As small molecule drugs for Duchenne muscular dystrophy (DMD), antisense oligonucleotides (AONs) have been shown to restore the disrupted reading frame of DMD transcripts by inducing specific exon skipping. This allows the synthesis of largely functional Becker muscular dystrophy (BMD)-like dystrophins and potential conversion of severe DMD into milder BMD phenotypes. Thus far we have used 2 0 -O-methyl phosphorothioate (2OMePS) AONs. Here, we assessed the skipping efficiencies of different AON analogs containing morpholino-phosphorodiamidate, locked nucleic acid (LNA) or peptide nucleic acid (PNA) backbones. In contrast to PNAs and morpholinos, LNAs have not yet been tested as splice modulators. Compared to the most effective 2OMePS AON directed at exon 46, the LNA induced higher skipping levels in myotubes from a human control (85 versus 20%) and an exon 45 deletion DMD patient (98 versus 75%). The morpholinoinduced skipping levels were only 5-6%, whereas the PNA appeared to be ineffective. Further comparative analysis of LNA and 2OMePS AONs containing up to three mismatches revealed that LNAs, while inducing higher skipping efficiencies, show much less sequence specificity. This limitation increases the risk of adverse effects elsewhere in the human genome. Awaiting further improvements in oligochemistry, we thus consider 2OMePS AONs currently the most favorable compounds, at least for targeted DMD exon 46 skipping.