Short antisense-locked nucleic acids (all-LNAs) correct alternative splicing abnormalities in myotonic dystrophy

Abstract

Myotonic dystrophy type 1 (DM1) is an autosomal dominant multisystemic disorder caused by expansion of CTG triplet repeats in 3'-untranslated region of DMPK gene. The pathomechanism of DM1 is driven by accumulation of toxic transcripts containing expanded CUG repeats (CUG(exp)) in nuclear foci which sequester several factors regulating RNA metabolism, such as Muscleblind-like proteins (MBNLs). In this work, we utilized very short chemically modified antisense oligonucleotides composed exclusively of locked nucleic acids (all-LNAs) complementary to CUG repeats, as potential therapeutic agents against DM1. Our in vitro data demonstrated that very short, 8- or 10-unit all-LNAs effectively bound the CUG repeat RNA and prevented the formation of CUG(exp)/MBNL complexes. In proliferating DM1 cells as well as in skeletal muscles of DM1 mouse model the all-LNAs induced the reduction of the number and size of CUG(exp) foci and corrected MBNL-sensitive alternative splicing defects with high efficacy and specificity. The all-LNAs had low impact on the cellular level of CUG(exp)-containing transcripts and did not affect the expression of other transcripts with short CUG repeats. Our data strongly indicate that short all-LNAs complementary to CUG repeats are a promising therapeutic tool against DM1.

Figure 1.

8, 10 or 12-unit antisense LNA oligomers efficiently bind stable CUG^exp hairpin and correct MBNL-sensitive alternative splicing. (a) EMSA on a native polyacrylamide gel showing the interaction between distinct PO-LNA-CAG oligomers used at indicated concentrations and a 5′-end radiolabeled (CUG)₁₀₀ used at a concentration of 0.1 nM. Migration of RNA hairpin (hairpin) and heteroduplexes of RNA with several oligomers (htd) is indicated. PO-LNA-CAG-6 is unable to form full complexes with (CUG)₁₀₀; however, PO-LNA-CAG-8 forms such complexes at much lower concentration than both PO-LNA-CAG-10 and PO-LNA-CAG-12. The K_d value for each interaction, based on at least three independent assays, is specified bellow each electrophoregram. 2′OMe-CAG-21 oligomer was used as a control. It shows that different modifications and length of molecules slightly change the strength of their interaction with long CUG repeat RNA. (b) The quantification of EMSA results based on the decline of free (CUG)₁₀₀ in favor of forming CUG₁₀₀/LNA-CAG complexes. (c) Prevention of the (CUG)₁₀₀/MBNL1 interaction in the presence of tested PO-LNA-CAGs. The curves show inhibition of MBNL1 binding by all tested oligomers. (d) RT-PCR results of alternative splicing changes of MBNL2 and MBNL1 transcript in non-DM1 fibroblasts (first lane) and (CUG)1000 fibroblasts treated with transfection reagent only (mock) or one of the four oligomers (PO-LNA-CAG-6, PO-LNA-CAG-8, PO-LNA-CAG-10 and PO-LNA-CAG-12) used at three different concentrations (5, 25, 125 nM). Upper bands represent exon inclusion isoforms (+ex7) while lower bands represent exon exclusion (−ex7). Lower level of alternative exon inclusion after treatment with highest concentrations of PO-LNA-CAG-8 and PO-LNA-CAG-12, and with all tested PO-LNA-CAG-10 concentrations, indicates alternative splicing correction. Statistical significance was determined using two-tailed Student's t-test by comparison of average results from three independent experiments to results obtained for mock samples (*P < 0.05, **P < 0.01 and ***P < 0.001).

Figure 2.

PO-LNA-CAG-10 corrects alternative splicing of MBNL-sensitive exons at low nanomolar concentrations. Results of RT-PCR analyses for alternatively spliced exons of INSR, NFIX, MBNL2, MBNL1, NCOR2, and PHKA1 transcripts in (CUG)1000 cells treated with 0.2–250-nM PO-LNA-CAG-10. The splicing pattern of all tested alternative exons in untreated non-DM1 cells is also shown. The EC₅₀ values were calculated for each transcript. Other details are the same as described in Figure 1d. Statistical significance was determined using two-tailed Student's t-test by comparison of average results from three independent experiments to results obtained for mock samples (*P < 0.05, **P < 0.01 and ***P < 0.001).

Figure 3.

Timeline of alternative splicing correction in PO-LNA-CAG-10-treated DM1-cells. Longitudinal analysis of alternative splicing changes in MBNL1, MBNL2, NFIX and NCOR2 transcripts in (CUG)1000 fibroblasts after 3, 6, 9, 12, 24 h and 2, 3, 5, 9, 15, 20 and 30 days from single transfection with 125-nM PO-LNA-CAG-10. Cells were passaged twice at days 16 and 21 (indicated by black arrows), at which the numbers of cells were four times and eight times higher, respectively, than at the time of transfection (confluence factor is indicated below timeline). The relative splicing effect shows the percentage of alternative exon inclusion at different experimental time points. Horizontal lines indicate the threshold characteristic for non-DM1 cells (value ‘100’) and not treated (CUG)1000 cells (DM1 non-treated; value ‘0’). The most significant splicing correction effect was observed between ∼24 and ∼216 h post transfection.

Figure 4.

The level of CUG^exp RNA is slightly decreased in PO-LNA-CAG-10-treated DM1 cells. (a) Allele-specific semi-quantitative RT-PCR assay measuring the relative level of transcripts from two DMPK alleles in DM1 fibroblasts transfected with 125 nM of each of the three control PO-LNAs (CTR1, CTR2 and CTR3) or PO-LNA-CAG-10. RT-PCR products containing BpmI polymorphism were either digested with BpmI (first five lanes) or undigested (-BpmI; last lane). DM1-specific RT-PCR product is sensitive to BpmI (DM1; lower band) and normal, non-DM1 variant is resistant to BpmI (N; upper band). (b) The relative level of the DMPK transcripts containing 960 CUG repeats in HeLa cells transiently transfected with DT960 expression or treated with transfection reagent only (mock). Cells were treated with either transfection reagents (lane 0) or PO-LNA-CAG-10 at two concentrations (125 and 250 nM). The results were normalized to GAPDH-specific RT-PCR product. Note that only slightly decreased levels of CUG^exp containing RNA were observed in (a) and (b) upon PO-LNA-CAG-10 treatment. (c) The relative level of transcripts containing short CUG repeat tracts (from 7 to 12 repeats) after PO-LNA-CAG-10 treatment was measured by real-time RT-PCR (normalized to GAPDH mRNA). The level of all transcripts in mock-treated samples was fixed as 1 and used for statistical analyses. Neither of the tested transcripts showed significant splicing change in the presence of LNA oligomer. (*P < 0.05, **P < 0.01 and ***P < 0.001).

Figure 5.

The number and size of CUG^exp containing nuclear foci are significantly decreased in PO-LNA-CAG-10-treated DM1 cell. (a) Representative FISH analyses of CUG^exp foci in (CUG)1000 fibroblasts treated with transfection reagent (mock) or PO-LNA-CAG-10 (125 nM). The average number of foci per nucleus (b) and their average volume (μm³) (c) were measured in 400 arbitrary-selected nuclei from cells transfected with 125 nM of either control PO-LNA-CTR3, PO-LNA-CAG-8 or PO-LNA-CAG-10 oligomers.

Figure 6.

PO-LNA-CAG-10 corrects missplicing in DM1 muscle cells and in mouse model of DM1. (a) Alternative splicing analyses after a single (1x) or double (2x) transfection of 125-nM control PO-LNAs (CTR1, CTR2, CTR3) and PO-LNA-CAG-10 in myoblasts cells derived from non-DM1 individual (non-DM1) and DM1 patient. The non-treated cells (N/T) and cells treated with transfection reagent (mock) were used as controls. Other details are the same as described in Figure 1d. Note that for control LNAs-treated cells there are no significant differences in splicing isoform distribution. (b) Tibialis anterior muscles of two HSA^LR mice were injected with either saline (sal) or 5 μg of PO-LNA-CAG-10 (LNA). The splicing pattern of exons regulated by Mbnl1 of three transcripts (Atp2a1, Titin and Nfix) was analyzed 14 days post-treatment. Non-treated muscle (N/T) of an HSA^LR mouse was used as a control. (c) Myotonic discharges were measured for each treated and non-treated muscle and the grade of myotonia was arbitrary classified using scale from 0 to 3. The results are presented as an average from two to four animals. (d) The relative level of CUG^exp containing transcript was quantified by norther blot hybridization. (*P < 0.05, **P < 0.01 and ***P < 0.001).