.

Cellular Distribution of ISIS-9389 in Liver. The liver contains several actively endocytosing cell types (Ashwell and Harford, 1982; Smedsrod et al., 1994). To identify the cell type(s) responsible for the hepatic uptake of ISIS-9389, rats were injected with the radiolabeled oligonucleotide. Parenchymal, endothelial, and Kupffer cells were isolated from the liver 60 min later, and assayed for radioactivity. The cell isolation procedure was performed at a low temperature (8°C) to prevent processing of internalized oligonucleotide. The results are shown in Table 4. Endothelial cells were the major site of uptake in the liver (51.9 ± 6.4% of the liver uptake), whereas Kupffer and parenchymal accounted each for approximately 25% of the liver uptake. When all injected oligonucleotide is cleared from the circulation, 87.7 ± 0.8% of the injected dose is taken up by the liver. It can thus be calculated that endothelial, Kupffer, and parenchymal cells accumulate 45.5 ± 5.6, 20.4 ± 2.2, and 21.8 ± 6.7% of the injected amount of oligonucleotide, respectively. Intracellular concentrations of ISIS-9388 in endothelial, Kupffer, and parenchymal cells can be calculated from these data and from the sizes of the three different cellular compartments. Parenchymal and Kupffer cells contribute to a similar extent to the hepatic uptake of ISIS-9389. However, parenchymal cells constitute >90% of the cellular mass and Kupffer cells only 2.5%. The concentration of ISIS-9389 in Kupffer cells is therefore much higher than in parenchymal cells (51.5 ± 5.5 versus 1.5 ± 0.5 µM). However, the highest concentration of ISIS-9389 is in endothelial cells (86.1 ± 10.6 µM).

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Uptake of intravenously injected [3H]ISIS-3082, [3H]ISIS-9388, and [3H]ISIS-9389 by liver cell types. Rats were intravenously injected with [3H]ISIS-3082 () [3H]ISIS-9388 (), or [3H]ISIS-9389 (), all at a dose of 1 mg/kg of body weight. After 60 min, parenchymal, endothelial, and Kupffer cells were isolated, and the association of radioactivity to each cell type was determined. The contribution of each cell type to the total liver uptake was calculated from the uptake per milligram of cell protein and the contribution of each cell type to the total liver protein (Nagelkerke et al., 1983). The percentage of the administered dose taken up by each cell type (when all oligonucleotide is cleared) was calculated from the contribution of each cell type to the total liver uptake and the contribution of the liver to the clearance (41.4 ± 1.4, 71.7 ± 3.7, and 87.7 ± 0.8% of the dose for ISIS-3082, ISIS-9388, and ISIS-9389, respectively). Values are means ± S.E.M. of three to four rats.

Implication of Scavenger Receptors in Liver Uptake of ISIS-9389. We demonstrated previously the role of scavenger receptors in the hepatic uptake of ISIS-3082 and ISIS-9388 (Bijsterbosch et al., 1997, 2000). Endothelial liver cells, and to a lesser extent Kupffer cells, express the type AI/II scavenger receptor (SR-AI/AII), which binds and internalizes a variety of polyanionic ligands (Nagelkerke et al., 1983; Krieger and Herz, 1994; van Berkel et al., 1998). Interaction of ligands with SR-AI/AII can be effectively inhibited by polyinosinic acid (poly-I), whereas polyadenylic acid (poly-A) is a poor inhibitor (Pierson et al., 1993). To study the possible role of SR-AI/II in the liver uptake of ISIS-9389, rats were preinjected with poly-I or poly-A shortly before injection of [³H]ISIS-9389. The uptake of ISIS-9389 by the liver was substantially inhibited by poly-I (Fig. 4). Poly-A also inhibited the hepatic uptake of ISIS-9388, but it was less effective than poly-I. These findings suggest that scavenger receptors play a major role in the hepatic uptake of ISIS-9388. The inhibition of uptake of the oligonucleotide by both poly-I and poly-A suggests that, in addition to SR-AI/II, alternative scavenger receptor systems are implicated.

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Liver uptake of [3H]ISIS-9389; effects of polyanions. Rats were intravenously injected with [3H]ISIS-9389 at a dose of 1 mg/kg of body weight. Shortly (1 min) before the injection of the labeled ligand, the animals received 10 mg/kg polyinosinic acid (poly-I; ), 10 mg/kg polyadenylic acid (poly-A; ), or an equal volume of saline solvent (; 2 ml/kg). At the indicated times, the amounts of radioactivity in the liver were determined. Values are means ± S.E.M. of three to four rats.

Association of ISIS-9389 with Plasma Proteins. PS-ODNs bind to plasma proteins, which is likely to affect their disposition (Cossum et al., 1993; Srinivasan et al., 1995; Nolting et al., 1997). We found that the association of cholesteryl-derivatized ISIS-9388 with plasma proteins differs from that of nonconjugated ISIS-3082 (Bijsterbosch et al., 2000). To study interaction of ISIS-9389 with plasma proteins, [³H]ISIS-9389 was incubated with rat plasma (at 20 µg/ml; the concentration immediately after injection). After 30 min, the mixture was subjected to size exclusion chromatography. For comparison, [³H]ISIS-3082 and [³H]ISIS-9388 were also incubated with plasma and subjected to size exclusion chromatography. Figure 5 shows the results. The chromatographic profile of [³H]ISIS-9389 was clearly different from those of [³H]ISIS-3082 and [³H]ISIS-9388. All three oligonucleotides were protein-bound, because no radioactivity was found at the position of free oligonucleotide. [³H]ISIS-9389 was predominantly recovered in fractions eluting at 0.90 to 1.50 ml. These fractions contain lipoproteins and high molecular weight plasma proteins (molecular weights >300,000), such as ₁-macroglobulin (Rump et al., 2000). Only a small proportion (15-20%) was recovered in fractions eluting at 1.50 to 1.90 ml, which contain the bulk of the plasma proteins.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Association of [3H]ISIS-3082, [3H] ISIS-9388, and [3H]ISIS-9389 with rat plasma components. [3H]ISIS-3082 (), [3H]ISIS-9388 (), or [3H]ISIS-9389 () were incubated at 37°C with rat plasma, at a concentration of 20 µg/ml. After 30 min, aliquots of the incubation mixtures were subjected to size exclusion chromatography on a Superose 6 column. Fractions of 0.1 ml were collected and assayed for radioactivity. The results are expressed as percentages of the recovered amounts (recoveries >95%). The void volume of the column, and the elution volumes of LDL, HDL, serum albumin (SA), and free oligodeoxynucleotide (ODN) are indicated by arrows.

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Association of [3H]ISIS-9389 with low-density lipoprotein and high-density lipoprotein. [3H]ISIS-9389 (20 µg/ml) was incubated at 37°C with 0.2 mg/ml rat 125I-LDL (A) or 1.0 mg/ml rat 125I-HDL (B). After 30 min, aliquots of the incubation mixtures were subjected to size exclusion chromatography on a Superose 6 column. The fractions (0.1 ml) were assayed for 3H radioactivity () and 125I radioactivity (O). The results are expressed as percentages of the recovered radioactivity (recoveries >80%). The elution volumes of LDL, HDL, and free oligodeoxynucleotide (ODN) are indicated by arrows.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we found that the conjugation of two cholesteryl residues to a PS-ODN results in almost complete hepatic uptake of the oligonucleotide in the rat. Spleen and bone marrow, which constitute together with the liver the reticuloendothelial system, also actively accumulated the bis-cholesteryl-conjugated oligonucleotide. Liver, spleen, and bone marrow together accounted for >95% of the clearance. Kidneys, which play a major role in the disposition of unconjugated oligonucleotide (Bijsterbosch et al., 1997), accumulated only a minor amount (<0.5% of the dose). It was found previously that conjugation of a single cholesterol at either end substantially increases the hepatic uptake of PS-ODNs (Crooke et al., 1996; Bijsterbosch et al., 2000). We show herein that an even higher targeting to the liver can be achieved by conjugation of cholesterol to both ends.

Within the liver, endothelial cells contribute most (51.9 ± 6.4%) to the uptake, whereas parenchymal and Kupffer cells accounted each for approximately 25% of the hepatic uptake. Endothelial cells contained the highest concentration of ISIS-9389 (86.1 ± 10.6 µM). Although parenchymal and Kupffer cells accumulated approximately equal amounts of ISIS-9389, the concentration of ISIS-9389 in Kupffer cells was much higher than in parenchymal cells (51.5 ± 5.5 versus 1.5 ± 0.5 µM), because Kupffer cells constitute a much smaller cellular compartment than parenchymal cells. We showed earlier that conjugation of a single cholesterol to ISIS-3082 induces a shift in the intrahepatic distribution of the oligonucleotide. Compared with the unconjugated ISIS-3082, uptake of the 3'-cholesteryl-conjugated ISIS-9388 by Kupffer cells increased approximately 5-fold (Bijsterbosch et al., 2000). The present data indicate that derivatization with a second cholesterol does not appreciably affect the intrahepatic distribution, because the distribution of ISIS-9389 over liver cell types is very similar to that of ISIS-9388.

We reported previously that scavenger receptors are implicated in the hepatic uptake of ISIS-3082 and ISIS-9388 (Bijsterbosch et al., 1997, 2000). The term scavenger receptor is used for a variety of proteins that bind negatively charged ligands. Six classes of receptors are distinguished within the scavenger receptor family (Krieger, 1997; Greaves et al., 1998; Terpstra et al., 2000). Class A receptors were the first to be identified and cloned and are the best characterized. Endothelial cells, and to a lesser extent Kupffer cells, abundantly express the SR-AI/AII. These receptors bind a wide variety of polyanionic ligands, including modified (lipo)proteins, polynucleotides, and polysaccharides (Krieger and Herz, 1994). The three-dimensional structure of polynucleotides is an important determinant for their affinity for SR-AI/AII (Pearson et al., 1993; van Berkel et al., 1998). Poly-I binds tightly to the receptors, whereas poly-A is poorly bound. The different abilities of poly-I and poly-A to inhibit the interaction of ligands with SR-AI/AII were used to examine the role of SR-AI/AII in liver uptake. Poly-I and poly-A both inhibited the liver uptake of ISIS-9389, albeit that the latter was less effective. This indicates that at least part of the hepatic uptake occurs via scavenger receptors that are sensitive to poly-A inhibition. However, the molecular nature of these receptors remains to be established. Several candidate receptors have been described. Endothelial cells of various species have been found to express scavenger receptors that are different from the SR-AI/AII (Krieger, 1997; van Berkel et al., 1998; Greaves et al., 1998; Terpstra et al., 2000). Furthermore, it was found that Kupffer cells express scavenger receptors that recognize oxidized LDL and are likely related to macrosialin and CD68 (van Velzen et al., 1997).

The different biological behavior of ISIS-9389, compared with ISIS-3082 and ISIS-9388, may be due to the different binding of the oligonucleotides to plasma proteins. Analysis by size exclusion chromatography indicates that the binding of ISIS-9389 to plasma proteins is different from that of ISIS-3082 and ISIS-9388. ISIS-9389 elutes predominantly with high molecular weight proteins (molecular weights >300,000). The plasma components that bind ISIS-9389 remain to be identified. ISIS-9389 binds to LDL and HDL, when incubated with these lipoproteins alone. However, it is not likely that ISIS-9389 is also associated with lipoproteins in blood plasma. HDL is the major lipoprotein in rat plasma, but we detected no appreciable binding of ISIS-9389 to HDL after incubation with rat plasma. There are also no indications that the bis-cholesteryl-derivatized ISIS-9389 induces cross-linkage of LDL and HDL. Analysis by size exclusion chromatography of complexes of ISIS-9389 with both lipoproteins showed that the complexes eluted at the same position as native LDL and HDL. It is, however, possible that ISIS-9389 induces multimer formation of other, as yet unidentified plasma components. Further experiments need to be performed to identify the plasma proteins involved in the binding of ISIS-9389.

Our results indicate that bis-cholesteryl conjugation has implications for the therapeutic application of PS-ODNs. Bis-cholesteryl-conjugated PS-ODNs are almost exclusively taken up by liver, spleen, and bone marrow. In the liver, by far the highest concentrations were found in the cells of the endothelial lining: Kupffer and endothelial cells. We did not examine the cellular localization of ISIS-9389 in spleen and bone marrow, but it is likely that ISIS-9389 accumulates in these organs also in reticuloendothelial cells. The selective accumulation of bis-cholesteryl-conjugated oligonucleotide in reticuloendothelial cells is beneficial when genes expressed in these cells are targeted. For example, ICAM-1 is up-regulated on Kupffer cells and liver endothelial cells under inflammatory conditions, which results in the harmful infiltration of neutrophils into the liver (van Oosten et al., 1995; Jaeschke et al., 1996). It has been shown that systemically administered unconjugated PS-ODNs specific for ICAM-1 reduce the adherence of neutrophils to the cells of the endothelial lining in the liver, and consequently exert a therapeutic effect (Wong et al., 1997). A higher cellular accumulation of these oligonucleotides, achieved by cholesterol conjugation, is expected to result in an improved therapeutic effect. Moreover, the reduced uptake by nontarget tissues, in particular kidneys, will reduce side effects. After administration of unconjugated PS-ODNs, kidneys contain the highest concentrations of PS-ODNs, which results in several morphological changes (Henry et al., 1999). Conjugation of two cholesteryl residues to PS-ODNS decreases their accumulation in kidneys approximately 50-fold and is consequently expected to substantially reduce side effects. It was shown previously that conjugation of PS-ODNs with a single cholesterol already substantially increases hepatic uptake and reduces extrahepatic uptake (Crooke et al., 1996; Bijsterbosch et al., 2000). However, conjugation of a second cholesterol results in an almost exclusive targeting to the liver, spleen, and bone marrow. Furthermore, bis-cholesteryl-conjugated PS-ODN is eliminated more rapidly from the circulation than mono-cholesteryl-conjugated PS-ODN, which reduces the risk of side effects in the circulation, such as complement activation and impaired coagulation (Henry et al., 1999). For extrahepatic targets, such as ICAM-1 in Crohn's disease or nonhepatic transplants (Stepkowski et al., 1994, 1998; Bennett et al., 1997; Yacyshyn et al., 1998), the benefits of bis-cholesterol modification are questionable. Several in vitro studies with extrahepatic cells demonstrated a higher antisense activity of cholesteryl-conjugated PS-ODNs, compared with nonconjugated PS-ODNs (Alahari et al., 1996; Epa et al., 1998). However, the present data indicate that in vivo the bis-cholesteryl-conjugated PS-ODNs are hardly taken up by extrahepatic tissues.

In conclusion, we show in this study that attachment of two cholesterol residues to a PS-ODN results in an almost complete hepatic uptake of the oligonucleotide when injected in the rat. The highest concentrations of the oligonucleotide were found in Kupffer cells and endothelial cells. In addition, we established the predominant role of scavenger receptors in the hepatic uptake. We further show that the bis-cholesterol conjugation affects the interaction with plasma proteins, which may explain its biological fate. The high and selective liver accumulation of the bis-cholesteryl-conjugated PS-ODNs is expected to result in higher in vivo efficacy against hepatic targets. Furthermore, the efficacy of bis-cholesteryl-conjugated PS-ODNs may improve by other mechanisms, such as more favorable intracellular trafficking.