improved leishmanicidal effect of phosphorotioate antisense oligonucleotides by ldl-mediated...

ELSEVIER Biochimica et Biophysica Acta 1264 (1995) 229-237

i

BB. Biochi~ic~a et Biophysica A~ta

Improved leishmanicidal effect of phosphorotioate antisense oligonucleotides by LDL-mediated delivery

Rakesh Kumar Mishra a,1 Catherine Moreau b Claude Ramazeilles a, Serge Moreau a Jacques Bonnet b, Jean-Jacques T o u l m 6 a,*

a INSERM U.386, Laboratoire de Biophysique Mol~culaire, Unicersit£ Bordeaux I1, 146 rue L~o Saignat, 33076 Bordeaux c~dex, France b INSERM U.8, Avenue du Haut L~t,~que, 33600 Pessac, France

Received 23 January 1995; revised 24 April 1995; accepted 30 June 1995

Abstract

We have designed antisense oligonucleotides that can interact with lipoproteins in order to use them as vectors to facilitate the uptake by those cells expressing the corresponding receptor. Phosphorothioate (PS) oligonucleotides were linked at the 5' end to a palmityl group giving rise to PSPal conjugates. Such a modification enables the oligonucleotide to form a stable non-covalent complex with low density lipoproteins (LDL) through hydrophobic interactions. The antisense effect of LDL-oligonucleotide complexes was assayed by targeting the mini-exon sequence of Leishmania amazonensis in infected mouse peritoneal macrophages. A 16-mer antisense PSPal oligonucleo- t ide/LDL complex exerted a more pronounced sequence-specific effect than the free oligomer: about 25% and 10% of infected macrophages were cured by a 48 h incubation in the presence of 2.5 /xM of the complexed and the free oligomer, respectively. When oxidized LDL was used instead of the native one for complexation, a further 2-fold increase in the antisense effect was observed suggesting that alternative (unregulated) scavenger receptor can be used for more efficient delivery of antisense oligonucleotides into macrophages. In addition, a significant reduction of the parasitic load was observed in those cells that were not fully cured.

Keywords: Oligonucleotide conjugate; Protozoan parasite; Macrophage; Drug delivery; (Leishmania amazonensis)

1. Introduction

Antisense oligonucleotides provide a versatile method to regulate gene expression post-transcriptionally. While such molecules have the potential to selectivity target a gene in the organism of choice, the susceptibility of unmodified compounds to cellular nucleases and their poor uptake by cells need to be overcome to fully exploit the antisense approach of gene targeting (see [1] for a review). Keeping such objectives in mind, a variety of chemical modifications have been reported which confer new properties to oligonucleotides while retaining the ability to recognize the target sequence [2].

There have been various reports on using vectors to deliver oligonucleotides into the cells [3]. Linking

Abbreviations: LDL, low density lipoproteins; PS, phosphorothioates; PO, phosphodiester; Tin, melting temperature.

* Corresponding author. Fax: +33 57 571015. i Present address: Health Center, St. Louis University, St. Louis, MO

63104, USA.

0167-4781/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 1 6 7 - 4 7 8 1 ( 9 5 ) 0 0 1 4 5 - X

oligomers to poly(L-lysine) [4], to cholesterol [5,6] or to biotin-avidin complexes [7,8] has been shown to promote the penetration of cell membranes. The use of vehicles like liposomes was also reported to increase uptake of oligonucleotide analogues [9,10].

Receptor internalization-mediated delivery has attracted particular attention: the oligomer, suitably modified with a ligand, is internalized by the cell and delivered to specific compartments, viz., lysosomes upon binding to the cognate receptor. For instance, the uptake of neoglycoprotein- oligonucleotide conjugates was shown to be mediated by selective binding to and subsequent internalisation of membrane lectins [11,12]. Interleukin-1/3 conjugate [13] and immunoliposomes [14,15] were also used for receptor-mediated delivery of oligonucleotides to a specific cell type. Delivery of lipophilic drugs using lipoproteins has been studied previously [16] and preliminary attempts have been made to adopt this mode of delivery to antisense oligonucleotides linked to a cholesterol group [17,18].

In trypanosomatids a trans-splicing event leads to the presence of a 39 nucleotide long mini-exon sequence, at

230 R.K. Mishra et al. / Biochimica et Biophysica Acta 1264 (1995) 229-237

the 5' end of every message [19]. Therefore, oligomers complementary to this sequence can potentially block the total protein synthesis of the parasite [20-23]. Indeed, we have reported that antisense phosphorothioate oligonucleotides targeted against the mini exon sequence of Leishma- nia amazonensis can selectively kill the parasite in the host cell [24]. As these protozoan parasites reside in the para- sitophorous vacuole of the infected macrophage, they can receive the content of lysosomes by direct fusion to this vacuole. In this context, delivery of oligonucleotide via low density lipoprotein (LDL) is of special interest.

We describe here the use of phosphorothioate oligonucleotides, further modified by linking a hydrophobic palmityl chain to the 5' end, against L. amazonensis. These oligomers form stable complex with LDL. Such an oligonucleotide-LDL complex, targeted to the mini exon sequence of L. amazonensis, has been assayed for leishmanicidal efficiency in a culture of infected mouse macrophages. The antisense oligonucleotide alone, its complex with native or with oxidized LDL showed increasing sequence-specific effect, in that order.

2. Materials and methods

2.1. Chemicals and enzymes

Palmitic anhydride was from Aldrich. Ethylene diamine, 1,1'-carbonyldiimidazole, pyridine and other or- ganic solvents were from Fluka. Lipofilm kit was from Sebia. [a-32p]ddATP (110 TBq/mmol ) was from Amer- sham. Nucleotidyl terminal transferase was purchased from Boehringer-Mannheim.

2.2. Synthesis o f oligonucleotides

Oligonucleotides listed in Table 1 were synthesised on solid phase using a Milligen 7500 automatic DNA synthe- siser. The oxidation step was performed with Beaucage's reagent for the phosphorothioate modification [25]. Oligo-

Table 1 Oligonucleotides used in this study

Oligonucleotide Sequence (5' ~ 3') Modifications

1 6 a s CTGATACTTATATAGC PO, PS, POPal and PSPal 1 6 m m CTGATACAAATATAGC PS and PSPal 1 6 i n v CGATATATTCATAGTC PO, PS, POPal and PSPal 1 6 t b CTGTTCTAATAATAGC PS and PSPal

Abbreviations: as = anti sense sequence; mm=mismatch (the mis- matched bases are underlined); inv = im'erted and tb = an unrelated sequence, complementary to the mini exon of T~.panosoma brucei. Modifications: PO = unmodified phosphodiester; PS = phophorothioate; PSPal or POPal = phophorothioate or unmodified phosphodiester with a palmityl group at the 5' side. Numbers in front indicate the length of the oligonucleotide.

nucleotides were purified by reverse phase HPLC using a 0-50% gradient of acetonitrile, as previously described [24].

2.3. Synthesis and analysis o f 5'-palmityl derivatiues

The scheme for synthesis is described in Fig. 1. The oligonucleotide to be modified at the 5' end, was detrity- lated while on the solid support after the last cycle of synthesis. The cartridge containing the oligonucleotide linked to the solid support was removed from the synthe- siser and dried extensively in vacuum. The rest of the synthesis was carried out manually on the column, with the help of two 1 ml syringes, following in part a method published previously [26]. Briefly, a 1,1'-carbonyldiimidazole solution (60 mg in 1 ml dry dioxan) was added to the column and reaction was carried out at room temperature for 1 h with occasional mixing with the help of the syringe plungers. The residual carbonyldiimidazole solution was withdrawn and the column was washed with dioxan (I ml, 5 times), dried again in vacuum and then reacted with a solution of ethylene diamine (25/zl in 1 ml dioxane/water; 9:1). The reaction was carried out at room temperature with occasional mixing for 2 h after which the column was washed with dioxan, methanol, acetonitrile, in that order (1 ml, 2 times each) and dried in vacuum overnight. A saturated solution of palmitic anhydride in pyridine (100 mg in 1 ml) was added to the column and the reaction was carried out at 40°C (to avoid precipitation of the anhydride). After 3 h of incubation with occasional mixing, the solution was removed and the column was washed first with pyridine (1 ml X 5) then with acetonitrile (I ml X 2) and dried in vacuum.

The deprotection and removal of the oligonucleotide from the support was done by treatment with 5 ml of aqueous ammonia at room temperature for about 40 h. The ammonia solution was dried and the product was analysed by HPLC using a C 18 reversed phase column. Elution was done in 50 mM triethylammonium acetate buffer (pH 7.0), using a 0-80% linear gradient of acetonitrile in 60 min. The peaks were collected and analysed further.

The oligonucleotide was 3' end-labelled with T4 nucleotidyl terminal transferase and [ c~-32 P]dideoxy ATP (as per vendor's prescription) and purified on a spun column of Biogel P2. Residual traces of unincorporated label after this procedure were detected following analysis of the labelled oligonucleotide on a 20% polyacrylamide gel containing 7 M urea. This background was taken into account for quantitative evaluation of oligonucleotide LDL-complexes (see below).

2.4. Low density lipoprotein isolation and modification

LDL (density 1.019-1.063 g /ml ) was isolated from male human donors, as previously described [27] and the

R.K. Mishra et aL / Biochimica et Biophysica Acta 1264 (1995) 229-237 231

HO : : o ~ ~ o ~

~ . s J | 5

{ CDI )

f B 0 '~ B

~ , o J / It ICon,==,

H~~N~

o

R~C~ ( PALMITIC c / \

- - 0 ANHYDRIDE/

Ir O-

I {l " o Lo-V-oL o ! [~ ,. J , ' ~

I DEPROTECTION ( NH 3 )

H 0 (" B ~ B

-; -~--; -.o_~o_i_+q-o.

Fig. 1. Scheme for the synthesis of the 5' palmityl derivative of oligonucleotides.


protein concentration was determined by the method of Lowry. When needed, LDL (100 mg of protein/ml) was oxidized by incubation with 5 /~M CuSO 4 in Ham's F10 medium (Gibco BRL) at 37°C for 24 h. Oxidized LDL was dialysed against phosphate-buffered saline (PBS) and checked by electrophoresis in comparison to native LDL. The degree of oxidation, immediately determined by analysis of thiobarbituric acid-reactive substances according to [28], varied between 7.7 and 11 nmol malondialdehyde equivalent per mg of protein. LDL was routinely examined for endotoxin content and contained < 5 pg of bacterial lipopolysaccharide per mg of LDL protein.

2.5. Interaction of 5' palmityl-modified oligonucleotides with LDL

2.7. Sensitivity of modified oligonucleotide / LDL complex to nucleases

3' End-labelled 16PSPal or 16POPal oligonucleotide (0.1 /xM in 10 /xl reaction mixtures), free or complexed with native LDL (40 /xg/ml, final concentration), was incubated with fetal calf serum (inactivated by heating at 560C for 1 h) at 37°C. Denaturing dye mix was added and samples were placed on ice to stop the reaction. The degradation was estimated by analysing the mixtures on a 20% polyacrylamide-7 M urea gel and autoradiography. The autoradiograms were quantitated using video image analysis with the 'Image 72' Software (Appligbne, France).

2.8. Leishmanicidal effect of oligonucleotides

The palmityl-modified oligonucleotide 16asPSPal, 32 p_ labelled at 3' end, was used for monitoring its interaction with human LDL. Oligonucleotide (0.1 /~M) and LDL (5 mg/ml , protein concentration), in PBS containing 2 mM EDTA, were incubated at 37°C for 2 h, although the complexation reaction was complete in less than one h (not shown). The mixture was then analysed by gel electrophoresis using the precast polyacrylamide gels of Lipofilm kit from Sebia (France) in the supplied buffer at a constant current of 6 mA (0.7 mA/cm). The samples were premixed with Coomassie blue to monitor the migration of protein band. The gel was dried and autoradiographed.

32p End-labelled palmityl-modified oligonucleotide- LDL mixtures were also analysed by gel filtration on Sephadex G50 (fine) packed in 1 ml tips and eluted with PBS. I00 /xl fractions were manually collected and each fraction was monitored for oligonucleotide (by counting the fractions for associated radioactivity) and for LDL (by estimating the protein content by Bradford reagent). The residual label present in the oligonucleotide solution was subtracted for each value. It corresponded to the fraction which did not bind even to a large excess (50-fold) of LDL.

To estimate the stability of the oligonucleotide-LDL association, the purified complex was re-loaded on a Sephadex G-50 column and protein and radioactivity were monitored as above.

2.6. Melting curues

Peritoneal macrophages were collected from Balb/c mice and infected with the parasite to obtain 55-75% of infected cells, adhering to the coverslip, as previously described [24]. Oligonucleotides at the desired concentration, either free or complexed with LDL (40 /xg/ml or 15 /.~g/ml of native and oxidized LDL, respectively) were added to this infected cell population, in complete RPMI- Hepes medium (RPMI medium/25 mM Hepes, pH 7.3, containing 5% fetal calf serum and 50 /xg of gentamycin per ml). The incubation was carried on for 48 h at 34°C after which the medium was removed and ceils were incubated in fresh medium, without oligomers, for another 18 h at 34°C. Cells were then fixed with methanol, stained with Giemsa and counted for the level of infection and parasitic load. The level of infection in treated cells was compared to that in untreated cells taken as 100% infected.

2.9. Uptake of oligos by macrophages

32p 3' end-radiolabelled oligonucleotide (0.5 /xM) free or complexed with native LDL (40 /xg/ml, final concentration) was added to a final volume of 100 /xl containing about 10 6 cells in complete medium as above, except that it contained 5% of heat inactivated fetal calf serum. The cell suspension was incubated at 37°C or 4°C for varied time and then passed through an oil cushion by spinning 10000 rpm; 5 min) in a microfuge. The bottom of the tube containing the cell pellet was cut and counted for the associated radioactivity.

Modified or unmodified oligonucleotide (2 /zM) was mixed with an equimolar amount of complementary sequence in 10 mM Tris-HC1 (pH 7.5) containing 100 mM NaC1, heated at 80°C for a few minutes and allowed to reanneal slowly. The stabilities of the duplexes were determined by recording melting curves on a Uvikon 940 spectrophotometer. Absorbance at 260 nm was monitored with increasing temperature at a rate of 0.5°C per rain by a Hiiber Ministat, driven by a Hiiber PD 410 controller.

3. Results

A phosphorothioate oligonucleotide was conjugated at its 5' end with a palmityl group and purified by reversed phase HPLC, as described in Section 2. The non-conjugated oligonucleotide 16asPS eluted at 22.62 min ('Peak A', Fig. 2). The delayed 'Peak B' (at 41.15 min; yield 55%) displayed an absorption spectrum characteristic of nucleic acids, estimated to correspond to the palmityl-con-

R.K. Mishra et al. / Biochimica et Biophysica Acta 1264 (1995) 229-237 233

250-

200

i50.

mV

i00-

A B

i, ii l i !! ,, !! i

0 . . . . . . . . . . . . "

, , , , , ,

O0 I 0 3 0 45.0 600

Time (minute)

Fig. 2. Purification of 5' palmityl oligonucleotide conjugate by reversed phase HPLC. The unmodified oligonucleotide 16asPS elutes at 22.62 rain (peak A) and the more lipophilic oligonucleotide (due to 5' palmityl-modification) 16asPSPal elutes at 41.15 min (peak B) of the gradient (see text for details). The inset shows an autoradiogram of the electrophoretic analysis of 3' end-radiolabelled oligonucleotides from the two peaks (labelled as such) on a 20% polyacrylamide gel containing 7 M urea.

jugate 16asPSPal, since the highly hydrophobic covalent modification results in increased retention (Fig. 2). In contrast to the unmodified one, this oligonucleotide could not be precipitated by ethanol. When ~4C-labelled eth- ylenediamine was used (Fig. 1, step 2), the 'Peak B' contained the radiolabelled substrate. As expected, this oligonucleotide fraction was not a substrate for polynu- cleotide kinase. Finally, the compound of peak B moved slower in urea-polyacrylamide gel than the unmodified oligonucleotide of identical size and sequence (peak A) (Fig. 2, inset). These results strongly suggest that the late eluting fraction (peak B) indeed contained the modified oligonucleotide 16asPSPal.

We performed such a modification with both phosphorothioate and phosphodiester oligomers of different sequences to obtain PSPal and POPal oligomers, respectively (see Table 1). To determine the affinity of modified oligonucleotides towards the complementary RNA or DNA sequences, we measured the T m and compared it with that of non-conjugated ones. The results show that 5' palmityl modification did not change the affinity of the oligonucleotide towards the target sequence. However, as previously reported [29-31], complexes with PS oligomers exhibited a lower stability than those with phosphodiester ones (Table 2). A similar behaviour was observed for both DNA and RNA complementary sequences.

3.1. Interaction of 5' palmityl-oligonucleotide conjugates with LDL

Unlike the non-conjugated oligonucleotides, those modified at the 5' end with palmityl group show the ability to complex with LDL. This was determined using electrophoresis as well as gel filtration chromatography. The mixture of LDL with 32p end-labelled 16asPSPal oligonucleotide resulted in the appearance of a slow-migrating species detected after autoradiography which corresponded to LDL as shown by protein staining (Fig. 3). No retarded band was seen with non-conjugated oligomers 16asPS or 16asPO. 32p end-labelled oligonucleotide and LDL mixtures were also analysed on a G50 Sephadex column. The palmitate-oligonucleotide conjugate 16asPSPal was eluted with the lipoprotein peak at the void volume whereas the non-conjugated oligomer 16asPS was eluted much later (Fig. 4A), indicating that no interaction took place between this non-conjugated parent compound and LDL particles. Varying the oligonucleotide to LDL ratio and subsequent analysis by gel filtration chromatography indicated that 10-15 molecules of oligonucleotide can be loaded onto one LDL particle (Fig. 4B). The stability of the purified complex was tested by analysing it similarly after incubation at 37°C for 2-3 h. The complex did not show any detectable dissociation, suggesting that once formed, it was stable for, at least, a few hours.

We monitored the nuclease sensitivity of oligonucleotide-LDL complexes. We compared the behavior of 3' end-labelled PO or PS oligomers upon incubation in cell growth medium. We observed about a 5-fold reduction in the amount of degraded palmityl-PO oligonucleotide in the LDL complex compared to the free conjugate, following incubation in complete RPMI/Hepes medium for 30 rain at 37°C (not shown). LDL had no effect on the lifetime of the non-conjugated oligonucleotide. A lower effect was seen with PS oligonucleotides as these derivatives display a better resistance to nucleases than PO ones.

3.2. Uptake of oligonucleotides by macrophages

In the uptake experiments, 5' palmityl-phosphorothioate oligonucleotide 16asPSPal complexed with LDL, under

Table 2 Melting temperatures (T m) of oligonucleotide-target hybrids

Oligonucleotide Target Tm(°C)

16asPS DNA 34.9 16asPS RNA 41.2 16asPO DNA 45.2 16asPO RNA 51.7 16asPSPal DNA 34.2 16asPSPal RNA 39.6

Targets were synthetic 18-mers 5'-CGC UAU AUA AGU AUC AGU or 5'-CGC TAT ATA AGT ATC AGT. Melting curves were obtained as described in Section 2.


1 2 3 4 5 1 2 3 4 5

(A) (B) Fig. 3. Electrophoretic analysis of LDL-oligonucleotide complex. Oligo- nucleotides were 32 P-labelled at the 3' end. Panel A is protein profile as seen by Coomassie blue. The band in lane 1 corresponds to the free dye which is not seen in lanes 2-5 due to the use of a limiting amount of dye in the loading buffer; Panel B is the same gel dried and autoradiographed. Lane 1 is 16asPSPal; lanes 2 -4 are LDL with 16asPO, 16asPS and 16asPSPal, respectively; lane 5 is LDL alone.

6oo A II o . 9 -

o

0 0 5 10 15 2 0

FRACTION NUMBER

~ 1~'

F, ~ 80'

0 ~ 40"

i • | , | - -

0 1 0 2 0 3 0 4'0 5'0

OLIGONUCLEOTIDE / LDL (molar ratio)

Fig. 4. Analysis of LDL-oligonucleotide complex by Sephadex G50 gel filtration column chromatography. (A) LDL was mixed with either 3' end-labelled 16asPS ((3) or 16asPSPal ( 0 ) and analysed on the column; each fraction was counted for associated radioactivity, Protein content (11) was estimated for LDL alone (see Section 2.5). (B) LDL 16asPSPal complexes were made by mixing the two in varying molar ratios with 0.1 /xM oligonucleotide. The ratio of free and complexed oligonucleotide was estimated by comparing the early (for bound) and late (for free) eluting peak, see Fig. 4A.

5 0 ,

40' ~ .T.

(6 -J 30'

o a 20'

o 10'

0 0 2 4 6 8 10

OLIGONUCLEOTIDE, ixM

Fig. 5. Leishmanicidal effect of oligonucleotides, Cells infected with the parasite as described in Section 2 were incubated in the presence of 16asPS (O); 16asPSPal ( 0 ) ; 16asPSPal complexed with LDL (Ill); 16asPSPal complexed with oxidized LDL ( • ) ; 16 mmPSPal complexed with LDL ( [] ); and 16 mmPSPal complexed with oxidized LDL ( zx ).

conditions indicated in Section 2.9, showed a 2-fold increased association with macrophages at 37°C in comparison to the free oligonucleotide 16asPS. This increased uptake was not seen at 4°C and was also more prominent at early time points, e.g., 30 min (not shown). No significant degradation of the oligonucleotide was detected under these conditions. This suggested that the LDL-oligonucleotide complex is taken up by the cells by an active mecha- nism, presumably the LDL receptor-mediated endocytosis.

3.3. Leishmanicidal effect of phosphorothioate oligonucleotides

We then investigated the in vitro effect of oligomers on L. amazonensis. The antisense oligonucleotides targeted to the mini-exon sequence were effective to varying degrees, in curing mouse peritoneal macrophages infected by L. amazonensis amastigotes depending upon the modification, as shown in Fig. 5. At lower concentrations, 16asPSal showed an increased biological effect compared to 16asPS. However, at concentration of l0 ~M and above, both oligonucleotides exhibited similar efficiency in the absence of LDL. An improvement was seen with LDL/oligonucleotide complexes: at 2.5 ~M oligonucleotide associated with 40 p.g/ml native LDL, the number of cured cells was twice as large with the LDL-delivered antisense oligonucleotide compared to the free 16asPS. This effect was further increased when oxidized LDL was used for the complexation: 45% of cured macrophages were obtained in the presence of 2.5 /~M antisense oligonucleotide associated with 15 /zg/ml oxidized LDL. This should be compared to 10% for the parent compound 16asPS. In addition, the upper limit (45%) obtained with the oligomer delivered with oxidized LDL was not reached with the non-conjugated 16asPS for which the effect plateaued at about 25% in the presence of 10 ~M oligonucleotide. We have shown previously that 45% cured cells were reached only in the presence of 25 /xM 16asPS [24]. Cells treated


t/.I i- tO

el

,,r- l--

tO =,A ..d kl.I ¢,)

,-e

80'

60

40'

2 0

....i r'~ ..A

+

o.. 03 o_

+ to

_ 13 . .

L.

O

Fig. 6. Effect o f ant isense ol igonucleot ides on the parasi t ic load o f

macrophages . Cells were treated with 2.5 tzM ol igonucleot ide , in the

presence o f LDL, oxidized L D L or nothing, as indicated. The percentage

of cells with 0 - 5 parasi tes (of the total infected cells as seen in control ,

i.e., wi thout t reatment) were scored.

with non complementary 16mmPS, 16tbPS, 16invPS or with 16invPSPal complexed with LDL or oxidized LDL, did not show any decrease in the parasitemia indicating that the effect of 16asPS sequences is due to the sequence-dependent binding of the antisense oligomers to the mini-exon target. Unloaded LDL or oxidized LDL were not toxic to the macrophages at the concentrations used; toxic effects on cells were seen above 100 /zg/ml native LDL or above 25 /zg /ml oxidized LDL.

Over and above the effect scored in terms of fraction of infected cells getting cured completely, we also observed a significant decrease in the parasitic load (Fig. 6). Incuba- tion of infected cells with 2.5 /zM of 16asPS led to 25 __+ 4% of cells with fewer than 5 parasites. This should be compared to only 6 + 3% in the non-treated control population. At this concentration, the palmityl conjugate 16asPSPal induced a low parasitemia ( < 5 parasites/cell) in 40 + 4% of the cells. This decrease in parasitemia was seen in as much as 52 + 4% and 73 + 5% of cells when this oligonucleotide conjugate was carried by native and oxidized LDL, respectively. This, and also the observation that cultures with lesser degree of infection (lower parasitemia) respond to the antisense treatment better than the cells with higher parasitemia, suggest that the antisense oligonucleotides indeed are active in almost all the cells, however only a fraction gets cured under the experimental conditions used here.

4. D i s c u s s i o n

We have reported previously that phosphorothioate oligonucleotides, targeted to the mini-exon sequence of the protozoan parasite L. amazonensis, can block protein syn-

thesis in vitro [23] and control the growth of the parasite in cultured cells [24]. Despite these positive results, anti- leishmanial chemotherapy by antisense oligomers is ham- pered by the location of protozoan inside the host macrophage. Indeed large amounts (25 /zM) of a 16-mer antisense oligonucleotide (16asPS) had to be used to cure less than 50% of infected murine cells. We describe here the properties of oligonucleotide-LDL complexes and their ability to control Leishmania development in cultured cells. We compared also the efficiency of both native and oxidized LDL, in order to take advantage of two classes of receptors on the macrophage membrane [16].

Oligonucleotides were first modified at the 5' end with a palmitic chain. The end-modified oligonucleotide 16asPSPal was able to form complexes with LDL up to 10-15 molecules per LDL particle (Fig. 4B). Conflicting results were previously reported according to which satura- tion occurred at an oligonucleotide/LDL molar ratio of either 50:1 [17] or 7.5:1 [18]. These differences may reflect either a variability in LDL samples or methods used for quantification. The association of the oligonucleotide to LDL depends on the presence of the hydrophobic ligand: the non-conjugated oligomer 16asPS did not significantly bind to LDL. The non-specific association of phosphorothioate oligonucleotides to proteins [32] does not seem to be significant in our case.

The use of the oligonucleotide-palmitate conjugate re- vealed a slight but reproducible improvement of the leishmanicidal effect of the phosphorothioate anti-mini-exon sequence 16asPSPal compared to the non-conjugated oligomer 16asPS, at concentration below 10 /zM. This suggests that the presence of the hydrophobic ligand pro- motes the interaction with the cells or the internalisation of oligonucleotide associated with serum components. The improvement displayed by LDL/oligonucleotide complexes is likely related to an improved uptake of the oligomer, as suggested by experiments performed at low temperature and by results obtained with cholesterol-conjugates [17,18].

Oxidized LDL is ligand to scavenger receptors, whose endocytosis is not regulated by the internal cholesterol concentration unlike that of LDL receptor. We speculated that oxidized LDL may deliver oligonucleotides more effi- ciently than native one and, if so, the same oligonucleotide should have a more pronounced effect. In accordance with these expectations, the oligomer 16asPSPal/oxidized LDL complex did show a significant improvement in its leishmanicidal effect at lower concentrations: the use of oxidized LDL allowed to reduce, by a factor of about ten, the amount of oligonucleotide required to get the highest effect of this antisense sequence. A simultaneous reduction of the parasitic load was induced by 16asPSPal/oxidized LDL complex. At 2.5 /zM oligonucleotide fewer treated cells (about 25% of the infected cells) contained more than 5 parasites whereas this figure amounted to about 95% for non-treated macrophages. These effects were selective: no


significant ( < 5%) reduction was observed with unloaded native or oxidized LDL. There was no effect either with LDL loaded with phosphorothioate 16-mers palmitate conjugates non complementary to the mini-exon sequence. Therefore, the killing of L. amazonensis amastigotes in murine peritoneal macrophages resulted from a sequence- specific effect of the antisense oligonucleotide, i.e., from an interaction between 16asPS and its target on the mini- exon sequence. Whether this occurred on the pre-RNA or matured RNAs, by interfering with trans-splicing or trans- lation, respectively, remains to be determined.

We were never able to cure Leishmania-infected macrophage cultures beyond 50%. This phenomenon was also observed when free 16asPS was used. We previously demonstrated that this was not due to the presence of a parasite sub-population resistant to the antisense oligonucleotide treatment [24]. Clearly, delivery of the antisense oligomer by LDL did not overcome this limitation although decrease in parasitemia was seen in more cells. However, cells that apparently did not respond to the treatment exhibited a very high parasitemia and drastic morphological changes. This heavily infected population could correspond to macrophages which have reduced uptake properties or leishmanicidal potency in the presence of the internalized oligomer. This may also be attributed to the state of cells. We did not observe any significant effect of antisense phosphorothioate oligomers on promastigotes and axenic amastigotes, which are free stages of Leishma- nia. The use of LDL did not improve either the antipara- sitic efficiency.

Various delivery vehicles have been used against Leish- mania amastigotes. More than fifteen years ago Black et al. used liposome-entrapped antimonials in vivo [33]: sev- eral teams obtained significant improvement of classical drugs like Pentostam or Glucantim against L. donovani in mice [34,35]. A lipophilic derivative of amphotericin was also shown to exhibit a higher efficiency in vivo when delivered with LDL [36]. Our work clearly demonstrates that an increased uptake of the LDL-associated oligonucleotide, leading to its accumulation in the cell compart- ment where the parasite is growing improves its biological effect. This is further supported by the fact that different efficiencies were observed with native and oxidized LDL. A reduced sensitivity to serum nucleases of oligonucleotides bound to LDLs could also partly contribute to the increased antisense effect. This approach is of special interest to reduce the quantity of costly and/or sophisti- cated chemically-modified oligonucleotides required for cellular studies.

Acknowledgements

This work was in part supported by the Direction Gtntrale des Etudes et Techniques and by the P61e M~di- cament Aquitaine.

References

[1] Htl~ne, C. and Toulmt, J.J. (1990) Biochim. Biophys. Acta 1049, 99-125.

[2] Goodchild, J. (1990) Bioconj. Chem. 1, 165-187. [3] IAonetti, J.P. and Leserman, L. (1993) in Antisense Research and

Applications (Lebleu, B. and Crooke, S.T., ed.), pp. 493-504, CRC Press, Boca Raton.

[4] Lema~tre, M., Bayard, B. and Lebleu, B. (1987) Proc. Natl. Acad. Sci. USA 84, 648-652.

[5] Boutorin, A.S., Guskova, L.V., Ivanova, E.M., Kobetz, N.D., Zary- tova, V.F., Ryte, A.S., Yurchenko, L.V. and Vlassov, V.V. (1989) FEBS Lett. 254, 129-132.

[6] Letsinger, R.L., Zhang, G., Sun, D.K., Ikeuchi, T. and Satin, P.S. (1989) Proc. Natl. Acad. Sci. USA 86, 6553-6556.

[7] Temsamani, J., Agrawal, S. and Pederson, T. (1991) J. Biol. Chem. 266, 468-472.

[8] Pardridge, W.M. and Boado, R.J. (1991) FEBS Lett. 288, 30-32. [9] Ropert, C., Lavignon, M., Dubernet, C., Couvreur, P. and Malvy, C.

(1992) Biochem. Biophys. Res. Commun. 183, 879-885. [10] Thierry, A.R., Rahman, A. and Dritschilo, A. (1993) Biochem.

Biophys. Res. Commun. 190, 952-960. [11] Bonfils, E., Depierreux, C., Midoux, P., Thuong, N.T., Monsigny,

M. and Roche, A.C. (1992) Nucleic Acids Res. 20, 4621-4629. [12] N~gre, E., Chance, M.L., Hanboula, S.Y., Monsigny, M., Roche,

A.C., Mayer, R.M. and Hommel, M. (1992) Antimicrob. Agents Chemother. 36, 2228-2232.

[13] Chollet, A. (1990) Nucleos, Nucleot. 9, 957-966. [14] IAonetti, J.-P., Machy, P., Degols, G., Lebleu, B. and Leserman, L.

(1990) Proc. Natl. Acad. Sci. USA 87, 2448-2451. [15] Zelphati, O., Zon, G. and Leserman, L. (1993) Antisense Res. Dev.

3, 323-338. [16] Shaw, J.M., Shaw, K.V., Yanovich, S., Iwanik, M., Futch, W.S.,

Rosowsky, A. and Schook, L.B. (1988) Ann. N.Y. Acad. Sci. 507, 252-271.

[17] Desmidt, P.C., Le Doan, T., Defalco, S. and Vanberkel, T.J.C. (1991) Nucleic Acids Res. 19, 4695-4700.

[18] Krieg, A.M., Tonkinson, J., Matson, S., Zhao, Q.Y., Saxon, M., Zhang, L.M., Bhanja, U., Yakubov, L. and Stein, C.A. (1993) Proc. Natl. Acad. Sci. USA 90, 1048-1052.

[19] Borst, P. (1986) Annu. Rev. Biochem. 55, 701-732. [20] Cornelissen, A.W.C.A., Verspieren, P., Toulmt, J.J., Swinkels, B.W.

and Borst, P, (1986) Nucleic Acids Res. 14, 5605-5614. [21] Walder, J.A., Eder, P.S., Engman, D.M., Brentano, S.T., Walder,

R.Y., Knutzon, D.S., Dorfman, D.M. and Donelson, J.E. (1986) Science 233, 569-571.

[22] Verspieren, P., Cornelissen, A.W.C.A., Thuong, N.T., Htl~ne, C. and Toulm~, J.J. (1987) Gene 61,307-315.

[23] Pascolo, E., Blonski, C., Shire, D. and Toulmt, J.-L (1993) Biochimie 75, 43-47.

[24] Ramazeilles, C., Mishra, K.R., Moreau, S., Pascolo, E. and Toulm~, J.J. (1994) Proc. Natl. Acad Sci. USA 91, 7859-7863.

[25] Iyer, P.I., Lawrence, R.P., Egan, W., Regan, J.B. and Beaucage, S.L. (1990) J. Org. Chem. 55, 4693-4698.

[26] Wachter, L., Jablonski, LA. and Ramachandran, K.L. (1986) Nu- cleic Acids Res. 14, 7985-7994.

[27] Haval, J., Eder, H.A. and Bragdon, J.H. (1955) J. Clin. Invest, 34, 1345-1351.

[28] Streinbecher, U.P. (1987) J. Biol. Chem. 262, 3603-3611. [29] Morvan, F., Porumb, H., Degols, G., Lefebvre, I., Pompon, A.,

Sproat, B.S., Rayner, B., Malvy, C., Lebleu, B. and Imbach~ J.L. (1993) J. Med. Chem. 36, 280-287.

[30] Kiblerherzog, L., Zon, G., Uznanski, B., Whittier, G. and Wilson, W.D. (1991) Nucleic Acids Res. 19, 2979-2986.

[31] Cazenave, C., Stein, C.A., Loreau, N., Thuong, N.T., Neckers, L.M.,


Subasinghe, C., H61bne, C,, Cohen, J.S. and Toulm~, J.J. (1989) Nucleic Acids Res. 17, 4255-4273.

[32] Stein, C.A. and Cheng, Y.C. (1993) Science 261, 1004-1012. [33] Black, C.D.V., Watson, G.J. and Ward, R.J. (1977) Trans. Roy. Soc.

Trop. Med. Hyg. 71, 550-552. [34] Alving, C.R., Steck, E.A., Chapman, W.L., Waits, V.B., Hendricks,

L.D., Swartz, G.M. and Hanson, W.L. (1978) Proc. Natl. Acad. Sci. USA 75, 2959-2963.

[35] New, R.R.C., Chance, M.L., Thomas, S.C. and Peters, W. (1978) Nature 272, 55-56.

[36] New, R.R.C., Chance, M.L. and Heath, S. (1981) J. Antimicrob. Chemother. 8, 371-381.

improved leishmanicidal effect of phosphorotioate antisense oligonucleotides by ldl-mediated...

Documents