Lavigne C, Yelle J, Sauve G, Thierry A. Is Antisense an Appropriate Nomenclature or Design for Oligodeoxynucleotides Aimed at the Inhibition of HIV-1 Replication?.
AAPS PharmSci. 2002; 4 (2): article 9. DOI: 10.1208/ps040209

Is Antisense an Appropriate Nomenclature or Design for Oligodeoxynucleotides Aimed at the Inhibition of HIV-1 Replication?
Carole Lavigne,^1,2  Jocelyn Yelle,²  Gilles Sauvé,²  and Alain R. Thierry^3,4 

¹Département de Microbiologie et Immunologie, Faculté de Médecine, Université de Montréal, Montréal, Québec, Canada H3C 3J7
²Institut Armand-Frappier, Université du Québec, Laval, Québec, Canada H7N 4Z3
³Laboratory of Tumor Cell Biology, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892-4255
⁴MedinCell Project, Laboratoire des Défenses Antivirales et Antitumorales, UMR 5124, 1919 route de Mende, 34293 Montpellier Cedex 5, France

Correspondence to:
Alain R. Thierry
Tel:
Fax:
Email: thierry1@micronet.fr

Submitted: December 15, 2001; Accepted: February 27, 2002; Published: April 30, 2002

Keywords:  Antisense, Oligodeoxynucleotides, Oligonucleotides, HIV, Specificity, Drug Delivery, Lipoplexes

Abstract

We have evaluated the specificity and the variation in activity against human immunodeficiency virus (HIV) infection of antisense oligodeoxynucleotides (ODNs) with regard to factors such as dose-response range, number and choice of experimental controls, backbone modifications of the ODNs, type of cell infection, length of assays, and delivery approach. The highest level of inhibition was achieved in our long-term assay with MOLT-3 cells acutely infected with HIV-1 (IIIB) and treated with free phosphorothioate-modified ODNs (PS-ODNs). The highest level of specificity was observed in our short-term assay with MOLT-3 cells acutely infected with HIV-1 (IIIB) and treated with free PS-ODNs. The highest potency (IC50 level) was observed in our short-term chronic-infection model with (DLS)-delivered ODNs in which the DLS delivery improved the ODN activity up to 106 times compared to the activity of free ODNs. Thus, the near blocking of HIV replication obtained when using PS-ODNs appears because of the addition of extracellular and/or membranar effects. The higher efficacy of PS-ODNs compared to unmodified ODNs, when both are delivered with the DLS system, was demonstrated solely in our short-term assay with MOLT-3 cells. Important variations in the level of sequence specificity were observed and depended on the type of control used and the type of cell assay employed. It seems that all 3 groups of control-tested, random, sense sequence, and non-antisense T30177 ODNs might have distinct activity and, consequently, different modes of action in inhibiting HIV replication. Our data buttress the notion that the contribution of the sequence-specific mediated mode of action is minor compared to the other mechanisms involved in ODN antiviral activity.

Introduction

Antisense oligodeoxynucleotides (ODNs) are a novel class of therapeutic agents that offer an attractive approach for the treatment of viral infections, cancer, and genetic disorders by controlling cellular or viral gene expression at the mRNA and possibly gene levels. These molecules are designed to block the action of specific genes by binding to their RNA transcripts by a phenomenon called hybridization arrest through Watson-Crick base-pairing. Typically, antisense ODNs can selectively interfere with RNA splicing or processing, prevent initiation of translation, block progression of ribosomes along the mRNA, and cause RNA cleavage through activation of RNAse H.¹ Paul Zamecnik's laboratory published the first report on application of antisense ODNs in 1978,² along with a report on their application against HIV replication in 1986.³ The authors showed that a 13-mer synthetic ODN complementary to the 3'-end of the Rous sarcoma virus genome inhibited the formation of new virus and prevented transformation of chick fibroblasts into sarcoma cells when added exogenously to the infected cell cultures. Since then, many oligonucleotide analogs have been synthesized and studied as antisense agents, and antisense effects of ODNs in mammalian cells have been reported in numerous tissue culture experiments⁴ and in several recent in vivo studies.^5,6

The most widely used analogs of ODNs are phosphorothioate-modified ODNs (PS-ODNs), which have a sulfur-for-oxygen substitution for 1 of the nonbridged oxygen atoms of internucleotide linkages. PS-ODNs have good biological activity, pharmacology, pharmacokinetics, and safety.^7,8 PS-antisense ODNs are now evaluated in clinical trials for their therapeutic potential against several human diseases, including cancer, viral infections, and connective tissue disorders, and initial results are encouraging.⁹ Although results from recent clinical trials of first-generation PS-ODN drugs seem promising, the mechanism of action of these compounds remains somewhat speculative and may include antisense and non-antisense effects.¹⁰ PS-ODNs have been shown to inhibit the replication of human immunodeficiency virus (HIV-1) in vitro by both sequence-specific and non-sequence-specific activity, depending on the cell culture models used, the concentrations, and the assay's design.^10,11 Several studies have demonstrated that the potency and specificity of antisense ODNs can be greatly enhanced by the use of lipid-based carrier system to deliver the ODNs into cells.^12,13 Lipid-based carriers, such as the DLS system, increase cellular uptake and help antisense compounds accumulate in the nucleus, allowing activity at much lower concentrations.^14,15

In this work, we studied the antiviral activity of various ODN sequences in different HIV infection cell culture models. We compared the activity of free and DLS-associated PS-ODNs with that of DLS-delivered unmodified ODNs (PO-ODNs) and PS-ODNs, in both acute- and chronic-infection models. We looked at several factors that have been proposed to account for discrepancies in the literature on antisense technology such as dose-response range, number and choice of experimental controls, backbone modifications of the ODNs, and type of cell infection (acute or chronic), along with length of assays and delivery approach, in order to improve further protocol design in this field area. Although this paper describes observations already made independently elsewhere, it constitutes a unique comprehensive work, as no previous report has addressed the effect of all these parameters influencing ODN's anti-HIV activity. The overall goal and objective of this report are to present a thorough analysis of the question of sequence specificity and, consequently, to evaluate antisense ODN design.

Materials and Methods

PO-ODNs and PS-ODNs (ODN with a sulfur atom introduced at each phosphodiester linkage) were synthesized by using an automated DNA synthesizer (BioServe Biotechnologies, Laurel, MD) and purified by polyacrylamide gel electrophoresis. In this study, we used 4 antisense sequences known to have anti-HIV-1 activity in either phosphorothioate (PS-ODN) or unmodified (PO-ODN) form: antisenses DIS and Pac, which are complementary to a portion of the 5'-long terminal repeat of the HIV-1 genome and are considered to be essential for HIV-1 RNA encapsidation^16,17; antisense GEM 91 (gene expression modulator 91), a 25-mer complement to the gag initiation site of HIV-1¹⁸; and antisense rev, a 28-mer complementary to the 5'-end sequence of HIV-1 rev mRNA.¹⁹ The ODN sequences are shown in Table 1. As a control, a 28-mer random phosphorothioate (RS) or phosphodiester (RD) sequence was made from a mixture of all 4 nucleotides at each synthesis step, in order to verify the sequence specificity of the antisense ODNs. As controls for PS-ODNs, we also used 2 sense sequences named SSDIS and SSPac for sense SDIS and sense SPac, respectively (Table 1). In addition, an ODN sequence named T30177 was used here as a non-antisense control. T30177 is a 17-mer ODN that comprises only deoxyguanosine and thymidine residues stabilized by an intramolecular guanosine octet.²⁰ T30177 contains single phosphorothioate internucleoside linkages at its 5' and 3' ends, and has shown strong anti-HIV activity in infected cells on the basis of its 3-dimensional structure.

The human lymphoid cell line MOLT-3 was kindly provided by Dr R.P. Sekaly (Clinical Research Institute of Montreal, Québec, Canada). H9 cells chronically infected with HIV-1 (IIIB) (H9/human T-lymphotropic virus [HTLV]-IIIB [NIH] 1983)^21,22 were obtained from the NIH Acquired Immunodeficiency Syndrome (AIDS) Research and Reference Reagent Program (Rockville, MD). Uninfected and infected cells were cultivated in RPMI 1640 culture medium (Invitrogen / Life Technologies, Burlington, Ontario, Canada) supplemented with 10% heat-inactivated fetal calf serum, L-glutamine (4 mM), and gentamycin (50 µg/mL) at 37°C in a 5% CO2 atmosphere. Peripheral blood mononuclear cells (PBMCs) from healthy HIV-1-seronegative donors were isolated by Ficoll-Hypaque (Amersham Biosciences, Piscataway, NJ) gradient centrifugation of heparinized venous blood. The cells were collected, washed, and stimulated with phytohemagglutinin-P (PHA-P) (1 µg/mL; Pharmacia Biotech) for 24 hours. The cells were then washed and maintained in the same culture medium as above, supplemented with recombinant human interleukin-2 (10-20 U/mL; ZeptoMetrix Corporation, Buffalo, NY). HIV-1 laboratory strain IIIB was obtained from Advanced BioScience Laboratories Inc (Kensington, MD) and was used to infect MOLT-3 cells and primary cells.

DLS liposomes are a mixture of equal amounts of dioctadecylamidoglycylspermidine (Promega, Madison, WI) and dioleoyl phosphatidylethanolamine (Sigma-Aldrich Corp. St-Louis, MO) and consist of small unilamellar vesicles, which can complex with ODNs in an interactive molecular manner. Liposomes were prepared as previously described.^23,24 Oligonucleotides were first complexed to DLS preparation in sterile deionized water as described earlier,¹⁵ and the preparation was incubated at room temperature for at least 30 minutes just prior to addition to the cells. Dilution in deionized water was made to obtain appropriate concentrations. The lipoplexes used here are specifically formulated to obtain highly stable and reproducible preparations. These preparations showed high homogeneity in size (polydispersity ~ 0.20) as determined by dynamic light scattering. The coefficient of variation of the mean size (~120 nm) was found to be 19.5%. ODN-liposome complexes were stored at 4°C up to the next treatment 3 or 4 days later. Fresh ODN-liposome complexes were prepared every week (every 2 treatments).

Antiviral activity of ODNs was evaluated in 2 different cell systems. MOLT-3 cells or PBMCs were infected with HIV-1 (IIIB) at a multiplicity of infection of 0.1. Cells were cultured with various concentrations of ODNs either added free in culture medium or complexed with DLS liposomes for 7 days. The supernatants were then harvested and examined for virus production.

To test the antiviral activity of the antisense ODNs in a long-term model treatment in acutely infected cells, MOLT-3 cells were infected at a multiplicity of infection of 0.01 with HIV-1 laboratory strain IIIB as previously described.¹⁵ Cells were treated for up to 21 days with PS-ODNs added free to the cultures at 0.01, 0.1, or 1 µM concentrations, or for up to 28 days with either PS-ODNs or PO-ODNs complexed with DLS liposomes at 0.001, 0.01, 0.1, and 5 nM concentrations. Every 3 or 4 days, cells were split to 4 x 105 cells/mL and supernatants were collected to determine the HIV-1 titer. Except for the initial 2-hour period of virus adsorption, the oligonucleotides were always present in the culture medium during the course of the experiments.

To study the inhibition of HIV-1 by antisense ODNs and their efficiency in chronically infected cells, we used the H9 cell line chronically infected with HIV-1 (IIIB). Cells were plaed into 96-well microtiter plates at a concentration of 4 x 105 cells/mL, and antisense ODNs were either added free or complexed with DLS liposomes at 1.5 and 18.5 µM concentrations or at 0.001 and 0.01 nM concentrations, respectively. The cells were kept in culture for 3 to 4 days, and HIV-1 replication was determined by the p24 antigen assay.

Virus replication was determined by detection of p24 HIV-1 viral core antigen in cell-free supernatants by a p24 antigen-capture assay (Beckman Coulter Immunology, Fullerton, CA or SAIC Frederick, Frederick, MD). Cell viability and cytotoxicity were monitored by the tetrazolium-based colorimetric cell proliferation assay (MTT).²⁵

Data for experimental groups were expressed as mean ± SD and compared with those of control groups or different treatment groups using the single-factor analysis of variance. When statistical significance (P < .05) was reached with the F test, comparisons of the means were then performed using either the Tukey-HSD test or the Student t test. A P value lower than 0.05 was considered significant.

Experimental protocol design

The remarkable ability of ODNs to inhibit genetic expression in a sequence-specific manner has become, in recent years, the subject of increased scrutiny in the field of antisense technology. This issue arises because the mechanism of action of these compounds remains unclear and can apparently include antisense and a variety of non-antisense mechanisms simultaneously.^11,26,27,28 Here, we have investigated the influence of several factors - such as backbone modification, cell culture model, dose, sequence of the ODN, type of control used, and delivery approach - on the sequence specificity of activity of antisense ODNs against HIV-1. Antiviral activity of antisense ODNs could be evaluated by measuring reverse transcriptase activity as well. The inhibitory activity can be more precisely determined by quantifying the specific protein targeted by the antisense molecule and could be best suited to estimate the true antisense mechanism. Non-antisense mechanisms may be determined by evaluating ODNs' capacity to hybridize with viral and/or host sequences of a proteic or nucleic nature. ODN activity, especially non-antisense activity, could furthermore be explored by investigating biological mechanisms such as measuring cellular, metabolically active proteins to appreciate health of treated cells. This study focused only on the contribution of the sequence-specific anti-HIV activity of antisense ODNs. Inhibition of HIV replication was conceived as the strategic end point for treating infection with antisense ODNs. Consequently, p24 assay was considered solely for sensitive monitoring of ODN activity. In addition, this study did not attempt to investigate why several ODN sequences tested were more active than others.

Antisense compounds are single-stranded nucleic acids that, in principle, disrupt the synthesis of a targeted protein by hybridizing in a sequence-dependent manner to the mRNAs that encode it. This mechanism is conventionally termed in the literature and in this report as "antisense." It is well established that^26-32 antisense ODNs might have, in addition, activity that is related to a "non-antisense" effect (ie, not related to an antisense effect) as we have referred to it here. This non-antisense effect can be due to multiple mechanisms comprising those originating from the activity of a specific ODN sequence (termed here as a "sequence-specific" effect) and those originating from the activity of an ODN that is not specific to a particular sequence (termed here as a "non-sequence-specific" effect). Alternatively, we can distinguish these effects as being either the sequence-specific effect (comprising antisense and non-antisense sequence-specific effects) or the non-sequence-specific effect.

The experimental design of this work aimed at dissecting the contributions of both sequence-specific and non-sequence-specific interactions to the HIV-inhibitory properties of ODNs. The results are expressed as percent inhibition of p24 production in culture cell medium. We, arbitrarily, defined the sequence-specific activity as the percent of the difference between the inhibition level of the most active antisense ODNs and the random control ODNs; and the non-sequence-specific activity as the percent of the inhibition level of the random control ODN to the inhibition level of the most active ODN.

In our study, part of the activity was dependent on the sequence of antisense sequences and of some control sequences. Non-sequence-specific activity was observed with the random control sequences, and this activity can be attributed to all ODNs. The concentration of a single sequence in a preparation of random 28-mer ODNs is negligible, as it is more than 7 x 1016 times lower than the concentration of total ODNs in the preparation. To study the effect of PO-ODNs and to avoid extracellular biodegradation and cell surface non-specific interference of ODNs with virus entry, we used a carrier system. We used the DLS delivery system, which efficiently protected PO-ODNs from nuclease degradation and increased intracellular availability, allowing them to exert their activity against HIV-1 at low concentration.³³ In order to understand the importance of specific inhibition by antisense molecules, we compared the antiviral activity of either free or DLS-associated PS-ODNs or PO-ODNs in 4 different cell assays, including 3 acute infection models (short-term and long-term assays with HIV-1 [IIIB]-infected MOLT-3 cells and a short-term assay with HIV-1 [IIIB]-infected PBMCs), with a chronic infection model (with H9/HTLV-III cell line).

Cell culture models and antiviral activity

Use of various cell culture models in this work allows for the quantitative comparison of ODN activity, as significant differences (P < .05) were obtained for numerous experimental groups.

All free ODNs showed significant inhibition (> 80%) at the highest dose tested. Nevertheless, viral replication appears to be affected differently depending on the cell culture model used, with the following growing order upon estimated IC50: chronically infected H9 cells < acutely infected PBMCs = acutely infected MOLT-3 cells in a short-term assay < acutely infected MOLT-3 cells in a long-term assay (Tables 2, cell culture models used in this work 3, 4, 5). All free PS-ODNs tested showed the highest activity in the long-term assay for MOLT-3 acutely infected cells with >99% inhibition at 1 µM; in addition, SPac and SDIS showed >99% inhibition at 0.1 µM.

Use of a delivery approach

Use of DLS to deliver ODN does not improve the optimal viral inhibition of ODN when compared with ODN in a free form. Substantial antiviral activity with DLS-associated PS-ODNs was obtained with up to 83% inhibition, although we never observed the very high level of inhibition of viral replication seen when using relatively large amounts of free PS-ODNs. In contrast, when using free ODNs, DLS-delivered ODNs affected viral replication (on the basis of estimated IC50) in an inverse manner: acutely infected MOLT-3 cells in a long-term assay = acutely infected PBMCs < acutely infected MOLT-3 cells in a short-term assay < chronically infected H9 cells. Overall, the results did not indicate any statistical difference between the activities of the ODNs tested, and there was no significant difference between PS and PO activities when they were delivered with DLS to cells. DLS liposomes used as controls at the equivalent concentrations used for DLS-ODNs at effective doses did not show any effect on HIV replication in MOLT-3 cells, and no major cytotoxicity could be observed with either free or DLS-complexed ODNs at effective doses in MOLT-3 and chronically infected H9 cells (Table 6).

In all in vitro systems tested, DLS delivery did not improve the level of antiviral activity of PS-ODNs compared to that of free PS-ODNs, and DLS-unmodified ODNs were not more potent than free PS-ODNs when maximal inhibition is considered. On the basis of our results, the use of free PS-ODNs seems to be the best approach to block HIV replication in vitro. However, at IC50, DLS delivery greatly improved cellular uptake since inhibition of viral replication with DLS-delivered ODNs could be achieved at subnanomolar concentrations compared to micromolar concentrations with free ODNs (Table 7). This confirms that the DLS carrier system appears to be an efficient approach for ODN delivery. An IC50 in the nanomolar range is clearly of high pharmacological interest. IC50 around 0.001 nM was never observed for any in vitro experimental protocol using either free ODN or ODN complexed with a carrier. The DLS formulation has also been tested for its in vivo efficiency for systemic gene delivery in mice. After a single injection of DLS-complexed plasmid DNA, the transgene was detected in lung, liver, spleen, and heart, and the transgene mRNA was detected in lung and spleen for as long as 3 months.²³ In addition, DLS- DNA showed a rapid cellular uptake, a high cytoplasmic and nuclear distribution of DNA in cell cultures, and a relatively low plasma clearance following intravenous administration in mice.²⁴ Therefore, the DLS system may represent a powerful tool for in vivo application of gene therapeutics.

Sequence-specific effect

In this study, the level of antiviral activity of ODNs varied depending on the cell assay used, and the level of the activity that was due to a sequence-specific effect varied depending on the type of control used to calculate the specificity of the inhibition. We found that the portion of the activity of the antisense molecules that could be attributed to a sequence-dependent mechanism was more considerable in our short-term assays with acutely infected cells than in our long-term assay with acutely infected MOLT-3 cells and our short-term assay with chronically infected cells (Table 7). When using the acutely infected MOLT-3 cell culture assay, specific inhibition was dose-dependent (Table 2). At 0.01 µM, SPac displayed antiviral activity (53% ± 2% inhibition of viral replication), whereas SDIS and random control completely failed to inhibit HIV-1 replication (0% inhibition for each). At 0.1 µM, all ODNs exhibited substantial antiviral activity varying from 76% ± 11% to 92% ± 3% inhibition. In contrast, no inhibition could be achieved with the random control at this same concentration. Therefore, at 0.1 µM, all the antiviral activity (100% of the activity) observed with antisense sequences could be attributed to a sequence-specific mechanism (P = .05). When oligonucleotide doses were increased to 1 µM, 81% ± 4% to 98% ± 2% inhibition was obtained with ODN sequences, compared to 96% ± 1% inhibition with the random control. Thus, at higher concentration, 100% of the inhibitory activity was due to a non-antisense mechanism. Moreover, at optimal viral inhibition, antiviral activity appeared essentially to result from non-sequence-specific effects in all other cell culture models used in this work (Tables 2, 3, 4, 5, 7concentrations compared to MOLT-3 cells ). Consequently, a sequence-specific effect was observed for free antisense ODN in the acutely infected MOLT-3 short-term and long-term assays at concentrations lower than 1 µM and 0.1 µM, respectively, and was observed for DLS-delivered antisense ODN in the acutely infected PBMCs and in the chronically infected H9 cells at concentrations of 10 nM and less than 0.01 nM, respectively.

Previously described non-specific effects

Non-specific effects of PS-ODNs have been reported in many studies. PS-ODNs are polyanions and, as a consequence, they are able to bind to various cellular proteins on the basis of charge interaction and base sequence (aptamer approach).^27,34 The presence of a G-quartet, along with particular sequences flanking the G-quartet, may also enhance non-specific effects.²⁸ Furthermore, CpG motifs have been found to be highly immunostimulatory in mice.^35,36 PS-ODNs containing the dinucleotide motif CpG can increase immunoglobulin secretion and expression of B-cell activation markers such as major histocompatibility complex class II, induce interferons (INFs), augment natural killer cell activity, and stimulate the release of several interleukins (ILs) from T cells. It is possible that release of ILs, such as INF-gamma, tumor necrosis factor α, or IL-12, may be involved in anti-HIV activity in the assays used in this study, thus contributing to the random ODNs' high activity. PS-ODNs have been shown to inhibit the replication of HIV-1 in vitro by both sequence-specific^19,29,37,38 and non-sequence-specific^29-32 processes depending on the cell culture model employed. In acute-infection models, non-antisense inhibition of PS-ODNs has been well documented¹ and might occur, because of interference with virus adsorption, by binding to the CD4 receptor or the V3 loop of viral gp120,^39-41 or with reverse transcription.⁴² In chronic-infection models, sequence-specific inhibition of HIV production has been reported,^19,32 but the precise mechanism of action of PS-ODNs still remains to be elucidated. Furthermore, it might be possible that non-antisense sequence-specific activity could be partly attributed to ODN interaction with cellular life proteins such as proteins interfering with transcription, phosphorylation, and cell cycle.

Cellular compartmentalization of ODN non-specific effect

In comparing the activity of free PS-ODNs with DLS-delivered PS-ODNs, it is possible to evaluate the amount of activity that can be attributed to extracellular and/or membranar effects such as interaction with virus by binding to the gp120 viral protein and interaction with cell surface molecules. In fact, DLS complexation prevents ODN interaction with virus and/or cellular membrane and then bypasses the non-specific effects that result from ODN binding to either the CD4 receptor or the gp120 viral protein. We found in our short-term acute-infection model using MOLT-3 cell line that approximately 25% of the activity of PS-ODNs can be attributed to extracellular and/or membranar effects (P = .03). It is interesting that in our PBMC assay and in our chronic-infection model, no significant difference between the level of the inhibition of free PS-ODNs and that of DLS-delivered PS-ODNs could be observed, suggesting that the portion of the activity that can be attributed to non-specific extracellular and/or membranar effects may be less important in primary cells and chronically infected cells than it is in acutely infected established cell lines. With regard to this hypothesis, the higher level of inhibition observed in MOLT-3 cells treated with free PS-ODNs compared to that in primary cells (Table 7) might be explained, in part, by the addition of extracellular and/or membranar effects on overall antiviral activity in immortalized cell lines. In chronically infected cells, PS-ODNs cannot interact with virus entry or reverse transcription steps that occur only in acute infections and, thus, can only interfere with post-integration steps. As such, free PS-ODNs and DLS- PS-ODNs might both exert their activity only in the cytoplasm and/or the nucleus, explaining in part why free PS-ODNs and DLS- PS-ODNs showed the same maximum level of activity in this model. Furthermore, the difference in the levels of the inhibition between DLS- PS-ODNs and DLS- PO-ODNs was more important in our assays using MOLT-3 cells than in our assays using either primary or chronically infected cells, suggesting that non-specific intracellular effects due to backbone modification are more apparent in immortalized cell lines.

In addition, intracellular uptake and bioavailability of ODNs may vary, on the one hand, between immortalized and primary cells⁴³ and, on the other hand, between acutely and chronically infected cells.¹⁵ The cellular uptake and the intracellular distribution of the free PS-ODN GEM 91 was evaluated in MOLT-3 cells and PBMCs by flow cytofluorometry and laser-assisted confocal microscopy for uptake and biodistribution, respectively.⁴³ The cellular uptake was slow and similar in both MOLT-3 cells and PBMCs. However, in MOLT-3 cells, the intracellular distribution of GEM 91 was mainly concentrated in cytoplasmic vesicles, in contrast to PBMCs, in which the intracellular distribution of GEM 91 was more diffuse and, thus, more available for cytoplasmic and nuclear activity. Complexation with DLS formulation may protect ODNs from degradative enzymes in endosomal vesicles and allow bypassing vesicular trafficking, resulting in a higher intracellular and intranuclear localization compared to free ODNs.²⁴ In this study, when PS-ODNs were delivered by DLS, the level of activity achieved with the DLS-delivered PS-ODNs was similar in both MOLT-3 cells and PBMCs in our short-term assays (Table 7), but in PBMCs, high antiviral activity was observed at lower concentrations compared to MOLT-3 cells (Tables 2,4). This difference suggests that the cellular uptake of DLS- PS-ODNs might be higher in PBMCs than in MOLT-3 cells or that the intracellular distribution of DLS- PS-ODNs might be more diffuse in PBMCs than in MOLT-3 cells, as previously observed.⁴³ The highest IC50 of DLS- ODNs was observed in chronically infected cells in which a high level of inhibition was observed at the very low concentration of 1 pM. This result might be explained, in part, by a higher level of cellular uptake and/or a higher intranuclear localization compared to acutely infected cells. In this chronic-infection model, an increased intranuclear localization may be advantageous compared to an acute-infection model and may result in increased inhibitory activity, since in chronically infected cells, the site of action of ODNs is expected to be predominantly in the nucleus where the viral transpose acid and nucleic acid is integrated into the cellular genome. Taken together, our results suggest that most of the PS-ODNs' activity should be due to intracellular activity. Assays on ODN's activity using primary cells should be promoted over assays using established cell lines, since primary cells are more representative of in vivo infections and since cellular uptake and intracellular distribution may vary among cell models employed.

Antisense-Dependent vs Non-Specific Activities

In the literature, most of the studies that report antisense-dependent activity of antisense molecules have been done in short-term acute-infection assays.^31,37,38,44 In the present study, the highest level of sequence-specific activity of our antisense PS-ODNs and PO-ODNs, either free or DLS-delivered, was observed in our short-term assay using acutely infected MOLT-3 cells with HIV-1 (IIIB) (Table 7). In agreement with our data, other investigators also observed, in a long-term acute-infection assay using the MOLT-3 cell line, a high level of antiviral activity in HIV-infected cell cultures treated with 28-mer random and mismatch sequences for up to 21 days postinfection.²⁹ However, in contrast to our results, antisense-dependent activity of GEM 91 was observed in another study⁴³ in which the activity of GEM 91 was compared to a random control in a long-term assay using acutely infected MOLT-3 cells with HIV-1 (IIIB).

In another experiment in which the efficacy of the 3 antisense PS-ODNs, Srev, SDIS, and Spac, to block the replication of HIV-1 clinical isolate VR2846A was evaluated in infected PBMCs, SDIS and SPac showed a significant sequence-specific activity for up to 14 days when compared to their sense control, but no antisense-dependent activity could be found when compared to the random control at 0.1 µM (Lavigne et al, unpublished observations, 2001). Non-antisense-dependent antiviral activity was also observed in PBMCs infected with the HIV-1 clinical isolate 571 with 1 µM GEM 91 compared to a random sequence 7 days after infection.⁴³ However, after 11 days, the level of inhibition started to decrease in cells treated with the random control (74% inhibition at day 11 and 0% at day 14) compared to cells treated with GEM 91 (>99% inhibition up to day 14).

In our chronic-infection model, we did not observe significant sequence-specific activity of our free antisense PS-ODNs. Non-sequence-specific inhibition observed in our chronic-infection model might be explained, in part, by direct interactions with proteins or by hybridization to other mRNA targets.^4,28 In a recent study, inhibition of virus production in chronically infected H9 cells was evaluated at about 60% (as measured by p24 determination) with either the antisense molecule GEM 91 or the mismatched oligo control at 1 µM, suggesting that inhibition of HIV-1 production in this chronic-infection model was also a non-sequence-specific phenomenon.⁴¹ However, these authors found that treatment of the chronically infected H9 cells with GEM 91 caused a significant decrease in gag mRNA expression (60%-70% inhibition), while mismatched oligo treatment resulted in gag mRNA expression similar to the control, suggesting that GEM 91 had a sequence-specific effect at this stage of the viral-replicative cycle. Nevertheless, a reduction in mRNA level does not necessarily result in biological activity since a lower mRNA level might be sufficient for protein function. Taken together, the above considerations indicate that the nature of the activity of ODNs (antisense-dependent vs non-specific activity) may vary depending on the type of infection and length of the assay, and conclusions regarding ODNs' activity may differ depending on the stage of HIV replication investigated.

Inhibition of HIV-1 replication by antisense ODNs depends on various factors in vitro that should be taken into consideration for the design of any study on antisense ODNs. For instance, the use of an efficient delivery system has made it possible to point out extracellular and/or membranar activity of PS-ODNs that appears as the most effective presentation among those tested for the antisense ODN-mediated inhibition of HIV replication. In light of our results and previous observations, the contribution of sequence-specific effect and, consequently, antisense-mediated mode of action appears minor compared to the other mechanisms involved in ODN antiviral activity. As long as the inhibition of HIV replication is concerned, the term "antisense" does not appear appropriate to identify this class of compound. As the antisense sequences used in this study, and in others, were specifically designed for eliciting the true antisense mechanism, it appears more promising to shift the oligonucleotide compound design to a design adapted for other modes of action such as the aptamer^27,45 or the sequence-specific immunostimulation approaches.^35,36 This might result in better adjustment to the universal pharmacological paradigm of the structure/function rationale and, consequently, in development of more effective ODN therapeutics.

This work was supported by grants from the Medical Research Council of Canada. C. Lavigne benefited from postgraduate scholarships from the Fonds pour la formation de chercheurs et l'aide à la recherche (FCAR). We thank Mrs M. Fauvel for giving access to the Laboratoire de Santé Publique du Québec's P3 facilities that are used for virus manipulation.

1.   Stein CA, Tonkinson JL, Yakubov L. Phosphorothioate oligodeoxynucleotides-antisense inhibitors of gene expression? Pharmacol Ther. 1991;52:365-384.
 

2.   Zamecnik PC, Stephenson M. Inhibition of Rous sarcoma virus replication and transformation by a specific oligodeoxynucleotide. Proc Natl Acad Sci U S A. 1978;75:280-284.
 

3.   Zamecnik PC, Goodchild J, Taguchi Y, Sarin PS. Inhibition of replication and expression of human T-cell lymphotropic type III in cultured cells by exogenous synthetic oligonucleotides complementary to viral RNA. Proc Natl Acad Sci U S A. 1986;83:4143-4146.
 

4.   Stein CA, Cheng Y-C. Antisense oligonucleotides as therapeutics agents: is the bullet really magical? Science. 1993;261:1004-1012.
 

5.   Agrawal S, Temsamani J, Tang J-Y. Pharmacokinetics biodistribution, and stability of oligodeoxynucleotide phosphorothioates in mice. Proc Natl Acad Sci USA. 1991;88:7595-7599.
 

6.   Bennett CF, Dean N, Ecker DJ, Monia BP. Pharmacology of antisense therapeutic agents: cancer and inflammation. In: Agrawal S, ed. Antisense Therapeutics. [city?], NJ: Humana Press; 1996:13-46.
 

7.   Iversen PL, Mata J, Tracewell WG, Zon G. Pharmacokinetics of an antisense phosphorothioate oligodeoxynucleotide against rev from human immunodeficiency virus type 1 in adult male rat following single injections and continuous infusion. Antisense Res Dev. 1994;4:43-52.
 

8.   Agrawal S, Temsamini J, Galbraith W, Tang J-Y. Pharmacokinetics of antisense oligonucleotides. Clin Pharmacokinet. 1995;28:7-16.
 

9.   Roush W. Antisense aims for a renaissance. Science. 1997;276:1192-1193.
 

10.   Wagner RW. The state of the art in antisense research. Nature Medecine. 1995;1:1116-1118.
 

11.   Stein CA, Kreig AM. Problems in interpretation of data derived from in vitro and in vivo use of antisense oligodeoxynucleotides. Antisense Res Dev. 1994;4:67-69.
 

12.   Wagner RW. Gene inhibition using antisense oligodeoxynucleotides. Nature. 1994;372:333-335.
 

13.   Thierry AR, Tackle GB. Liposomes as a delivery system for antisense and rybozyme compounds. In: Akhtar S, ed. Delivery Strategies for Antisense Oligonucleotide Therapeutics. Boca Raton, FL: CRC Press; 1995:199-221.
 

14.   Thierry AR, Dritschilo A. Intracellular availability of unmodified, phosphorothioated and liposomally encapsulated oligodeoxynucleotides for antisense activity. Nucleic Acids Res. 1992;21:5691-5698.
 

15.   Lavigne C, Thierry AR. Enhanced antisense inhibition of human immunodeficiency virus type 1 in cell cultures by DLS delivery system. Biochem Biophys Res Commun. 1997;237:566-571.
 

16.   Kim H-J, Lee K, O'Rear JJ. A short sequence upstream of the 5' Major splice site is important for encapsidation of HIV-1 genomic RNA. Virology. 1994;198:336-340.
 

17.   Muriaux D, Girard P-M, Bonnet-Mathonière B, Paoletti J. Dimerization of HIV-1Lai RNA at low ionic strength. J Biol Chem. 1995;270:8209-8216.
 

18.   Agrawal S, Tang JY. GEM 91-an antisense oligonucleotide phosphorothioate as a therapeutic agent for AIDS. Antisense Res Dev. 1992;2:261-266.
 

19.   Matsukura M, Zon G, Shinozuka K, et al. Regulation of viral expression of human immunodeficiency virus in vitro by an antisense phosphorothioate oligodeoxynucleotide against rev (art/trs) in chronically infected cells. Proc Natl Acad Sci U S A. 1989;86:4244-4248.
 

20.   Ojwang JO, Buckheit RW, Pommier Y, et al. T30177, an oligonucleotide stabilized by an intramolecular guanosine octet, is a potent inhibitor of laboratory strains and clinical isolates of human immunodeficiency virus type 1. Antimicrob Agents Chemother. 1995;39:2426-2435.
 

21.   Popovic M, Read-Conole E, Gallo RC. T4 positive human neoplastic cell lines susceptible to and permissive for HTLV-III. Lancet. 1984;II:1472-1473.
 

22.   Popovic M, Sarngadharan MG, Read E, Gallo RC. Detection, isolation, and continuous production of cytopathic retroviruses (HTLV-III) from patients with AIDS and pre-AIDS. Science. 1984;224:497-500.
 

23.   Thierry AR, Lunardi-Iskandar Y, Bryant JL, Rabinovich P, Gallo RC, Mahan LC. Systemic gene therapy: biodistribution and long-term expression of a transgene in mice. Proc Natl Acad Sci USA. 1995;92:9742-9746.
 

24.   Thierry AR, Rabinovich P, Peng B, Mahan LC, Bryant JL, Gallo RC. Characterization of liposome-mediated gene delivery: expression, stability and pharmacokinetics of plasmid DNA. Gene Ther. 1997;4:226-237.
 

25.   Pauwels R, Balzarini J, Baba M, et al. Rapid and automated tetrazolium based colorimetric assay for the detection of anti-HIV compounds. J Virol Methods. 1988;20:309-321.
 

26.   Gura T. Antisense has growing pains. Science. 1995;270:575-577.
 

27.   Stein CA. Does antisense exist? Nature Medecine. 1995;1:1119-1121.
 

28.   Stein CA. Phosphorothioate antisense oligodeoxynucleotides: questions of specificity. Trends Biotechnol. 1996;14:147-149.
 

29.   Lisziewicz J, Sun D, Metelev V, Zamecnik P, Gallo RC, Agrawal S. Long-term treatment of human immunodeficiency virus-infected cells with antisense oligonucleotide phosphorothioates. Proc Natl Acad Sci U S A. 1993;90:3860-3864.
 

30.   Agrawal S, Goodchild J, Civeira M, Sarin PS, Zamecnik PC. Nucleosides Nucleotides. 1989;8: 819-823.
 

31.   Kinchington D, Galpin S, Jaroszewski JW, Ghosh K, Subasinghe C, Cohen JS. A comparison of gag, pol and rev antisense oligodeoxynucleotides as inhibitors of HIV-1. Antivir Res. 1992;17:53-62.
 

32.   Zelphati O, Imbach J-L, Signoret N, Zon G, Rayner B, Leserman L. Antisense oligonucleotides in solution or encapsulated in immunoliposomes inhibit replication of HIV-1 by several different mechanisms. Nucleic Acids Res. 1994;22:4307-4314.
 

33.   Lavigne C, Thierry AR. Lipid-based delivery of combinations of antisense oligodeoxynucleotides for the in vitro inhibition of HIV-1 replication. AAPS Pharmsci. 2001;3:article 7. Available at: http://www.aapspharmaceutica.com/scientificjournals/pharmsci/journal/01_07.html. Accessed February 12, 2001.
 

34.   Jendis J, Strack B, Volkmann S, Boni J, Molling K. Inhibition of replication of fresh HIV type 1 patient isolates by a polypurine tract-specific self-complementary oligodeoxynucleotide. AIDS Res Hum Retroviruses. 1996;12:1161-1168.
 

35.   Tamamoto T, Yamamoto S, Kataoka T, Tokunaga T. Ability of oligonucleotides with certain palindromes to induce interferon production and augment natural killer cell activity is associated with their base length. Antisense Res Dev. 1994;4:119-122.
 

36.   Krieg A, Yi,A-K, Matson,S, et al. CpG motifs in bacterial DNA trigger direct B-cell activation. Nature. 1995;374:546-549.
 

37.   Agrawal S, Ikeuchi T, Sun D, et al. Inhibition of human immunodeficiency virus in early infected and chronically infected cells by antisense oligodeoxynucleotide and its phosphorothioate analogue. Proc Natl Acad Sci U S A. 1989;86:7790-7794.
 

38.   Anazodo MI, Wainberg MA, Friesen AD, Wright JA. Sequence-specific inhibition of gene expression by a novel antisense oligodeoxynucleotide phosphorothioate directed against a nonregulatory region of the human immunodeficiency virus type 1 genome. J Virol. 1995;69:1794-1801.
 

39.   Stein CA, Cleary AM, Yakubov L, Lederman S. Phosphorothioate oligodeoxynucleotides bind to the third variable loop domain (v3) of human immunodeficiency virus type 1 gp120. Antisense Res Dev. 1993;3:9-31.
 

40.   Yakubov L, Khaled Z, Zhang L-M, Truneh A, Vlassov V, Stein CA. Mode of interaction of oligodeoxynucleotides with recombinant sCD4. J Biol Chem. 1993;268:18818-18823.
 

41.   Yamaguchi K, Papp B, Zhang D, Ali AN, Agrawal S, Byrn RA. The multiple inhibitory mechanisms of GEM 91, a gag antisense phosphorothioate oligonucleotide, for human immunodeficiency virus type 1. AIDS Res Hum Retroviruses. 1997;13:545-554.
 

42.   Majumdar C, Stein CA, Cohen JS, Broder S, Wilson SH. Stepwise mechanism of HIV reverse transcriptase: Primer function of phosphorothioate oligodeoxynucleotide. Biochemistry. 1989;28:1340-1346.
 

43.   Lisziewicz J, Sun D, Weichold FF, et al. Antisense oligodeoxynucleotide phosphorothioate complementary to Gag mRNA blocks replication of human immunodeficiency virus type 1 in human peripheral blood cells. Proc Natl Acad Sci U S A. 1994;91:7942-7946.
 

44.   Bordier B, Perala-Heale M, Degols G, et al. Sequence-specific inhibition of human immunodeficiency virus (HIV) reverse transcription by antisense oligonucleotides: Comparative study in cell-free assays and in HIV-infected cells. Proc Natl Acad Sci U S A. 1995;92:9383-9387.
 

45.   White RR, Sullenger BA, Rusconi CP. Developing aptamers into therapeutics. J Clin Invest. 2000;106:929-934.