#apaperaday: Rational Design of Chimeric Antisense Oligonucleotides on a Mixed PO–PS Backbone for Splice-Switching Applications

In today’s #apaperaday, Prof. Aartsma-Rus reads and comments on the paper titled: Rational Design of Chimeric Antisense Oligonucleotides on a Mixed PO–PS Backbone for Splice-Switching Applications

Today’s pick is from Biomol_MDPI by Le et al. DOI: 10.3390/biom14070883. The paper is on phosphorothioate (PS) and phosphorodiester (PO) backbones for splice modulating antisense oligonucleotides (ASOs). The PS backbone is often used in therapeutic ASOs as it prevents renal clearance and thus is key to bioavailability. It also helps ASOs cross membranes.

However, the PS modification makes each bond between nucleotides chiral, meaning there is a left and a right version that is not symmetrical (like hands). The different versions have different characteristics with regards to stability and affinity.

Note that currently approved PS ASOs contain a mix of different isomers, where each PS position can have a left or right (Rp or Sp) linkage, so this leads to 2(x) variations, where x is the number of PS linkages (~half a million different individual isomers for 19 linkages).

Authors here aim to reduce the effect of isomers by reducing the number of PS bonds and using PO bonds instead. All ASOs have 2OMe RNA, but different types of PS/PO combinations. They use a full PS and a full PO ASO for mouse dystrophin exon 23 skipping as references.

First, authors notice that the Tm (affinity) for the full PO is lower than the full PS. That is not unexpected, as POs reduce affinity for RNA. For the ASOs with both PS and PO, the ASOs having longer stretches of POS have a lower melting temperature than the mix (POPSPOPSetc).

Authors then look into the exon skipping efficiency, finding that the full PS ASO induces the most skip, followed by the mixmers (where the ASOs with PO flanks did worse than PS flanks or the POPSPOPS ASO) and the full PO ASO performed worst.

Again, this is not surprising, because the PS backbone helps with delivery of the ASOs even with transfection. So the differences likely reflect differences in delivery rather than stereoisomer differences (although authors try to argue the isomer angle is the cause).

Authors assessed the stability of ASOs, and not surprisingly, ASOs with PSs at their ends are more stable than ASOs with POs at their ends. Authors discuss that Fritz Eckstein developed the PS backbone—indeed, and this made him the first OTSociety lifetime achievement award recipient.

Authors argue the Rp bond is more stable than the Sp bond and that this difference may underlie the difference they observe between the mixmers and the full PS. I do not think this is the case as mentioned, more likely the effect is purely due to PS being taken up better by cells.

Authors also argue that using mixmers may lead to better ASOs. However, for clinical translation, they should have also discussed the fact that we need a specific number of PS bonds to prevent renal clearance, as otherwise there will be less/no protein binding and filtration.