Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

When activated under physiologically challenging conditions, such as endotoxemia or immune reactions, macrophages release large amounts of cytokines, interleukins, and prostanoids, which may result in organ damage. Kupffer cells, the resident macrophages in the liver, play a major role in the pathogenesis of liver injury. It has been demonstrated that obliteration of Kupffer cells prior to administration of hepatotoxins prevents liver damage (Adachi et al., 1994; Ishiyama et al., 1995; Laskin et al., 1995). Tumor necrosis factor- (TNF-), a proinflammatory cytokine, exhibits pleiotropic effects on various cell types (Beutler and Cerami, 1986). Kupffer cells are the major producers of TNF- following exposure to lipopolysaccharide (LPS), the bacterial endotoxin (Decker, 1990). An overproduction of TNF- has been associated with the development of alcoholic liver injury (McClain and Cohen, 1989; Nanji et al., 1994; Kamimura and Tsukamoto, 1995), rheumatoid arthritis (Elliot et al., 1994), inflammatory bowel disease (Miurch et al., 1993), and septic shock (Michie et al., 1988). Antibodies that bind TNF- neutralize the effects of TNF- released by Kupffer cells in conditions such as ischemia reperfusion (Wanner et al., 1999) and experimental liver damage induced by chronic alcohol consumption (Iimuro et al., 1997).

In recent years, the use of antisense oligodeoxynucleotides (ASOs) has been an alternative approach to suppress the synthesis of specific proteins. ASOs contain sequences complementary to RNAs, which block or destroy the targeted mRNAs (Matteucci and Wagner, 1994; Tu et al., 1998). Recently, our laboratory developed a highly effective phosphorothioate-modified ASO, TJU-2755, against rat TNF- (Tu et al., 1998). Although TJU-2755 was highly effective in primary Kupffer cells ex vivo (>90% inhibition of LPS-stimulated TNF- production), the in vivo efficacy has not been determined.

The in vivo efficacy of any drug depends upon the efficiency with which it is delivered to the relevant cellular compartment in the target cells. Recently, we demonstrated that anionic liposomes (which entrap the ASO inside lipidic membranes) can be used as an efficient delivery vehicle for targeting ASOs to Kupffer cells (Ponnappa et al., 1998). Ninety minutes postintravenous injection, over 50% of liposome-encapsulated ASO was distributed in macrophage-rich organs, like the liver (40%) and spleen (10%); incorporation into other organs, such as muscle, heart, brain, lungs, kidneys, and testes, was minimal (Ponnappa et al., 1998). In the liver, over 65% of the ASO was found in Kupffer cells, accounting for a 200-fold enrichment of ASO in Kupffer cells versus that in the combined body tissues (Ponnappa et al., 1998).

In addition to targeting primarily TNF- producing cells, a second challenge in the delivery of ASOs lies in the release of the ASOs from the lysosomes, to allow for their entry into the nucleus, where the ASOs appear to have their main action (Chin et al., 1990; Spiller and Tidd, 1995; Ponnappa et al., 1998; Tu et al., 1998). The anionic liposomal preparations used in our previous studies (Ponnappa et al., 1998), although efficiently taken up by the macrophages, contained phosphatidyl choline, which renders the liposomes insensitive to the acidic pH (Ellens et al., 1984) in the lysosomal vesicles. pH insensitivity makes them unsuitable for cytoplasmic delivery of ASO, since, following phagocytosis, these liposomes remain trapped in the lysosomes until the contents are digested at acidic pHs by the powerful lysosomal hydrolases. Our initial studies with these pH-stable liposomal vesicles showed no effects of ASO TJU-2755 (B. C. Ponnappa, I. Dey, F. Zhou, Q.-n. Cao, and Y. Israel, unpublished data; also see Results). Hence, it was necessary to develop alternative formulations of anionic liposomes, which at low pHs would destabilize the endosomal/lysosomal barrier such that pharmacologically relevant concentrations of the ASO could be released into the cytosol and eventually to the nucleus. In our new formulation, we used phosphatidyl ethanolamine (PE), which is fusogenic at acidic pHs (Ellens et al., 1984) and cholesteryl hemisuccinate (CHEM) that stabilizes the liposomes at physiologic pH. Based on previous observations (Ellens et al., 1984; Connor and Huang, 1985; Straubinger, 1993; Tschaikowsky and Brain, 1994), it is expected that, upon endocytosis and acidification by a proton pump in the membrane, pH-sensitive liposomes fuse with the endosomal membrane and destabilize the endosomal compartment, resulting in the release of the contents into the cytosol. The mechanism of delivery of macromolecules into the cytoplasm by pH-sensitive liposomes has been reviewed (Straubinger, 1993). Rapid entry of ASOs from the cytoplasm into the nucleus has been reported (Fisher et al., 1993).

In the studies described here, the in vivo efficacy of anti-TNF- ASO, TJU-2755, was tested in rats. Following encapsulation in pH-sensitive liposomes, in vivo delivery was accomplished by intravenous injection. Rats were subsequently administered LPS, following which, plasma TNF- levels and the ability of liver slices incubated ex vivo to produce TNF- were determined. Results showed that ASO TJU-2755 effectively inhibited the ability of liver to produce TNF- and lowered plasma TNF- levels, demonstrating a therapeutic potential of this delivery system and the antisense molecule in inhibiting TNF- synthesis in vivo.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. All animal experiments were conducted in male Sprague-Dawley rats purchased from Harlan Sprague-Dawley, Inc. (Indianapolis, IN). Animals were maintained on laboratory chow. The body weights of the animals ranged from 250 to 300 g. The experimental protocol was approved by the Institutional Animal Care and Use Committee of Thomas Jefferson University (Philadelphia, PA).

Chemicals Cholesterol, CHEMS, dipalmitoyl phosphatidyl choline, and dipalmitoyl phosphatidyl glycerol were purchased from Sigma Chemical (St. Louis, MO). PE (transphosphatidylated from egg lecithin) was purchased from Avanti Polar Lipids, Inc. (Alabaster, AL). Radioisotope [-³²P]ATP was purchased from Amersham Pharmacia Biotech (Arlington Heights, IL). The phosphorothioate oligonucleotides used were custom-synthesized either from Geneset, Inc. (La Jolla, CA) or from Hybridon, Inc. (Milford, MA). Essentially two types of oligonucleotides were used in the present study, and both were 21 nucleotides long. The first one, ASO TJU-2755, had a sequence (5'-TGATCCACTCCCCCCTCCACT-3'), complementary to the 3'-untranslated region of rat TNF- mRNA (Tu et al., 1998). The second oligo, TJU-2755SS, was a "sense" oligonucleotide, complementary to TJU-2755 (Tu et al., 1998). All other chemicals used were of reagent grade, purchased either from Sigma Chemical or from Fisher Scientific (Pittsburgh, PA).

Preparation of Labeled Phosphorothioate ASO. Where indicated, the ³²P-ASO used in this study was labeled at the 5'-end with [-³²P]ATP using a DNA 5'-end labeling system kit from Promega (Madison, WI). Routinely, 200 ng of the oligo was incubated with [-³²P]ATP (3,000-5,000 Ci/mmol) and T4-polynucleotide kinase at 37°C for 30 min. The reaction was stopped by the addition of EDTA to a final concentration of 0.1 M and precipitated with ethanol overnight at 20°C. The precipitate was sedimented by centrifugation, washed twice with 80% ice-cold ethanol, dried, and dissolved in TE buffer (10 mM Tris and 1 mM EDTA, pH 8.0).

Preparation of ASO-Encapsulated pH-Sensitive Liposomes. An amphipathic lipid such as PE can form a stable bilayer only at pH greater than 9 or in media of low ionic strength. Consequently, the bilayer collapses in a physiologcal (pH) environment. A secondary amphipath, such as CHEMS, provides a conditional stability, such that the liposomes are stable at pH 7 to 8, but collapse in acidic environment (pH <6).

pH-Sensitivity of Liposomes. Liposomes were prepared as described encapsulating trace amounts of [³²P]TJU-2755 + 100 µg of the unlabeled ASO. Aliquots of the liposomal suspension were exposed to varying pH conditions, in the range of pH 5 to 7.4 at 37°C for 15 min. The suspensions were centrifuged at 100,000g for 45 min at 4°C to separate the supernatant from the liposomes. The amount of radioactivity released into the supernatant was expressed as percentage of the radioactivity that was present in the total suspension.

Liposomal Storage Stability and Size Determination Studies. To determine the stability of the liposomes upon storage, trace amounts of TJU-2755 were labeled with [-³²P]ATP as previously described and mixed with 2 mg of the unlabeled ASO. Liposomes were prepared as described and stored at 0 to 4°C. At various intervals up to 4 weeks, aliquots of the liposomal suspension were centrifuged at 100,000g for 45 min to separate the liposomes from the medium. The amount of radioactivity retained in the liposomes was compared with that present on the day of preparation, "day zero".

Stability of pH-Sensitive Liposomes in Plasma. Plasma was obtained from heparinized rat blood. Liposomes containing labeled TJU-2755 were prepared as previously described for the determination of pH stability. Aliquots of the liposomes were mixed with rat plasma and incubated at 37°C, to a final concentration range of 0 to 4 mg of lipid/ml. The final concentration of plasma was at least 90%. In the control group, plasma was replaced by PBS. The mixture was incubated at 37°C for 5, 30, and 60 min. At the end of incubation, the contents were diluted 10-fold with PBS and ultracentrifuged to separate the supernatant from the liposomes. The amount of radioactivity associated with the pellet, as well as the supernatant, was determined. The amount of radioactivity retained in the pellet was expressed as a percentage of the total amount of radioactivity that was present in the original liposomal suspension.

Preparation of pH-Stable Liposomes. pH-Stable liposomes were prepared by the reverse-phase method exactly as described for pH-sensitive liposomes, except that a mixture of dipalmitoyl phosphatidyl choline, dipalmitoyl phosphatidyl glycerol, and cholesterol (molar ratio of 4:1:5) was substituted for the lipids. Routinely, 15 mg of ASO and 25 mg of the lipid mixture were taken for the preparation of pH-stable liposomes. The encapsulation efficiency ranged from 16 to 20%.

In Vivo Studies. Liposomes encapsulated with TJU-2755 were i.v. injected (in a volume of PBS not exceeding 1 ml) into the animal by the tail vein route. In all of the experiments, the concentration of the liposomal lipid injected was maintained in the range of 15 to 17 mg/kg b.wt. Because the encapsulation efficiency varied from preparation to preparation, the amount of TJU-2755 injected also varied. However, unless indicated otherwise, the amount of TJU-2755 injected ranged between 1.5 to 2.0 mg/kg b.wt. Rats were injected single or multiple daily doses of the ASO, and sacrificed 24, 48, or 72 h postinjection, as per specific experiments detailed in the legends to the figures. Ninety minutes prior to sacrifice, animals were administered LPS (50 µg/kg b.wt.). Venous blood was collected in a heparinized tube 90 min postinjection and was kept on ice until processed for TNF-. Liver, and when necessary, spleen samples were taken for processing as described.

TNF- Secretion in the Liver Following the in vivo administration of LPS, the production of TNF- in the liver was assessed by measuring the amount of TNF- secreted by liver slices. Liver slices were prepared and incubated in culture medium according to the procedure described by Videla and Israel (1970), with modifications that include the addition of insulin and fetal calf serum to the culture medium. Briefly, liver slices of uniform thickness (10 × 4 × 0.4 mm) were prepared from the midlobe and initially rinsed in ice-cold Williams' E medium. The slices (2 slices/dish) were transferred to culture dishes (35 mm) containing 2 ml of fresh medium (90% Williams' E, 10% fetal calf serum, and 2 units of insulin/100 ml) and incubated at 37°C for 2 h. The amount of TNF- released into the medium was measured as described.

ELISA of TNF-. ELISA was conducted by using Cytoscreen KRC3012 kits (BioSource International, Camarillo, CA) according to the manufacturer's specifications. Supernatants containing high TNF- levels were diluted prior to the assay to ensure assay results within the standard curve.

Assay of ASO by Southern Hybridization

Extraction. The ASO TJU-2755, was extracted from the tissues and quantitated by Southern hybridization as described (Ponnappa et al., 1998). About 100 mg of the wet tissue was homogenized in an ice-cold buffer (50 mM Tris, pH 7.5, 10 mM EDTA), hereafter referred to as the homogenization buffer. Five-fold and 20-fold volumes of the homogenizing buffer were used for liver and spleen tissues, respectively. Subsequent digestion and extraction of the oligonucleotide was conducted according to the procedure of Temsamani et al. (1993), with modifications. Briefly, 100 µl of the homogenate were mixed with proteinase K (Sigma) solution (3.6 mg/ml) to a final volume of 150 µl. To this, an equal volume of an extraction buffer (0.5% SDS, 10 mM NaCl, 20 mM Tris-HCl, pH 7.6, 10 mM EDTA) was added and incubated at 55°C for 2 h. The digest (0.3 ml) was then mixed with an equal volume of the extraction solvent (phenol:chloroform:isoamyl alcohol; 25:24:1 v/v; Life Technologies, Gaithersburg, MD), vortexed vigorously, and centrifuged in a Microfuge for 2 min. The aqueous upper phase was transferred to a fresh tube and the organic lower phase was re-extracted once with 100 µl of TE buffer. The aqueous supernatants were combined and further extracted with an equal volume of chloroform. The ASO was precipitated (20°C/overnight) from the final aqueous extract by the addition of ethanol (70% v/v final) in the presence of 0.3 M sodium acetate (pH 5.3). The precipitate was sedimented by centrifugation, washed once with ice-cold 80% ethanol, dried, and dissolved in 100 µl of TE buffer.

Quantitation. The ASO TJU-2755 was separated by polyacrylamide gel electrophoresis (20% acrylamide/7 M urea), transferred to a nylon membrane (Hybond), and cross-linked (Stratalinker). The membrane was prehybridized in a sealed bag for 1 h at 40°C using Expresshyb Hybridization Solution (CLONTECH, Palo Alto, CA) and hybridized for 3 h in fresh buffer containing labeled (P-32) probe (TJU-2755ss). Routinely, the isotope concentration in the hybridization buffer was 2 to 3 × 10⁶ cpm/ml. Following hybridization, the membrane was washed in 5× SSC (10 min) and 5× SSC/0.05% SDS (5 min) at 40°C. The blotted membrane was placed between the folds of a Saran wrap and exposed to X-ray film overnight at 70°C. The autoradiograms were scanned (model JX-330, Sharp Corporation, Osaka, Japan, equipped with Image Scan software), and the densities associated with bands (contours) were compared. In each gel, along with the samples, varying concentrations of the standard TJU-2755 (2.5-20 ng) were also included. Concentration of the unknown was extrapolated from the standard curve.

Tissue Levels of TNF- mRNA by Northern Hybridization

Extraction of RNA. Liver and spleen samples (100 mg) were homogenized in 3 volumes of cold TRI Reagent (Molecular Research Center, Inc., Cincinnati, OH). The resultant homogenate was treated with chloroform for phase separation, precipitated with isopropanol, and washed with ethanol (70% v/v). The purified RNA pellet was resuspended in diethylpyrocarbonate-treated water and kept frozen until further analysis. The intactness of total RNA was measured by agarose (1%) gel electrophoresis by visualizing the intactness of ribosomal RNA (28S, 18S). Isolation of mRNA from total RNA was accomplished using Qiagen's Oligotex mRNA Spin-Column Midi kit, as per protocol. The purity of the isolated mRNA was monitored by analyzing the absorption ratios of 260 to 280 (nm), which always ranged from 1.8 to 2.0.

Quantitation of mRNA. Samples of mRNA were run on 1% agarose gel electrophoresis, vacuum-transferred (Bio-Rad model 785 Vacuum Blotter; Bio-Rad, Hercules, CA) onto a nitrocellulose membrane (Amersham), and crosslinked (Statagene UV Stratalinker 2400; Stratagene, La Jolla, CA). Subsequently, the membrane was prehybridized overnight at 50°C (PerfectHyb Plus Hybridization Buffer, Sigma). The next day, the DNA probe (322 bp polymerase chain reaction fragment from the open reading frame of rat TNF- gene) was labeled with ³²P using a random labeling system (Amersham Multiprime DNA Labeling System RPN.1601). Hybridization was performed overnight at 50°C in fresh hybridization solution + the labeled DNA probe (specific activity  10⁹ cpm/µg). The membrane was washed at 55°C, sequentially in 4× SSC, 2× SSC, 1× SSC/0.1% SDS, 0.2× SSC/0.1% SDS, and 0.1× SSC/0.1% SDS. Each wash lasted at least 10 min. The membrane was briefly air-dried and exposed to X-ray film to obtain the autoradiograms. The autoradiograms were quantitated using an ultrascan laser densitometer. The membrane was then stripped and rehybridized with cDNA probe for rat glyceraldehyde phosphate dehydrogenase used as the housekeeping gene. The intensities of the TNF- mRNA bands were normalized to values obtained with the housekeeping gene.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Size, pH-Sensitivity, and Long-Term Stability of pH-Sensitive Liposomes. In a typical preparation, 85% of the liposomes were 0.2 to 1.0 µm in diameter; 5% were <0.2 µm and 10% were >1 µm. The liposomes were stable at pH 7.4. As expected, when the pH was lowered, the liposomes became unstable, losing about 20% of the oligonucleotide content at pH 6 and 100% at pH 5.5 (Fig. 1). The liposomes displayed great stability at pH 7.4 at 4°C; when stored under these conditions for 4 weeks, over 95% of the ASO remained encapsulated in the liposomes (Fig. 2). The stability of liposomes in plasma was also tested. As shown in Fig. 3, the stability of liposomes was dependent upon the concentration of lipid in plasma; the higher the concentration, the greater the instability. At 2 mg of lipid/ml of plasma, >80% of the liposomes remained intact after 60 min incubation at 37°C. Since the concentration of liposomes attained in vivo after intravenous injection was in a much lower range (0.5-0.6 mg of lipid/ml of plasma), they are likely to remain intact during the period of rapid sequestration by the macrophages, similar to that reported for the anionic liposomes (Ponnappa et al., 1998).

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Stability of pH-sensitive liposomes. Liposomes encapsulated with trace amounts of labeled (32P) ASO TJU-2755 + 100 µg of unlabeled TJU-2755 were prepared by the reverse-phase method described under Materials and Methods. Aliquots of the liposomal suspension were exposed to varying pH values (pH 5.0-7.4) in PBS for 15 min at 37°C and centrifuged at 100,000g for 45 min. The amount of radioactivity (ASO) remaining in the liposomes was expressed as percentage of the total radioactivity present in the suspension at the beginning of the experiment, at pH 7.4. Values are the means ± S.E. of (n = 4) separate preparations.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Long-term stability of pH-sensitive liposomes. Liposomes were prepared in the presence of 2 mg of ASO TJU-2755 + trace amounts of [32P]TJU-2755 as described under Materials and Methods. The liposomal preparation was divided into several aliquots and stored at 4°C. At the times indicated, aliquots of the suspension were centrifuged at 100,000g to separate the supernatant from the pellet. The amount of radioactivity associated with the pellet was expressed as percentage of that present in the pellet on the day of the preparation (day "zero"). Values are the means of two separate preparations, each of which were prepared on different days.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Stability of pH-sensitive liposomes in plasma. Liposomes containing trace amounts of labeled (32P) ASO TJU-2755 was prepared as described in the legend to Fig. 1. Aliquots of the liposomal preparation were mixed with rat plasma (10% dilution of plasma) and incubated at 37°C for 60 min. The final concentration of liposomal lipid ranged from 0 to 4 mg/ml of plasma. After incubation, the suspensions were diluted with PBS and ultracentrifuged to separate the supernatant from the liposomes. The data points are the means of two closely agreeing values obtained from separate liposomal preparations.

In Vivo Efficacy of ASO TJU-2755. The efficacy of TJU-2755, administered in vivo, was assessed by two different methods. One of them was to measure the amount of TNF- produced by the liver tissue itself, and the other one was to determine plasma TNF- levels. In both cases, the animals were administered LPS prior to determination of TNF-. Based on preliminary observations, it was expected that there would be a time lag between the phagocytosis of liposomes by the macrophages and the entry of the oligonucleotides into the nucleus. Therefore, in the initial studies, the effect of a single intravenous administration of ASO TJU-2755 on LPS-induced production of TNF- was determined in liver slices. The studies showed that maximal efficacy (~30% inhibition) was observed at 48 h postinjection (Fig. 4). Therefore, in subsequent studies, the effect of multiple daily injections of TJU-2755 was tested 48 h after the last injection. As shown in Fig. 5, the extent of inhibition increased with the number of daily doses reaching 50% after two daily doses (p < 0.01). A more pronounced reduction (68%, p < 0.001, n = 8) in TNF- levels was observed in the corresponding plasma samples following two doses of ASO (Fig. 6). The plasma TNF- values (LPS treated) for rats administered "empty" liposomes at <17 mg of lipid/kg of b.wt., as in the present studies, were not significantly different from those for PBS-injected controls, but higher lipid concentrations showed inhibitory effects (data not shown).

View larger version (13K): [in this window] [in a new window]   Fig. 4.   In vivo efficacy of ASO TJU-2755 against TNF- in the liver: effect of duration postinjection. Rats were intravenously injected with pH-sensitive liposomes encapsulated with TJU-2755 (1.5-1.75 mg/kg b.wt.). At different times (24, 48, and 72 h) after ASO administration, LPS (50 µg/kg for 90 min) was administered i.v., and the animals were sacrificed 90 min after this challenge. Liver was excised, slices prepared, and incubated for 2 h as described under Materials and Methods. The amount of TNF- released into the medium was measured by ELISA. TNF- values (bars) from the ASO-treated animals were normalized to control values (control = 100) obtained from body weight-matched rats treated similarly, except that they were injected with "empty" liposomes. Values indicated are means ± S.E. of (n) determinations. Number of animals: 24 h, n = 3; 48 h, n = 4, 72 h, n = 3. *p < 0.05 (versus control).

View larger version (12K): [in this window] [in a new window]   Fig. 5.   In vivo efficacy of ASO TJU-2755 against TNF- in the liver: effect of multiple injections. Rats were injected either one or two daily doses of ASO (1.5-1.75 mg/kg b.wt.) encapsulated liposomes. Forty-eight hours after the last injection, they were administered LPS (50 µg/kg), and the animals were sacrificed 90 min after this challenge. The amount of TNF- produced by liver slices was determined as described in the legend to Fig. 4. Bars indicate the amount of TNF- produced by ASO-treated animals, expressed as percentage of controls injected with empty liposomes and are the means ± S.E. of (n = 4 for single dose and n = 7 for two doses). *p < 0.05, **p < 0.01 (versus control). The absolute values (mean) of TNF- for control and TJU-2755-treated livers (for two daily doses) were 17,748 and 7,500 (pg/g liver), respectively.

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Effect of multiple doses of ASO TJU-2755 on plasma TNF-. Rats were injected one (n = 3), two (n = 8), or three (n = 1) daily doses of TJU-2755 and administered LPS (50 µg/kg b.wt.) at 48 h after the last injection of the antisense formulation as described in the legend to Fig. 5. Ninety minutes after LPS injection, rats were sacrificed and blood was collected in heparinized tubes. Plasma TNF- was determined by ELISA. Bars indicate values from ASO-treated animals expressed as percentage of corresponding controls (animals dosed with empty liposomes) and are the means ± S.E. of (n) determinations. ***p < 0.001 between rats injected two daily doses of ASO and controls. The absolute (mean) values of TNF- for control and TJU-2755-treated plasma samples (for two daily doses) were 34,375 and 9,731 (pg/ml of plasma), respectively.

View this table: [in this window] [in a new window]   TABLE 1 pH-Stable liposomes and antisense action Rats were i.v. injected with two daily doses (1.5 mg/kg b.wt.) of ASO TJU-2755 encapsulated in pH-stable liposomes and challenged with LPS (50 µg/kg b.wt.) at 48 h postinjection. Animals were sacrificed 90 min post-LPS. The amount of TNF- in plasma, as well as that secreted by liver slices, was determined exactly as described for pH-sensitive liposomes under Materials and Methods and in Figs. 4 and 5. Body-weight-matched rats injected with two doses of "empty" liposomes and challenged similarly with LPS constituted the control group. All TNF- values were normalized to the corresponding control groups (=100). Values are expressed as the means ± S.E of (n) determinations.

Tissue Levels of ASO TJU-2755. Multiple doses of the ASO had cumulative effect on tissue levels of intact TJU-2755, as measured by Southern hybridization. As shown in Fig. 7, using pH-sensitive liposomes, at 48 h postinjection, a single dose (1.75 mg/kg b.wt.) of TJU-2755 resulted in a concentration of 5.8 µg/g of tissue of liver. Under similar conditions, two daily doses of the ASO led to an accumulation of 14.7 µg/g of tissue, and three daily doses had an accumulation of 32.3 µg/g of tissue.

View larger version (12K): [in this window] [in a new window]   Fig. 7.   Effect of multiple doses of ASO TJU-2755 on tissue levels of the oligonucleotide. Rats were injected one (n = 4), two (n = 5), and three (n = 1) daily doses of liposome-encapsulated ASO. Forty-eight hours after the last injection, they were administered LPS (50 µg/kg b.wt.). Ninety minutes after LPS injection, rats were sacrificed and a portion of the liver was processed for the extraction and estimation of the ASO as described under Materials and Methods.

TNF- mRNA Levels following the Administration of ASO TJU-2755. Since, it has been reported that binding of ASOs to the target mRNA transcripts initiated RNase H-mediated degradation of mRNA (Giles et al., 1995), we determined whether the inhibitory effects of TJU-2755 on LPS-induced TNF- production in vivo could also be due to reduced levels of TNF- mRNA, as reported earlier in isolated Kupffer cells (Tu et al., 1998). As shown in Fig. 8, in the liver tissue, there was a 35% reduction (p < 0.01) in the steady-state levels of TNF- mRNA in the group pretreated with TJU-2755, compared with controls (empty liposomes). As discussed later, it is reasonable to assume that LPS-induced changes in TNF- mRNA in the whole liver tissue reflects to a large extent, but not exclusively, the effects on Kupffer cells. Similar to the liver, the spleen also contains macrophages that are targeted by anionic liposomes (Ponnappa et al., 1998). For comparative purposes, TNF- mRNA was determined in the spleen following LPS administration and found to be similarly affected by TJU-2755, as shown by a 36.5% reduction (p < 0.05) in mRNA levels (Fig. 8).

View larger version (9K): [in this window] [in a new window]   Fig. 8.   Effect of ASO TJU-2755 in vivo on TNF- mRNA. Rats were injected with two daily doses (1.5 mg/kg b.wt.) of ASO and injected LPS (50 µg/kg b.wt.) 48 h post-ASO administration as described in the legend to Fig. 5. Liver (n = 5) and spleen (n = 4) samples were processed for the extraction of mRNA fractions. Steady-state levels of TNF- mRNA were determined by Northern hybridization and normalized to the levels of glyceraldehyde phosphate dehydrogenase mRNA. Bars indicate mRNA (means ± S.E.) of (n) values from ASO-treated samples, expressed as percentage of control values derived from rats treated with empty liposomes. *p < 0.05, **p < 0.01 (versus control).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The objective of this study was to assess the in vivo efficacy of antisense oligonucleotides to suppress LPS-induced production of TNF- in Kupffer cells. As previously discussed, the efficacy of any antisense molecule depends upon its ability to cross the lysosomal barrier and be released into the cytoplasm. Cationic liposomes have been successfully used for DNA transfection in cell culture systems. However, because of their almost complete instability in undiluted plasma or sera (typically, in the in vitro studies, serum is diluted 5-20%); their usefulness for in vivo delivery has been limited.

In an earlier study, we reported that pH-stable anionic liposomes could be used as an efficient delivery vehicle for targeting oligonucleotides to Kupffer cells in vivo (Ponnappa et al., 1998). However, the liposomes used in that study did not contain the pH-sensitive fusogenic lipid, phosphatidyl ethanolamine, and therefore did not possess the ability to destabilize the endosomal/lysosomal membrane barrier. As shown in Table 1, ASO TJU-2755 was ineffective when delivered in that fashion.

Macrophagic Kupffer cells contain scavenger receptors that recognize an array of negative charges, such as in anionic liposomes, which are efficiently sequestered by these cells (Bautista et al., 1994; Ponnappa et al., 1998). Sequestration by Kupffer cells is further facilitated by the large size of the liposomes (Alino et al., 1993). The method used in this study results in the formation of liposomes, in which about 90% of the liposomes had diameters larger than 200 nm, thus preventing them from crossing the fenestrations in the endothelial barrier of the liver sinusoids. The pH-sensitive liposomes used in the current study were formulated to both, to be taken up by Kupffer cells and to rapidly destabilize below pH 6 (Fig. 1). Consequently, fusion of the liposomes with the endosomal membrane would release the contents into the cytosol, much before the lysosomal enzymes become optimally active.

An important consideration in the formulation of anionic liposomes is their stability in plasma. Plasma proteins are known to have a high affinity for fatty acids and lipids, which can potentially destabilize the liposomes before they are sequestered by the tissues. However, as shown in Fig. 3, our data show that phosphatidyl ethanolamine-cholesteryl hemisuccinate-based liposomes are stable in undiluted plasma at concentrations of the liposomal lipid below 1 mg/ml. Therefore, under i.v. delivery conditions, at least 90% of these are likely to remain intact before they are sequestered by tissues. Furthermore, the liposomal preparation was also most stable at 4°C, losing only about 5% of its intraliposomal content after 4 weeks, a property that should allow its formulation in pharmaceutical preparations.

The in vivo efficacy of the antisense oligonucleotide was assessed by measuring plasma levels of TNF-, as well as by the ability of incubated liver slices to produce TNF- following LPS challenge. Whereas plasma TNF is a measure of LPS-induced secretions from various organs, including the liver and spleen, those secreted by liver slices primarily reflect TNF- secretions from Kupffer cells (Decker, 1990). Thus, data on liver slices essentially reflect the response of the Kupffer cells following liposomal delivery of ASO 2755. The in vivo effects of the ASO were time-dependent, exhibiting maximal inhibition of TNF- production at 48 h postinjection. We assume that the slow onset of inhibition is due to a slow release of the liposomal contents from the endolysosomal compartment.

Our studies on the effect of duration, postinjection, of antisense oligonucleotides (Fig. 4) show that following a single injection, the effect of the ASO begins to fade after 72 h, suggesting gradual degradation of the oligonucleotide. Although the exact half-life of TJU-2755 is not known, others have reported half-lives in the range of 48 h for similar phosphorothioate antisense oligonucleotides (Saijo et al., 1994). Consequently, multiple doses of the antisense were expected to be more effective. Indeed, successive injections of the ASO resulted in a cumulative increase in the tissue levels of the oligonucleotide, which was also associated with a significant inhibition of LPS-induced TNF- production, both in the liver and plasma (Figs. 5-7). Interestingly, with two daily doses of ASO TJU-2755, the effect was more pronounced in the plasma (68% inhibition) than in the liver (50% inhibition). Increasing the dose or the number of daily administrations is expected to display greater effects both on liver and plasma levels, because a plateau was not observed. Preliminary reports from our laboratory (G.-c. Tu, Q.-n. Cao, F. Zhou, and Y. Israel, unpublished observations) show that, similar to the liver slices (Kupffer cells), spleen slices (splenic macrophages) have a high capacity to produce TNF- in response to LPS stimulation, and therefore, can also raise plasma levels of TNF. Since the spleen is simultaneously targeted during the in vivo delivery of anionic liposomes (Ponnappa et al., 1998), spleen-associated production of TNF is also likely to be inhibited by the delivery of ASOs. Release of mediators from the spleen into the portal circulation appears to be relevant to the development of hepatocellular injury mediated by LPS. It has been reported that splenectomy significantly reduces the liver damage induced by the administration of large doses of LPS (Hiraoka et al., 1995).

In the liver tissue, the decrease (50%) in LPS-induced production of TNF- was also associated with a decrease (35%) in the levels of TNF- mRNA. This suggested that the reduced production of TNF- was due in part to the oligonucleotide-induced degradation of mRNA by antisense mechanisms. Additionally, it is possible that some of the ASOs may remain bound to the mRNA, thus preventing translation of TNF- mRNA without altering their levels. The specificity and the hydrolysis of TNF- mRNA by TJU-2755 have already been demonstrated in the extensive studies conducted by Tu et al. (1998) in primary cultures of rat Kupffer cells. Overall, our data show that by using ASO TJU-2755 encapsulated in pH-sensitive liposomes, it is possible to effectively inhibit endotoxin-induced production of TNF- in vivo.

An aspect that should be indicated is that the 68% reduction in plasma TNF- attained by the liposomal administration of TJU-2755 occurs despite the marked elevations in plasma levels of TNF- (35,000 pg/ml) following LPS administration. These levels are two orders of magnitude greater than those (10-200 pg/ml) present in patients with advanced liver disease stages (Felver et al., 1990; Sheron et al., 1991), in whom the levels are strongly predictive of mortality and are more in line with those present in patients with severe sepsis (Damas et al., 1989; Atici et al., 1997). These considerations suggest that liposome-mediated delivery of anti-TNF- antisense oligonucleotide may be very valuable in a number of conditions in which TNF- is a pathogenic mediator.

In summary, although several strategies/drugs have been developed to suppress the chronic in vivo production of TNF- to control inflammatory diseases, such as alcohol liver disease, rheumatoid arthritis, Crohn's disease, and septic shock, the discovery of an ideal drug still remains a challenge. Antibodies targeted against TNF- have been successfully used in some of the clinical trials, but the potential for antigenicity and toxicity remains (Eigler et al., 1997). Against this background, we believe that the effect of antisense oligonucleotides against TNF- demonstrated in this study is of pharmacologic significance.