Antisense Oligonucleotide Chemistry

The most advanced oligonucleotide chemistry used for antisense drugs is phosphorothioate oligodeoxynucleotides. These differ from natural DNA in that 1 of the nonbridging oxygen atoms in phosphodiester linkage is substituted with sulfur (Fig. 1). Phosphorothioate oligodeoxynucleotides are commercially available, easily synthesized, support RNase H activity, exhibit acceptable pharmacokinetics for systemic and local delivery, and have not exhibited major toxicities that would prevent their use in humans. There have been significant resources employed to identify chemical modifications that further improve upon the properties of phosphorothioate oli-godeoxynucleotides. The primary objectives of the effort are similar to medicinal chemistry efforts for other types of pharmacological agents (i.e., to increase potency, improve pharmacokinetics, and decrease toxicity).

A dimer of an oligonucleotide depicting subunits that may be modified to enhance oligonucleotide drug properties is depicted in Fig. 3. In naturally occurring nucleic acids, these subunits are composed of heterocycles, carbohydrate sugars, and phosphodiester-based linkages between the sugars. The combination of the carbohydrate sugar (ribose in RNA, 2'-deoxyribose in DNA) and the linkage provides the backbone of the oligonucleotide polymer. Many modifications have been made on the individual base, sugar, and linkage subunits, and the sugar-phosphate backbone has been completely replaced with an appropriate substitute. In Addition, many diverse moieties have been conjugated to various positions in the subunits, mainly in an attempt to alter the biophysical properties of the polymer. Finally, prodrug modifications may be employed to enhance drug properties. Most of the positions available in a nucleoside dimer (approximately 25 positions for each dimer that do not directly interfere with Watson-Crick base pair-hydrogen bonding) have been modified and studied for their effects on the properties of the resulting oligonucleotides.

The nucleobases or heterocycles of nucleic acids provide the recognition points for the Watson-Crick base pairing rules and any oligonucleotide modification must maintain these specific hydrogen-bonding interactions. Thus, the scope of heterocyclic modifications is somewhat limited. The relevant heterocyclic modifications can be grouped into 2 structural classes: (1) those that enhance base stacking, and (2) those that provide additional hydrogen bonding. The primary objective of heterocyclic modifications being to enhance hybridization, resulting in increased affinity (Fig. 4). Modifications that enhance base stacking by expanding the ^-electron cloud are represented by lipophilic modifications in the 5-position of pyrimidines, such as propynes, hexynes, azoles, and a simple methyl group (49-52) and the 7 position of 7-deaza-purines position, including iodo, propynyl, and cyano groups (53-55). Investigators have continued to build out of the 5-position of cytosine by going from the propynes to 5-membered heterocy-cles to tricyclic fused systems emanating from the 4 and 5-positions of (Fig. 4) (56-59). A second type of heterocycle modification is represented by the 2-aminoadenine (Fig. 4), where the additional amino group provides another hydrogen bond in the A-T base pair, analogous to the 3 hydrogen bonds in a G-C base pair. Heterocycle modifications providing a combination of effects are represented by 2-amino-7-deaza-7-modified A (55) and the G-clamp, a tricyclic cytosine analog having hydrogen-bonding capabilities in the major groove of heteroduplexes (58) (Fig. 4). Furthermore, N2-modified 2-amino purine oligonucleotides have exhibited interesting binding properties (60,61). All these modification are positioned to lie in the major or minor groove of the heteroduplex, do not affect sugar conformation of the heteroduplex, and provide little nuclease resistances, but will generally support an RNase H cleavage mechanism.

Figure 4 Examples of different heterocycle modifications that support antisense activity.

Modifications in the ribofuranosyl moiety have provided the most value in the quest to enhance oligonucleotide drug properties (Fig. 5). In particular, certain 2'-O- modifications have greatly increased binding affinity and nuclease resistance, altered pharmacokinetics, and are potentially less toxic (62). Preorganization of the sugar into a 3'-endo pucker conformation is responsible for the increased binding affinity (63-65). The 2'-O-methoxyethoxy (MOE) and 2'-O-methyl modifications (Fig. 5) are the most advanced of the 2'-modi-fied series, and have entered clinical trials. The cationic 2 -

O-aminopropyl (66) and 2 -O-(dimethylaminooxyethyl) (67,68) have shown favorable binding affinity, with dramatically improved nuclease resistance. In an attempt to extend on the increased nuclease resistance of these cationic modifications to the high affinity seen with MOE, a dimethylaminoe-thyl version (DMAEOE) was prepared. This modification displays hybridization properties equal to or superior to those of MOE, and nuclease resistance equal to that of the 2 -O-aminopropyl modification. The modification showing the largest known improvement in binding affinity is a bicyclic

Antisense Oligonucleotides
Figure 5 Examples of different sugar and backbone modifications that support antisense activity.

system having the 4'-carbon tethered to the 2'-hydroxyl group. As this modification ''locks'' the conformation of the ribose sugar into an RNA-like (3'-endo) conformation, it is referred to as locked nucleic acid (LNA) (69,70). LNA shows dramatically improved hybridization properties with regard to a reference DNA:RNA duplex, and has extremely high nuclease resistance. Although extremely promising from early biophysical and in vitro data, whether these properties will translate into improved efficacy in vivo remains to be seen.

It is now well known that uniformly 2'-O-modified oligonucleotides do not support an RNase H mechanism (71). The heteroduplex formed has been shown to present a structural conformation that is recognized by the enzyme, but cleavage is not supported (72-74). Thus, uniformly modified, ''RNA-like'' oligonucleotides (3'-endo sugar conformation) will be unable to effect cleavage of the target mRNA, and must therefore exert their effects via other means. This has led to the development of a chimeric strategy (3,71,75-77), which focuses on the design of high-affinity, nuclease-resis-tant antisense oligonucleotides that contain a ''gap'' of contiguous phosphorothioate- modified oligodeoxynucleotides (Fig. 6). On hybridization to target RNA, a heteroduplex is presented that supports an RNase H-mediated cleavage of the RNA strand via interaction with the 2 -deoxy gap region. The stretch of the modified oligonucleotide-RNA heteroduplex, which is recognized by RNase H may be placed anywhere within the modified oligonucleotide. The modifications in the flanking regions of the gap should not only provide nuclease resistance to exo- and endonucleases, but also not compromise binding affinity and base pair specificity. There are several types of structures that have been successfully developed (Fig. 6), with the most advance being ''gapmers,'' having a 7- to 10-base oligodeoxynucleotide gap flanked by 2 regions of 2'-modified nucleosides. These oligomers, in particular, 2'-MOE modified, show reduced toxicity, increased potency, and superior pharmacokinetics relative to the parent unmodified 20-mer phosphorothioate oligodeoxynucleotide (77-81).

Several possible mechanisms exist for uniformly modified, non-RNase H activating oligonucleotides to show efficacy, such as prevention of assembly of the ribosome through binding in the 5'-UTR, ''translation arrest,'' or ribosome stalling by blocking the reading of the mRNA ribosome, and modulation of splicing events by binding to splice junctions. Although all these strategies have been pursued, no uniformly modified oligonucleotides have advanced beyond gapmer oligonucleo-tides. However, much recent progress has been made with non-RNase H active oligonucleotides, and there remains much potential for these modifications. LNA and MOE have been used in a uniform context in addition to the gapmer strategy, and early studies show promise. Another interesting uniform

Figure 6 Examples of different oligonucleotide structures.

modification is the phosphoramidiate modification, which substitutes an amino group for the 3' oxygen atom of the deoxyribose sugar of DNA. This results in a preference for the RNA-like (3' endo) sugar conformation, and results in increased affinity as is seen with the 2'-O-alkyl modifications (82,83).

One of the most intriguing backbone oligonucleotide modifications is peptide nucleic acid (PNA). PNA is unique in that the sugar-phosphate backbone is completely replaced with a peptide-based backbone (Fig. 5) (84). This results in a polymer with a neutral backbone that has high affinity for complimentary nucleic acids. PNA has been extensively investigated as an antisense agent, but these efforts have generally been frustrated by the poor cellular penetration and in vivo pharmacoki-netic properties of PNA (85). Recently, a 4-lysine peptide conjugated to a PNA was found to provide robust in vivo activity when targeted to a splice junction (47). These data are highly encouraging because they may provide a path to realizing the promise of PNA as an antisense therapeutic agent.

The most advanced uniform modification is the ''mor-pholino'' modification (Fig. 5), which is currently in phase II clinical trials for restenosis, cancer, and polycystic kidney disease. The morpholino modification simultaneously replaces the ribofuranosyl sugar with a morpholine ring, and the negatively charged phosphate ester with a neutral phos-phorodiamidate linkage (86,87). Morpholinos are generally used around the translation initiation start codon, and are believed to function via translation arrest. A morpholino oligonucleotide has shown in vivo activity (88), as well as oral bioavailability in rats (89), which would be a major advance if studies proved general and translated to larger mammalian species.

In addition to heterocycle, backbone, and sugar modification discussed above, various pendant groups have been attached to oligonucleotides, such as cholesterol, folic acid, fatty acids, etc., to alter pharmacokinetic properties (90,91). The reader is referred to several recent reviews that discuss the chemistry of oligonucleotides in more detail (92-95). It should be noted that there is no single modification that covers all the desired properties for a modified oligonucleo-tide. Modifications have been identified that increase hybridization affinity of the oligonucleotide for its target RNA, increase nuclease resistance, decrease toxicity, and alter the pharmacokinetics (Table 2). Furthermore, the ideal oligonu-cleotide will differ for different applications. Therefore, it is important to be able to mix and match the various modifications to obtain the optimal oligonucleotide for the task at hand.

Table 2 Attributes of Various Modified Oligonucleotides



Increased affinity

2'-0-methyl, 2'-fluoro, MOE,

for RNA


0 0

Post a comment