Ensayo de Concreto Endurecido

There are two basic strategies for the application of antisense RNA (or DNA) to cells. The first involves synthesis of the desired molecule in vitro followed by its introduction to cells via microinjection or bulk addition to the culture medium. For technical reasons, microinjection has been used primarily in oocyte and egg systems to study developmental issues. Synthetic oligonucleotides, on the other hand, have been used to inhibit gene expression in a wide range of cell types since they were first described 18 years ago (Zamecnik and Stephenson 1978, Stephenson and Zamecnik 1978). The second approach has been developed more recently (Izant and Weitraub 1984) and involves the creation of eukaryotic expression vectors which can transcribe a desired antisense sequence within transfected target cells. Stable transfection of cultured cells with antisense RNA-expressing vectors allows continuous generation of antisense sequences within the cell itself.

1.3.1.1 The oligonucleotide-based approach

The key parameter for antisense inhibition by an oligonucleotide is its intracellular concentration. In this respect two major obstacles must be overcome: poor penetration across cell membranes and degradation of antisense molecules in the growth medium and the cell (Wagner 1994).

The precise mechanism of oligonucleotide uptake is unclear (Budker et al 1992). Uptake efficiency is considerably higher at low oligonucleotide concentrations, suggesting that the mechanism is saturatable and there are a limited number of nucleic acid binding centres on the cell surface (Yakubov et al 1989). An 80kDa protein has been isolated from HL-60 cells and shown to bind oligo- and polynucleotides (Loke et al 1989). Similar proteins have been detected in different cell lines (Vlassov et al 1991) which may have a role in endocytosis of oligonucleotides. In lymphoid cells uptake can only be partially inhibited with competitor DNA, suggesting that oligonucleotide uptake can occur by two or more mechanisms, one of which is not competed and therefore not receptor mediated (Krieg et al 1991).

Attachment of endocytosable or lipophyllic groups can improve uptake of an oligonucleotide. Conjugation with poly-L-lysine (Leonetti et al 1990), cholesteryl (Boutorin et al 1989), alkyl groups (Kabanov et al 1990) or lipid residues (Shea et al

1990) can increase the capacity of an oligonucleotide to penetrate cell membranes. Alternatively, delivery of oligonucleotides can be enhanced by encapsulation in membrane carriers such as liposomes (Bennet et al 1992a), viral envelopes or erythrocyte ghosts (Vlassov et al 1989).

In addition to inefficient uptake of oligonucleotide, there is no guarantee that the oligonucleotide will be sequestered in the desired intracellular compartment once inside the cell. In some cells intact radio-labeled or fluorescent oligomers have been detected in the nucleus and the cytoplasm (Teichman-Weinberg et al 1988), but in others the oligonucleotides appeared to be trapped in the endosomes and not available at the site of translation (C. Hélène unpublished data).

To avoid nuclease attack on oligonucleotides in culture, medium is often used with no serum, or serum which has been heat-treated at a high temperature to inactivate the nucleases. Also it is possible to chemically modify oligonucleotides to increase their resistance to nuclease attack. Possible backbone modifications include synthesis of phosphotriester, methylphosphonate, phosphoramidates, a-anom ers, phosphoroselenoate and phosphorothioate oligonucleotide analogues.

Phosphorothioate analogues have been investigated the most extensively (Stein et al 1991). In these compounds a sulphur atom replaces one of the non-bridging oxygen atoms at each interbase phosphorous atom (Figure 1.10). Substitution of S- for O- retains the original aqueous solubility properties but confers nuclease resistance. Phosphorothioate oligonucleotides also have a higher affinity for the SOkDa putative transport protein (Neckers 1989). Finally, from the hst of possible backbone modifications, only native and phosphorothioate oligos form substrates for RNase H when they hybridize to their target sequence (section 1.3.2.3).

Non-specific effects can be a problem when using antisense oligonucleotides. Charged oligodeoxynucleotides are polyanions and may mimic the effect of biological polyanions such as heparin and heparan, and dermatan and chondroitin sulphates, and interact with other cellular proteins (Stein and Krieg 1994). These effects can be sequence specific. For example oligonucleotides containing four contiguous guanine residues may be antiproliferative (Yaswen et al 1992). Palindromic regions of six or more bases can induce interferon a and y production (Yamamoto et al 1992b). Non­ specific cellular activation of the SPl transcription factor by a phosphorothioate oligonucleotide has recently been reported (Perez et al 1994). Antisense oligonucleotides have been shown in Xenopus to cause cleavage of imperfectly matched target sites (Woolf et al 1992). Finally, nucleosides and nucleotides, the degradation products of oligonucleotides, can affect cell proliferation and differentiation (Rathbone et al 1992, Kamano et al 1992). To demonstrate genuine antisense sequence-specific inhibition of genetic expression it is therefore important to include many control oligonucleotides and to show a decrease in target protein directly. Commonly used control oligonucleotides include: (i) sense, which are in the reverse orientation to the antisense oligonucleotide, (ii) scrambled, in which the bases are in a random order and (iii) mismatch which have the same sequence as the antisense oligonucleotide apart from one or two mismatches in the central section of the molecule.

HOCH CH 0 -phosphodiester S -phosphorothioate OEt -phosphotriester Me -m ethylphosphonate OH

Figure 1.10 Oligonucleotide backbone modifications

1.3.1.2 The vector-based approach

Intracellular production o f antisense RNA by a transfected vector bypasses the problem of uptake and incorrect intracellular location of antisense sequences. Inducible vector systems have been developed which enable reversible antisense production within cells. This is particularly useful for studying genes that have a role in cell growth or division, as constitutive antisense expression of such genes may be lethal for the transfectants. Inducible systems often employ a heavy metal-inducible metallothionein promoter (Mayo et al 1982) or the glucocorticoid-inducible murine mammary tumor virus long terminal repeat promoter (Lee et al 1981). However both these promoters have significant baseline transcription in the absence of induction. A new system has recently been developed to overcome this leaky expression, using the E.coli tetracycline resistance operon (Gossen et al 1995). In this system the promoter for the antisense sequence contains a tet operon upstream of its promoter sequence which is recognized by the tetracycline-controlled transactivator (a fusion between the Tet repressor and a transcription activation domain) (Figure 7.1). Addition of tetracycline prevents the transactivator fi^om binding the tet operon, so no antisense sequence is expressed. The interaction between the Tet repressor and both its target sequence (tetO) and tetracycline is highly specific, so the system has minimal baseline activity.

Vector integration associated with stable transfection can produce dramatic positional effects on vector transcriptional activity and cause heterogeneous antisense expression. This necessitates the screening o f multiple transfected clones. Episomally-replicating vectors can be used to avoid the positional effects associated with chromosomal integration and also to increase the vector copy number within cells (Hambor et al

1988). With episomal and even conventional integrating vector systems, a gradual loss o f antisense is often observed during prolonged culture (Kasid et al 1989, Bolen et al

1987). To avoid such loss, it is often necessary to study low passage clones.

The chosen length for ohgonucleotides is usually fairly short (less than 30 bases) due to limitations of cellular uptake and the expense associated with long oligonucleotides.

produced. However, there is little information available on the optimal length for transcribed antisense RNA. Long antisense sequences may have greater specificity, but they may also be unwieldy and form complicating secondary structures. If there is a race between degradation of antisense RNA and binding to the specific target sequence, a longer antisense would have a better chance of reaching its target before it is completely degraded. It has even been suggested that if long antisense RNAs are partially degraded intracellularly, each fragment can still act as an individual antisense inhibitor (Nellen and Lichtenstein 1993). Many vector-based antisense experiments use the entire coding sequence of the target protein in an antisense orientation. Paradoxically, this approach can lead to increased production of the target protein due to transcription of the opposite ("sense") strand of the transfected antisense construct (Williard et al 1994).