Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

The ability of antisense oligonucleotides to interdict, sequence-specifically, the expression of pathogenic genes affords an exciting new strategy for therapeutic intervention (1–3). Oligonucleotides with physiological phosphodiester internucleotide bonds are rapidly degraded, predominantly by exonucleases. Numerous oligonucleotide analogs have therefore been synthesized to confer resistance toward nuclease activity (3). The phosphorothioate analog is the most extensively studied, and phosphorothioate oligodeoxynucleotides have been shown to be potent inhibitors of the expression of their target genes in vitro and in vivo (1,3). However, phosphorothioate oligodeoxynucleotides also bind avidly and nonspecifically to proteins, thus provoking a variety of non-antisense effects (4). Oligonucleotide analogs that do not bind to proteins are therefore expected to display less nonantisense side effects. However, protein binding also affects the in vivo disposition of oligonucleotides. Nonphosphorothioate oligonucelotide analogs generally do not bind to serum proteins, and are therefore rapidly cleared from the circulation, protein-bound phosphorothioate oligodeoxynucelotides circulate much longer (5,6).

The authors’ aim is to prolong the circulation time of nonphosphorothioate oligonucleotides, in order to increase the exposure of target cells to the oligonucleotides. The approach is to conjugate oligonucleotides with lipids, with the objective of inducing association of the oligonucleotide with long-circulating lipid particles (lipoproteins and lipoprotein-like particles). The biological fate of the particle-associated oligonucleotide will then be determined by the lipid carrier. The authors’ studies focused on low-density lipoprotein (LDL), the main cholesterol-transporting vehicle in human circulation. LDL is a spherical particle (diameter 23 nm), consisting of an apolar core of cholesteryl esters and triglycerides, which is surrounded by a shell of cholesterol and phospholipids (7). A large part of the surface is covered by apoprotein B100, which is recognized by specific LDL receptors. LDL is slowly cleared from the circulation (7,8), which makes it suitable for prolonging the circulation of associated oligonucleotides. A second reason to choose LDL as carrier is that it has been shown that a variety of tumor cell types (e.g., leukemic cells) internalize large amounts of LDL via the LDL receptor (9). Association of an oncogene-specific antisense oligonucleotide with LDL may lead to a higher uptake of the oligonucleotide by tumors, and consequently higher therapeutic efficacy.

This chapter describes the conjugation of various lipids to a c-myb-directed oligonucleotide. The association of the lipid-oligonucleotide conjugates (lipid-ODNs) with LDL is characterized, as well as the stability of the lipid-ODN-LDL complexes in vitro in rat plasma and in vivo in rats.

2 Materials

2.1 Synthesis of Activated Lipid Structure

  1. 1.

    Lithocholic acid.

  2. 2.

    Oleic acid.

  3. 3.

    3α-(oleoyloxy)-5β-cholanic acid.

  4. 4.

    3α,7α-ZnXoleoyloxy)-5β-cholanic acid.

  5. 5.

    3β-(oleoyloxy)-5-cholenic acid.

  6. 6.

    Pentafluorophenol.

  7. 7.

    Dicyclohexylcarbodiimide.

  8. 8.

    Cholesterol chloroformate.

2.2 Synthesis and Purification of 3′-Lipid-ODN

  1. 1.

    An 18-mer antisense oligonucleotide complementary to c-myb (10) (5′-GTG CCG GGG TCT TCG GGC-3′) was from Eurogentec (Seraing, Belgium) and had a zphosphorothioate linkages at the 5′ end and a C7-amino linker at the 3′-end.

  2. 2.

    [3H]-Labeled 3′-amine 18-mer antisense oligonucleotide, radiolabeled with 3H by heat-catalyzed exchange at the C8 positions of the purine nucleotides (see Note 1 and refs. 11,12).

  3. 3.

    [3H]ISIS-9388, a 3′-cholesteryl-conjugatedphosphorothioate oligodeoxynucleotide specific for murine intercellular adhesion molecule-1 (13) was kindly provided by Dr. M. Manoharan (ISIS Pharmaceuticals, Carlsbad, CA).

  4. 4.

    Na125I (carrier-free) and 3H2O were from Amersham (Amersham, UK).

  5. 5.

    Low-melting multipurpose agarose and Agarase from Pseudomonas atlantica were obtained from Boehringer (Mannheim, Germany).

  6. 6.

    Tween-20 was from Merck (Darmstadt, Germany).

  7. 7.

    Lithium perchlorate (LiClO4).

  8. 8.

    Acetone.

  9. 9.

    Dimethylformamide.

  10. 10.

    Dioxane.

  11. 11.

    N,N-Diisopropylethylamine (DIPEA).

  12. 12.

    Dichloromethane.

  13. 13.

    50 mM Triethylamine ammonium acetate, pH 7.0.

  14. 14.

    Agarase.

  15. 15.

    30 mM Bis-Tris, containing 10 mM EDTA, pH 6.5.

  16. 16.

    Acetonitrile.

  17. 17.

    45 mM Tris-borate buffer, containing 0.1 mM EDTA, pH 8.4.

2.3 Determination of Melting Temperatures

  1. 1.

    A 28-mer sense oligonucleotide (5′-CCA TGG CCC GAA GAC CCC GGC ACA GCA T-3′) was from Eurogentec and had a phosphodiester backbone.

  2. 2.

    Phosphate-buffered saline (PBS): 10 mM sodium phosphate buffer, pH 7.4, containing 0.15 M NaCl.

2.4 LDL and Lipid-ODNs

  1. 1.

    Human LDL: density 1.024–1.063 g/mL.

  2. 2.

    Hionic Fluor™ scintillation cocktail.

  3. 3.

    Superose 6 Precision chromatography column, 3.2 × 300 mm (Pharmacia).

2.5 Plasma Clearance and Liver Association

  1. 1.

    Male Wistar rats (180–230 g).

  2. 2.

    Ketamine hydrochloride was from Eurovet (Bladel, the Netherlands).

  3. 3.

    Hypnorm™ (0.315 mg/mL fentanyl citrate and 10 mg/mL fluanisone) was from Janssen-Cilag (Sauderton, UK).

  4. 4.

    Thalomonal™ (0.05 mg/mL fentanyl and 2.5 mg/mL droperidol) was from Janssen-Cilag.

  5. 5.

    Hionic Fluor and Emulsifier Safe™ scintillation cocktails and Soluene−350™ were from Packard (Downers Grove, IL).

  6. 6.

    Heparinized collection tubes.

3 Methods

3.1 Conjugation of Oligonucleotides with Lipid Structures

The authors synthesized a series of conjugates of lipids with a 18-mer antisense oligonucleotide complementary to the c-myb proto-oncogene. The oligonucleotide had a phosphodiester backbone with three phosphorothioate linkages at the 5′ end as protection against 5′-exonucleases. At the 3′ end, the oligonucleotide was provided with an amino group to enable conjugation. Figure 1 shows the structures of the lipids that were used for conjugation. To enable coupling to the 3′-amino-tailed oligonucleotide, the lipids were activated. Except cholesterol (ODN-4), all lipids were activated at their carboxylic acid functionalities with pentafluorophenol (see ref. 11 for full details of the synthesis of the activated lipid structures). Cholesterol chloroformate was used for the conjugation of cholesterol. Conjugation of the amino-terminated oligonucleotide with the activated lipids was performed in a mixture of water and the aprotic solvents, dimethylformamide and 1,4-dioxane (1:4:4; v/v/v), in which the organic base DIPEA was present to create mild basic conditions. After conjugation, the lipid-ODNs were purified. ODN-2, ODN-3, and ODN-4 (conjugates of oligonucleotide with litocholic acid, oleic acid, and cholesterol, respectively) were purified by reversed-phase high-performance liquid chromatography (RP-HPLC). Purification of conjugates of the oligonucleotide with the oleoyl steroid ester structures (ODN-5, ODN-6, and ODN-7) was not possible by RP-HPLC; no conjugated products were recovered. These lipid-ODNs were separated from unconjugated oligonucleotide by electophoresis in a 1% agarose gel, containing 0.1% Tween-20. Separation is accomplished by the formation of micelles of lipid-ODNs and Tween-20. The lipid-ODNs were retrieved from the gel by melting the gel, followed by digestion of the gel material by Agarase.

Fig. 1.
figure 1

Lipid structures of lipid-ODNs.

Full size image

3.1.1 Synthesis of Activated Lipid Structure

5β-Cholanic acid 3α-ol pentafluorophenyl ester, oleic acid pentafluorphenyl ester, 3α-(oleoyloxy)-5β-cholanic acid pentaflorophenyl ester, 3α,7β-bis(oleoyloxy)-5β-cholanic acid pentaflorophenyl ester, and 3β-oleoyloxy)-5-cholenic acid pentaflorophenyl ester were synthesized as described in full detail earlier by activating lithocholic acid, oleic acid, 3α-(oleoyloxy)-5β-cholanic acid, 3β,7β-bis(oleoyloxy)-5β-cholanic acid, and 3β-(oleoyloxy)-5-cholenic acid, respectively, with pentafluorophenol (11). The identity of all compounds was verified by mass spectrometry and nuclear magnetic resonance.

3.1.2 Synthesis and Purification of 3′-Lipid-ODN

  1. 1.

    The amino-terminated antisense oligonucleotide is precipitated as a Li-salt with 10 vol of 3% LiClO4 in acetone.

  2. 2.

    The oligonucleotide is subsequently dissolved in H2O and precipitated again with 10 vol acetone to remove final traces of LiClO4.

  3. 3.

    In a typical derivatization experiment, 15 nmol of oligonucleotide is dissolved in 350 μL H2O/dimethylformamide/dioxane (1:4:2, v/v/v). Subsequently, 1 μmol activated lipid, dissolved in 100 μL dioxane, and 35μmol DIPEA are added. For the preparation of ODN-2, ODN-3, and ODN-4, the mixtures were incubated for 48 h at 37°C, and, for the preparation of ODN-5, ODN-6, and ODN-7, the mixtures were incubated for 48 h at 56°C.

  4. 4.

    Solvents are removed in a Speed-Vac concentrator, and the residue is taken up in 200 μL dichloromethane and 200 μL H°O. The layers are separated by centrifugation, and the organic layer is washed twice with 200 μL H2O.

  5. 5.

    The aqueous phases are combined and freeze-dried.

  6. 6.

    Derivatization of (3H)oligonucleotide is performed as described in steps 1–5 with 2 nmol labeled oligonucleotide and 4 nmol unlabeled oligonucleotide.

  7. 7.

    Oligonucleotides conjugated with lithocholic acid, oleic acid or cholesterol are purified by RP-HPLC on a Waters C4 column (5 μm, 300 A, 300 × 3.9 mm) by applying a gradient of 1 % CH3CN/min in 50 mM triethyl ammonium acetate (pH 7.0) at a flow rate of 1 mL/min. The gradient (5–50%) is started after elution for 5 min at 5% CH3CN. Oligonucleotides are detected at 260 nm.

  8. 8.

    All other lipid-conjugated oligonucleotides are purified by gel electrophoresis in a 1% (w/v) low-melting multi purpose agarose gel, containing 0.1% Tween-20, at pH 8.4 (45 mM Tris-borate buffer, containing 0.1 mM EDTA). Gel slices containing lipid-ODNs are melted for 5 min at 65°C. The agarose is digested for 2 h with Agarase (40 U/mL gel) at 45°C in 30 mM Bis-Tris, containing 10 mM EDTA. The lipid-ODNs are precipitated with 10 vol of acetone. To remove traces of undigested agarose, the precipitate is taken up in 200 μL H2O and passed over a filter paper (no. 589, Schleicher and Scüll). Lipid-ODNs are isolated in a yield of 35–75%.

3.2 Interaction of Lipid-ODNs with Their Target Sequences: Determination of Melting Temperatures

To ascertain that the attached lipid structures do not interfere with the interaction of the antisense oligonucleotide with the target sequence, the authors determined the melting temperatures of antisense-sense hybrids. A 28-mer sense oligonucleotide was utilized with five overhang nucleotides at both the 3′- and 5′-end, to assess a possible interaction of the lipid with the single-stranded part of the hybrid. The melting temperatures of the hybrids with the conjugated ODNs differed not appreciably (maximally 2°C) from that of the hybrid with the unconjugated ODN-1 ( Table 1 ). Thus, the bulky steroid lipids (particularly the oleoyl steroid esters) do not significantly interfere with the overhang nucleotides of the target sequence. These findings are consistent with reports of other lipid-ODNs (14).

Table 1 Melting Temperatures of Hybrids of Antisense Lipid-ONDs and Sense Oligonucleotide
Full size table

Melting temperatures of hybrids of the antisense lipid-ODNs with an 28-mer sense oligonucleotide were determined using a Perkin-Elmer spectrophotometer equipped with a PTP-6 thermal programmer.

  1. 1.

    Equimolar amounts of both oligonucleotides are dissolved to a concentration of 3.6μMinPBS.

  2. 2.

    The mixtures are placed for 2 min at 96°C and subsequently slowly cooled to room temperature to allow annealing.

  3. 3.

    Then, the temperature is adjusted to 35°C and hybrids are melted by increasing the temperature to 95°C at a rate of 0.5°C/min (see Note 2 ).

3.3 Interaction of Lipid-ODNs with LDL

To examine which of the conjugated lipid structures is able to associate the oligonucleotide with LDL, radiolabeled lipid-ODNs were incubated for 2 h at 37°C in PBS with equimolar amounts of radioiodinated LDL. Aliquots of the incubation mixtures were analyzed by agarose gel electrophoresis at pH 8.8 in 75 mM Tris-hippuric acid buffer ( Fig. 2 ). The electrophoretic mobility (R f) of LDL in this gel matrix is 0.2; oligonucleotides (nonconjugated as well as lipid-conjugated) have (R f) values close to 1.0.

Fig. 2.
figure 2

Association of lipid-ODNs with LDL; analysis by agarose gel electrophoresis. Equimolar amounts of 125I-LDL and lipid-(3H)ODNs (1.7 μM), dissolved in PBS + 1 mMEDTA, pH 7.4, were incubated for 2 h at 37°C. Aliquots of the incubation mixtures were subjected to gel electrophoresis in a 0.75 % (w/v) agarose gel in 75 vaM Tris-hippuric acid buffer (pH 8.8). After electrophoresis, the gel was cut into slices that were assayed for 3H-radioactivity (O), and 125I-radioactivity (●). The results are expressed as percentages of the recovered radioactivities. Recoveries were >95%.

Full size image

Figure 2 shows that >95% of the [3H]ODN−4, [3H]ODN−5, and [3H]ODN−7 comigrated in the gel with radioiodinated LDL (Rf, 0.2). Thus, conjugation of the oligonucleotide with cholesterol or the oleoyl esters of lithocholic acid or cholenic acid induces spontaneous and almost complete association of the oligonucleotide with LDL. These steroids meet the structural requirements for LDL anchors as defined by Firestone et al. (15; see Note 3 ). When lipid-ODNs were associated at higher molar ratios, the increase in the electrophoretic mobility was more clear (data not shown).

Figure 2 also shows that conjugation of lithocholic acid and oleic acid (ODN-2 and ODN-3) did not induce association of the oligonucleotide with LDL at all. Only a proportion of the [3H]oligonucleotide conjugated to the bis-oleoyl steroid ester (ODN-6) comigrated with LDL. A substantial proportion of 3H-radioactivity migrated at (R f ) 0.2–1.0, which suggests that the lipid-ODN does associate with LDL, but that the complex slowly dissociates during electrophoresis. The lipid moiety of ODN-6 probably does not partition with its complete steroid structure into the lipids of LDL, but only with the two oleoyl chains. The complete bisoleoyl steroid ester structure may be too bulky to associate spontaneously with LDL. However, ODN-6 may well be incorporated in the recently developed artificial LDL-like carrier systems (16,17; ee Note 4 )

3.3.1 Isolation and Radioiodination of LDL

Human LDL (density 1.024–1.063 g/mL) was isolated from the serum of fasted volunteers by density gradient ultracentrifugation (18) and were dialyzed against PBS containing 1 mMEDTA. Radioiodination was performed atpH 10.0 with carrier-free 125I as described by McFarlane (19). Protein concentrations of the lipoproteins were determined by the method of Lowry et al. (20) with bovine serum albumin as standard.

3.3.2 Determination of Association of Lipid-ODNs with LDL

  1. 1.

    Equimolar amounts of 125I-labeled LDL and 3H-labeled lipid-ODNs (1.7 μM), dissolved in PBS containing 1 mMEDTA, are incubated for 2 h at 37°C (see Note 5 ).

  2. 2.

    Aliquots of the incubation mixtures are subjected to gel electrophoresis in a 0.75% (w/v) agarose gel in 75 mM Tris-hippuric acid buffer, pH 8.8.

  3. 3.

    After electrophoresis, the gel is cut into slices, and the 125I radioactivity is counted after addition of 0.5 mL Soluene-350.

  4. 4.

    The gel slices are allowed to dissolve for 24 h at room temperature.

  5. 5.

    Then, 3 mL Hionic Fluor is added and samples are counted for 3H-radioactivity. The measured values of 3H-radioactivity are corrected for the contribution of 125I-radioactivity.

3.4 Exchange of Lipid-ODNs from LDL to Plasma Proteins

The preceding subheading demonstrated that only the conjugates of the oligonucleotide with cholesterol or oleoyl steroid esters (ODN-4, ODN-5, and ODN-7) associate quantitatively with LDL. To examine the stability of the complexes of the lipid-ODNs with LDL in a biological matrix, the authors studied the exchange of the lipid-ODNs from preformed lipid-ODN-LDL complexes to components in rat plasma. Lipid-[3H]ODN-LDL complexes were incubated for 5 or 25 min with rat plasma and subsequently subjected to size-exclusion chromatography. The fractions were monitored for 3H-radioactivity, and the results are depicted in Fig. 3 . Although HDL is the main lipoprotein present in rat plasma, no significant redistribution of ODN-4, ODN-5 and ODN-7 from LDL to HDL (elution vol 1.57 mL) was seen. The lipid-ODNs redistributed to some extent to triglyceride-rich lipoproteins (elution vol 0.9 mL). The oleoyl steroid ester-conjugated oligonucleotides (ODN-5 and ODN-7) ( Fig. 3B , D ) appeared to be most stably complexed with LDL, because after 25 min of incubation, >70% of the radioactivity still was associated with the LDL fractions. At that time, only approx 50% of the cholesteryl-conjugated ODN (ODN-4) ( Fig. 3A ) was found to be associated with LDL.

The lipid-ODNs, ODN-4, ODN-5, and ODN-7, have a partially phosphorothioate (PS)-modified backbone (17% PS linkages). Many studies have found that PS linkages in oligonucleotides induce binding to serum proteins (22,23). Therefore, the authors also examined the stability of complexes of LDL and [3H]ISIS-9388, a full PS oligonucleotide that is conjugated at the 3′ end with cholesterol. Upon incubation with rat plasma, ISIS-9388 dissociated more rapidly from a preformed complex with LDL than the partially PS-modified lipid-ODNs. After 25 min incubation, less then 20% of ISIS-9388 was found to be associated with LDL ( Fig. 3D ; see Note 6 ).

Fig. 3.
figure 3

Chromatographic profiles after incubation of lipid-ODN-LDL complexes with rat plasma. Preformed lipid-ODN/LDL complexes were incubated with rat plasma for 5 min (O) or for 25 min (●) at 37°C. The samples were subsequently injected onto a Superose 6 column (Pharmacia) and the fractions were assayed for 3H-radioactivity. The 3H-radioactivity in the fractions are expressed as percentage of recovered radioactivity (recoveries were >95%). The gray zones indicate the fractions containing 90% of the lipid-ODN-LDL complex at t = 0 min. (A) ODN-4; (B) ODN-5; (C) ODN-7; (D) ISIS-9388.

Full size image

3.5 Determination of Exchange of Lipid-ODNs from LDL to Proteins in Plasma

  1. 1.

    Equimolar amounts of 125I-labeled LDL and 3H-labeled lipid-ODNs (200 pmol), dissolved in 100 μL of PBS containing 1 mMEDTA, are incubated for 2 h at 37°C.

  2. 2.

    Subsequently, 50 μL of the mixtures are injected onto a Superose 6 Precision Column (3.2 × 300 mm) (Smart System, Pharmacia) and eluted at a flow rate of 50 μL/min with PBS.

  3. 3.

    50 μL Fractions are collected and the three main fractions containing the lipidODN-LDL complexes are pooled.

  4. 4.

    50 μL Aliquots of the pooled fractions are subsequently incubated at 37°C with 40 μL citrated rat plasma.

  5. 5.

    After incubation, the mixtures are injected onto the Superose 6 column and the column is eluted as described previously. The fractions are assayed for 3H-radioactivity.

3.6 Behavior of Lipid-ODN-LDL Complexes In Vivo

To establish which of the lipid steroid structures is the most effective LDL anchor in vivo, the authors examined in the rat the plasma clearance of complexes of LDL and lipid-(3H)ODNs, ODN-4, ODN-5, and ODN-7. When injected without LDL, these lipid-ODNs were rapidly cleared from the circulation ( Fig. 4A , C , E ). At 5 min after injection, more than 95% of the dose was cleared, and a significant proportion of the radioactivity (15–40% of the dose) was recovered in the liver ( Fig. 4B , D , F ). The plasma clearance of the lipid-ODNs complexed with LDL was studied utilizing double-labeled complexes (lipid-(3H)ODN/(125I)LDL), which allows monitoring of both the lipid-ODN and the LDL carrier.

Fig. 4.
figure 4

Plasma clearance and liver association of lipid-ODNs and lipid-ODN-LDL complexes. Rats were injected with the free lipid-[3H]ODNs (▲) or lipid-(3H)ODN/ (125I)LDL complexes (3H,O; 125I, ●). At the indicated times, the amounts of radioactivity in plasma (A,C,E) and liver (B,D,F) were determined. Values are means ± SEM of two (free lipid-ODNs) or three (ODN-LDL complexes) separate experiments. A and B: ODN-4; C and D: ODN-5; E and F: ODN-7.

Full size image

Fig. 4A , C , E show that (125I)LDL was slowly cleared from the circulation, with a concomitant low liver uptake. These findings are consistent with the previously reported half-life of 5–6 h for LDL in the rat (8). The lipid-ODNs in the lipid-[3H]ODN/[125I]LDL complexes were cleared much slower than the noncomplexed lipid-ODNs ( Fig. 4A , C , E ). At 5 min after injection of the LDL-compexed lipid-ODNs, only 51, 39, and 24% of ODN-4, ODN-7, and ODN-5, respectively, had been cleared from the circulation (vs >95% for the noncomplexed lipid-ODNs). Thus, when complexed with LDL, both oleoyl steroid ester-conjugated oligonucleotides (ODN-5 and ODN-7) were more slowly cleared than the cholesteryl-conjugated ODN-4. Accordingly, the plasma area under the curve (AUC) of the LDL-complexed ODN-4 (1.49 ± 0.37 μg/min/mL) was significantly lower that the LDL-associated ODN-5 and ODN-7 (6.82 ± 1.34 μg/min/mL and 4.61 ± 0.38 μg/min/mL, respectively). The noncomplexed lipid-ODNs had plasma AUCs < 0.4 μg/min/mL. Compared to the free lipid-ODNs, the liver uptake of LDL-complexed lipid-ODNs was reduced ( Fig. 4B , D , F ). However, the lipid-ODN-LDL complexes were not completely stable, since the clearance of the lipid-ODNs did not completely resemble the clearance of LDL ( Fig. 4A , C , E ). The reduction of the clearance rate of the lipid-ODNs, achieved by complexation with LDL, was most evident for the oleoyl steroid ester-conjugated ODNs (ODN-5 and ODN-7). The cholesteryl-conjugated ODN (ODN-4) displayed the highest leakage from LDL.

3.6.1 Determination of Plasma Clearance and Liver Association

  1. 1.

    Male Wistar rats are anesthetized by subcutaneous injection of a cocktail of ketamine-HCl, fentanyl, droperidol, and fluanisone (75,0.04,1.1, and 0.75 mg/kg, respectively).

  2. 2.

    The abdomen is opened. Free lipid-[3H]ODNs (4 μM in PBS), or complexes of lipid-[3H]ODNs with (125I)LDL (4 μM, complexes prepared as described above), are injected via the vena penis at a dose of 5 μg lipid-ODN/kg body wt.

  3. 3.

    At the indicated times, 250 μL blood samples are taken from the inferior vena cava and collected in heparinized tubes. The blood samples are centrifuged for 10 min at 500g, and the plasma is assayed for radioactivity.

  4. 4.

    The total amounts of radioactivity in plasma are calculated using the equation: Plasma vol (mL) = [0.0291 × body wt (g)] + 2.54 (24)

  5. 5.

    At the indicated times, liver lobules are tied off and excised, and at the end of the experiment, the remainder of the liver is removed.

  6. 6.

    The amount of radioactivity in the liver at each time-point is calculated from the radioactivities and weights of the liver samples and is corrected for radioactivity in plasma present in the tissue at the time of sampling (85 μL/g fresh weight).

  7. 7.

    The plasma concentration-time AUC is calculated by computerized nonlinear fitting (Graphpad Prism, Graphpad Software, San Diego, CA).

3.6.2 Determination of Radioactivity

Samples containing 3H are counted in a Packard Tri-Carb 1500 liquid scintillation counter. Liquid samples are counted without further processing using Emulsifier Safe or Hionic Fluor scintillation cocktails. Agarose gel slices are first dissolved in Soluene-350. Tissue samples are processed using a Packard 306 Sample Oxidizer. Samples containing both 125I and 3H, are first assayed for 125I-radioactivity using a Packard Auto-Gamma 5000 counter. The 3H-radioactivity is subsequently measured as described above, and corrected for the contribution of 125I-radioactivity.

3.7 Discussion

This chapter describes the conjugation of an amino-terminated oligonucleotide with several lipid structures in solution phase. The conjugated lipid structures did not affect the association of the oligonucleotide with its target sequence, as judged by the lack of effects on the melting temperatures of the antisense-sense hybrids. Several of the lipid-ODNs, namely those conjugated with cholesterol or oleoyl steroid ester moieties, associate readily with LDL. The stability of the complexes of these lipid-oligonucleotides with LDL was investigated in vitro by incubation with rat plasma and in vivo in rats. Detailed reports of the experiments presented here were published earlier (11,25).

In vitro, the authors examined the exchange of the lipid-oligonucleotides from preformed lipid-ODN-LDL complexes to components in rat plasma. The more lipidic oleoyl steroid ester structures appeared to be better LDL-anchors than the cholesteryl moiety. ISIS-9388 (a 3′-cholesteryl-conjugated phosphorothioate oligonucleotide) redistributed to a much higher extent from LDL to plasma proteins than the partially phosphorothioate-modified lipid-ODNs (ODN-4, ODN-5, and ODN-7). This is likely to be the result of the high affinity of the full phosphorothioate oligonucleotide for plasma proteins (22,23).

When the lipid-ODNs were injected in rats without LDL, they were rapidly cleared from the circulation, meaning that lipid-conjugation alone is not sufficient to achieve a prolonged half-life in the circulation. A substantial amount of the lipid-ODNs was recovered in the liver, which may be ascribed to recognition by scavenger receptors on liver cells (13,26). When the lipid-ODNs were administered as complexes with LDL, the plasma clearance of the lipid-ODNs was considerably delayed and their liver uptake reduced. The oligonucleotides containing oleoyl steroid ester structures (ODN-5 and ODN-7) were more slowly cleared than the cholesteryl-conjugated ODN-4. The clearance and liver uptake of the LDL particles in the complexes was not altered, which indicates that complexation with the lipid-ODNs does not affect the integrity of the particle.

Taken together, of all the steroid structures tested, the lithocholic acid-3αoleate structure (present in ODN-5) most effectively reduced clearance of the oligonucleotide in vivo and the exchange from LDL in vitro. The strong association of this steroid ester lipid anchor with the lipids of LDL is probably primarily responsible for the prolonged circulation.The backbone chemistry of the oligonucleotide is also important. A modification with low affinity for plasma proteins (e.g., oligoncucleotide with little phosphorothioate linkages or nonphosphorothioate oligonucleotides, such as morpholino or peptide nucleic acid oligomers [27,28]), will reduce dissociation of lipid-conjugated oligonucleotides from lipoproteins. Oligonucleotides conjugated with the oleoyl steroid esters remain associated to LDL and the retarded plasma clearance results in a higher exposure to target cells. Oligonucleotides associated with LDL (or LDL-like particles [16,17]) may be taken up along with the particles via LDL receptors that are overexpressed on various types of tumor cells, e.g., leukemic cells (9). The firm association of lipid-ODNs with LDL or LDL-like particles thus has the additional advantage that enhanced exposure is accompanied by enhanced rate of uptake by LDL receptor-overexpressing tumor cells.

4 Notes

  1. 1.

    Radiolabeling of oligoncucleotides. The 3′-amine antisense oligonucleotide was radiolabeled with 3H by heat-catalyzed exchange at the C8 positions of the purine nucleotides as described previously (11,12). The radiolabeled oligonucleotide was stored at −20°C. No loss of radioactivity from the oligonucleotide was detected during 6 mo of storage at −20°C. The specific activity of the radiolabeled oligonucleotide was 80 × 106 dpm/mg.

  2. 2.

    Differences in melting temperatures measured in duplicate runs were <1°C.

  3. 3.

    Association of the lipid-ODNs with LDL also slightly broadened the band of LDL in the gel, probably because of an increase of the overall negative charge of the complex.

  4. 4.

    In these systems, lipid-ODNs can be incorporated during preparation of the particles, in contrast to the spontaneous association with LDL in this study.

  5. 5.

    For these experiments, the molecular weight of an LDL particle was related to the molecular weight of the apoprotein, B-100 (21) because each LDL particle contains only one copy of the apoprotein.

  6. 6.

    No radioactivity eluted at the column volume after incubation of ODN-4, ODN-5, or ODN-7 with plasma. This implies that no radiolabeled oligonucleotides were generated during the incubation, which indicates that the 3 PS-linkages at the 5′ end and the amino linker-steroid lipid structures at the 3′ end effectively protect the lipid-ODNs against nuclease activity in the plasma.