Main

Duchenne muscular dystrophy (DMD) is an X-linked recessive neuromuscular disease caused by mutations in the DMD gene, which encodes dystrophin—a protein that acts as a mechanical link between the cytoskeleton and the extracellular matrix, safeguarding muscle cells from contraction-induced damage1,2. FDA-approved treatments for DMD include exon skipping therapies for specific deletions9, stop codon read-through compounds such as ataluren10 and Elevidys, a recombinant gene therapy delivering micro-dystrophin11. Utrophin, encoded by UTRN, the paralogue of DMD, has been shown to be upregulated at the sarcolemma of skeletal muscles in patients carrying frameshift or nonsense DMD mutations that result in premature termination codons (PTCs)3,4, as well as in the mdx DMD mouse model12,13,14, which harbours a nonsense mutation in exon 23 of the Dmd gene. By contrast, patients with in-frame deletions in DMD have been reported not to show utrophin upregulation15. Preclinical studies have revealed an inverse correlation between utrophin expression and disease course severity in DMD5,16. Furthermore, mdx/Dmd mutant mice show a milder phenotype than the Utrn/Dmd double mutant mice12. Therefore, utrophin upregulation has been proposed to be a compensatory mechanism to partially counteract the lack of dystrophin, making it a promising therapeutic strategy for treating DMD patients. However, the mechanisms behind UTRN upregulation have remained largely elusive and are thought to be due to the loss of dystrophin protein17. Here, to investigate the mechanisms underlying UTRN upregulation in DMD patients carrying frameshift or nonsense alleles, we developed several genetic tools and show that DMD mutant messenger RNA (mRNA) decay plays a pivotal role in UTRN upregulation through a newly identified cellular response called transcriptional adaptation (TA). Furthermore, we reveal a new application for splice-switching antisense oligonucleotides (ASOs), as well as for ribozymes, to trigger genetic compensation via TA. During TA, mutant mRNA decay can lead to the increased transcription of so-called adapting genes6,7,8,18,19,20,21,22,23,24,25. This process is independent of the loss of protein function, and it can lead to functional compensation in some cases, potentially explaining why frameshift or nonsense mutations resulting in PTCs in critical genes do not cause obvious phenotypes6,7,8,18,19,20,21,22. TA has been identified thus far in zebrafish6,8,18, Caenorhabditis elegans19 and mouse cells in culture8, but as of yet TA has not been reported in human cells.

Enhancing DMD E37 inclusion

To study the role of DMD frameshift or nonsense mutations on UTRN expression, we first decided to investigate how to modulate the splicing of DMD exon 37 (E37). Previous studies have identified frameshift and nonsense mutations in this alternatively spliced exon associated with a milder form of DMD, termed Becker muscular dystrophy26,27. Furthermore, bioinformatics analysis has predicted that DMD E37 has a weak 3′ splice site and a low exonic splicing enhancer density, thereby leading to frequent exon skipping28. To evaluate whether the transcriptional elongation rate has an effect on DMD E37 inclusion, we first assessed the effect of camptothecin (CPT), a DNA topoisomerase I inhibitor, previously shown to indirectly inhibit transcriptional elongation29,30. We treated wild-type (WT) human embryonic kidney 293T (HEK293T) cells with 3 µM CPT for 6 h, a treatment condition that has been shown not to completely shut down transcription31,32, and observed E37 skipping (Fig. 1a and Extended Data Fig. 1a), suggesting that inclusion of this exon follows the kinetic model of cotranscriptional splicing31,33. As inhibiting elongation caused E37 skipping, we reasoned that stimulating RNAPII elongation would cause the opposite effect (Extended Data Fig. 1b). Therefore, we treated WT HEK293T cells with 1 µM of the histone deacetylase inhibitor trichostatin A (TSA) for 24 h (Extended Data Fig. 1c), a treatment condition that creates a more relaxed chromatin structure, thereby enhancing elongation31,33. We did indeed observe that TSA promotes E37 inclusion (Fig. 1a and Extended Data Fig. 1d).

Fig. 1: Promoting the inclusion of an alternatively spliced PTC-containing exon triggers DMD mutant mRNA decay and UTRN upregulation.
figure 1

a, Effects on endogenous DMD E37 alternative splicing in WT HEK293T cells after treatment with CPT or TSA; alternative splicing assessed by PCR with reverse transcription (RT–PCR) followed by agarose gel electrophoresis and SYBR Safe staining. Bars display means ± s.d. of the percentage of the intensity of the band containing E37 over the sum of intensity of the bands containing E37 and the skipped isoform (E37−) (n = 3 biologically independent samples). b, Schematic illustration of the DMD PTC allele generated with CRISPR–Cas9 in HEK293T cells; red indicates the PTC; gRNA, guide RNA. c,d, qPCR analysis of DMD mRNA levels in WT and DMDPTC/+ cells after treatment with DMSO (c) or TSA (d) (n = 4 biologically independent samples). e,f, qPCR analysis of UTRN mRNA levels in WT and DMDPTC/+ cells after treatment with DMSO (e) or TSA (f) (n = 5 biologically independent samples). g,h, Western blot analysis and protein quantification showing dystrophin and utrophin levels in WT and DMDPTC/+ cells treated with DMSO (g) or TSA (h) (n = 3 biologically independent samples). Data are normalized to WT and are mean ± s.d.; a two-tailed Student’s t-test was used to calculate P values. Threshold cycle (Ct) values are included in Supplementary Table 5.

Source Data

Full size image

Introduction of a PTC in DMD E37

The ability to induce the inclusion of DMD E37 at will makes it a promising exon in which to introduce a frameshift or nonsense mutation that would lead to the decay of the DMD transcript. Therefore, we targeted E37 in HEK293T cells using a guide RNA with a high fidelity Cas9 and generated a heterozygous DMD line, DMDPTC/+, that carries a 20-nt deletion in E37, which results in a PTC positioned 118 nt upstream of its 3′ end (Fig. 1b). We measured the effect of CPT and TSA treatment on DMD E37 splicing in these cells, and found that the impact of elongation rate on DMD E37 inclusion was similar to that previously observed in WT cells (Extended Data Fig. 2a–c). We also observed that DMD E37 skipping happened more frequently in DMDPTC/+ cells than in WT cells in control conditions (Extended Data Fig. 2d,e), an observation consistent with previous reports showing that frameshift indels can lead to the skipping of affected in-frame exons, with the resulting transcripts escaping from nonsense-mediated mRNA decay (NMD)34,35. We then used the splicing-factor binding-site prediction tool, ESEfinder36, and found that the deleted 20-nt sequence in E37 was indeed predicted to harbour many splicing enhancers (Extended Data Fig. 2f), potentially explaining the more frequent skipping of E37 in DMDPTC/+ cells compared with WT.

DMD E37PTC inclusion triggers UTRN upregulation

To determine whether DMD mutant mRNA decay was taking place upon inclusion of the PTC-containing exon, we measured the mRNA levels of a region of the DMD transcript where the inclusion of exons was not affected by TSA treatment (that is, DMD E39-40) (Extended Data Fig. 3a). In control conditions, DMD expression levels were similar between WT and DMDPTC/+ cells (Fig. 1c), whereas they were significantly reduced in DMDPTC/+ cells compared with WT when we induced the inclusion of the PTC-containing exon (Fig. 1d), which is consistent with PTC bearing transcripts being subjected to NMD37,38. Given that the mdx mutant mouse, which harbours a nonsense mutation in Dmd, displays increased expression of utrophin mRNA and protein12,13,14, we then wanted to test whether utrophin mRNA and protein levels would be similarly upregulated when inducing the inclusion of the PTC-containing E37 in DMDPTC/+ cells. In contrast to control conditions, in which there are no changes in UTRN mRNA levels between WT and DMDPTC/+ cells (Fig. 1e), we observed an increase in UTRN expression levels in DMDPTC/+ cells compared with WT when treated with 1 µM TSA for 24 h (Fig. 1f). The fold change of this UTRN upregulation was around 1.5–2, which is similar to that reported in mdx mutant mice39. To investigate the cause of this increase in UTRN mRNA levels, we measured its precursor mRNA (pre-mRNA) levels, and found that they were also increased upon inclusion of the PTC-containing exon (Extended Data Fig. 3b,c), indicating that UTRN upregulation is due to increased transcription. We then investigated whether these changes in DMD and UTRN mRNA levels resulted in protein level changes. We observed an upregulation of utrophin protein despite no obvious loss of dystrophin protein (Fig. 1g,h), the latter most probably due to a long protein half-life as previously observed in vivo40. To evaluate the half-life of dystrophin and utrophin proteins, we treated both WT and DMDPTC/+ cells with the protein translation inhibitor cycloheximide for 24 and 48 h. We found that whereas utrophin levels were severely reduced at 24 h and almost absent at 48 h, dystrophin levels were reduced but still present at 48 h (Extended Data Fig. 3d). Altogether, these findings suggest that UTRN upregulation is not due to the loss of dystrophin protein but to RNA feedback loops involving DMD mutant mRNA decay. They also show that increased UTRN mRNA levels lead to increased utrophin protein levels.

DMD mRNA decay precedes UTRN upregulation

The effect of TSA treatment on DMD E37 inclusion is dose and time dependent in both WT and DMDPTC/+ cells (Extended Data Fig. 4a–d). Increasing the concentration of TSA led to a more pronounced decrease in DMD mRNA levels and a more pronounced increase in UTRN mRNA levels (Extended Data Fig. 4e,f). In terms of time dependency, after 8 h of treatment with TSA, a significant decrease in DMD mRNA levels was already present in DMDPTC/+ cells compared with WT, whereas UTRN mRNA levels were only slightly altered at this time; however, UTRN mRNA levels were significantly upregulated (approximately 1.5-fold) after 16 h of treatment (Extended Data Fig. 4g,h). Together, these findings suggest that changes in UTRN mRNA levels are caused by DMD mutant mRNA decay. We could also reverse the effects of TSA by washing it away and letting the cells recover for 48 h. Recovery from TSA led to a reduction in DMD E37 inclusion levels (Extended Data Fig. 5a–d) as well as a normalization in DMD and UTRN mRNA levels (Extended Data Fig. 5e,f).

Blocking NMD normalizes UTRN upregulation

To investigate the role of the NMD surveillance machinery in DMDPTC/+ cells treated with TSA, we knocked down UPF1 and SMG6—two key NMD proteins37,38 (Extended Data Fig. 5g,h). Blocking NMD in DMDPTC/+ cells led to an increase in DMD mutant mRNA levels (Fig. 2a) as well as a loss of UTRN upregulation at the mRNA (Fig. 2b) and pre-mRNA (Fig. 2c) levels. Incidentally, although blocking NMD in WT cells had no effect on UTRN mRNA or pre-mRNA levels compared with a scrambled control (Fig. 2b,c), it did lead to a significant reduction in DMD mRNA levels (Fig. 2a). Together, the data from the DMDPTC/+ cells indicate that DMD mutant mRNA decay is required to trigger UTRN upregulation, which led us to explore whether TA, which is triggered by mutant mRNA decay and not protein loss6,8, was taking place.

Fig. 2: Blocking NMD results in the reversal of UTRN upregulation, whereas overexpressing DMD does not.
figure 2

a–c, Blocking NMD results in the normalization of DMD expression as well as a 62% loss of UTRN upregulation at the mRNA level and a 70% loss at the pre-mRNA level. qPCR analysis of DMD mRNA (a), UTRN mRNA (b) and UTRN pre-mRNA (c) levels after siRNA-mediated knockdown of UPF1 and SMG6 in WT and DMDPTC/+ cells treated with TSA (n = 5 biologically independent samples); siCTRL, scrambled control. d–f, Overexpression (OE) of dystrophin protein does not alter the upregulation of UTRN caused by the inclusion of the DMD E37PTC/+. d,e, Western blot analysis (d) and protein quantification (e) showing the OE of dystrophin protein following transfection of a DMD overexpression plasmid (n = 3 biologically independent samples). f,g, qPCR analysis of UTRN mRNA (f) and UTRN pre-mRNA (g) levels in WT and DMDPTC/+ HEK293T cells transfected with an empty vector or the DMD overexpression plasmid, followed by addition of 1 µM TSA for 24 h (n = 4 biologically independent samples). Data are normalized to WT and are mean ± s.d.; a two-tailed Student’s t-test was used to calculate P values. Ct values are included in Supplementary Table 5.

Source Data

Full size image

DMD OE does not reduce UTRN upregulation

To test the hypothesis that UTRN upregulation in DMDPTC/+ cells is caused by TA, we first overexpressed DMD in DMDPTC/+ cells treated with TSA. If UTRN upregulation in DMDPTC/+ cells is due to a protein feedback effect, dystrophin overexpression should normalize UTRN mRNA levels. However, if UTRN upregulation is due to TA, dystrophin overexpression should not affect it8. In line with the hypothesis that UTRN upregulation in DMDPTC/+ cells is caused by TA, dystrophin protein overexpression in these cells did not affect UTRN mRNA or pre-mRNA upregulation (Fig. 2d–g). Thus, DMD mutant mRNA decay, and not the loss of dystrophin protein, seems to play a key role in UTRN upregulation in DMDPTC/+ cells. We then sought to identify whether other genes were also affected by DMD mutant mRNA decay, potentially through TA. Therefore, we reanalysed a published RNA sequencing (RNA-seq) dataset from specific induced pluripotent stem (iPS) cells from DMD patients differentiated into myotubes41. These iPS cells were derived from two different patients, one with an E45 deletion (DMD1) and the other with an E51 deletion (DMD2), and the mutations were corrected through knock-ins. We focused our analysis on the DMD2 samples for several reasons including the upregulation of UTRN in mutant cells, which was reduced after mutation correction (Extended Data Fig. 6a,b and Methods). Given that these DMD2 cells seem to display TA, we proceeded to search for other potential adapting genes (that is, genes upregulated due to DMD mutant mRNA decay) and validated them by determining the transcriptome of DMDPTC/+ HEK293T cells treated with TSA, which exhibit UTRN upregulation by RNA-seq analysis as well (Extended Data Fig. 6c). We found that some of the commonly upregulated genes are involved, directly or indirectly, in muscle function, such as ANO5 (related to limb girdle muscular dystrophy type 2L), ACTA1 (related to congenital myopathy), SERAC1 (related to muscular dystonia), LIN28B (with functions in muscle development) and PPP1R14C (involved in muscle contraction) (Extended Data Fig. 6d), suggesting that more genes, beyond UTRN, may compensate for the loss of dystrophin via their upregulation. We found in another published RNA-seq dataset14 that Acta1 also showed a trend towards upregulation, although not significant, in mdx mutant skeletal muscle compared with WT.

DMD PTC minigenes trigger UTRN upregulation

To further test the hypothesis that TA can trigger UTRN upregulation and investigate whether one could induce UTRN upregulation in WT cells, we generated three different DMD minigenes, one consisting of E9 to E11 with intron 10, the second of E29 to E34 with intron 31, and the third of E34 to E36 with intron 35, hereafter referred to as the DMDWT minigenes (Fig. 3a). We then introduced nonsense mutations in E10 (E338X), E31 (E1421X) and E35 (Q1624X), respectively, to obtain the DMDPTC minigenes. We transfected the DMDWT and DMDPTC minigenes, or an empty vector—consisting of the plasmid backbone in which the minigenes were generated—into WT HEK293T, HAP1 and HeLa cells as well as into myotubes. For most cell lines, we observed lower exogenous DMD mRNA levels following transfection with the DMDPTC minigenes compared with transfection with the DMDWT minigenes (Extended Data Fig. 7a–c), suggesting that degradation of the DMD transcripts encoded by the PTC minigenes is taking place. For the DMDE1421X and DMDQ1624X constructs, we observed increased UTRN mRNA levels upon transfection of the DMDPTC minigenes, but not of the DMDWT minigenes, compared with transfections with the empty vector (Fig. 3c,d), in all cell lines except for HeLa cells. For the DMDE338X minigene, increased UTRN mRNA levels were observed only in HAP1 and HeLa cells (Fig. 3b). Altogether, these findings further indicate that DMD mutant mRNA decay can trigger UTRN upregulation; they also suggest that the UTRN upregulation response varies among different cell types and depending on the location of the PTC.

Fig. 3: UTRN upregulation induced by DMDPTC minigenes and by a DMDWT minigene containing a self-cleaving ribozyme.
figure 3

a, Schematic illustration of the DMDPTC minigenes. b–d, OE of the DMDPTC minigenes in WT cells results in UTRN upregulation. qPCR analysis of UTRN mRNA levels in WT HEK293T cells, myotubes, HAP1 cells and HeLa cells transfected with an empty vector, the DMDWT minigene or the DMDE338X (b), DMDE1421X (c) or DMDQ1624X (d) minigene (n = 3 biologically independent samples). e, Schematic illustration of the DMDWT minigene containing a self-cleaving ribozyme (T3H38-HHR) inserted in intron 31, flanked by two insulator sequences. Both an inactive and an active version of the ribozyme were used. f,g, OE of the DMDWT minigene containing T3H38-HHR in WT cells results in UTRN upregulation. qPCR analysis of DMD (exogenous) (f) and UTRN (g) mRNA levels in WT HEK293T cells, myotubes, HAP1 cells and HeLa cells transfected with the DMDWT minigene with an inactive ribozyme (T3H38-iHHR) or with an active ribozyme (T3H38-aHHR) (n = 4 biologically independent samples). Data are normalized to WT transfected with the empty vector (b–d) or with DMDT3H38-iHHR (f,g) and are mean ± s.d.; a two-tailed Student’s t-test was used to calculate P values; P values shown only when significant. Ct values are included in Supplementary Table 5.

Source Data

Full size image

A DMD HHR minigene triggers UTRN upregulation

We then investigated another approach to degrade DMD RNA without affecting dystrophin protein levels. To this end, we inserted a modified variant of the Schistosoma mansoni hammerhead ribozyme (HHR), hereafter referred to as T3H38-HHR (ref. 42), into the intron of a DMDWT minigene that consists of E29 to E34 with intron 31 (Fig. 3e). HHRs are small self-cleaving ribozymes that fold into defined structures and catalyse their own cleavage43. A single nucleotide substitution alters T3H38-HHR from its catalytically active form (denoted T3H38-aHHR) to an inactive form (denoted T3H38-iHHR)42, which we used as a control. WT HEK293T, HAP1 and HeLa cells as well as myotubes transfected with the DMDT3H38-aHHR construct displayed a decrease in exogenous DMD mRNA levels compared with cells transfected with the DMDT3H38-iHHR construct (Fig. 3f). Furthermore, UTRN mRNA levels were increased in most cell types transfected with the DMDT3H38-aHHR construct compared with cells transfected with the DMDT3H38-iHHR construct (Fig. 3g). In HeLa cells, we observed no upregulation of UTRN mRNA levels upon transfection with the DMDT3H38-aHHR construct compared with the DMDT3H38-iHHR construct (Fig. 3g), in line with the results obtained using the DMDE1421X minigene, the PTC counterpart (Fig. 3c). These findings with the PTC and ribozyme minigenes, which cause an increase in UTRN mRNA levels in the presence of DMD minigene mRNA/pre-mRNA decay despite no loss of endogenous dystrophin protein, further indicate that DMD mRNA decay is the main trigger for UTRN upregulation in DMDPTC/+ cells, placing TA as the mechanism behind UTRN upregulation in some DMD patients.

DMD PTC myotubes display UTRN upregulation

To place these findings in a disease context, we first measured DMD and UTRN mRNA levels in myotubes derived from four DMD patients, each carrying a different DMD lesion predicted to lead to PTCs: a nonsense mutation in E26 (DMDE118X), a deletion spanning from E51 to E54 (DMDΔE51-54), an E52 deletion (DMDΔE52) and a nonsense mutation in E59 (DMDR2905X) (Extended Data Fig. 8a). We observed reduced DMD mRNA levels and increased UTRN mRNA levels in all DMD derived myotubes compared with myotubes derived from a healthy control (Fig. 4a,b). Furthermore, as previously observed in HEK293T cells (Extended Data Fig. 3b,c), we also found increased UTRN pre-mRNA levels in myotubes derived from the DMD patients compared with myotubes derived from a healthy control participant (Extended Data Fig. 8b), indicating that UTRN upregulation in these cases is also due to increased transcription.

Fig. 4: Myotubes from DMDPTC patients display UTRN upregulation, and restoring the DMD reading frame in DMDΔE52 myotubes reduces this upregulation.
figure 4

a,b, qPCR analysis of DMD (a) and UTRN (b) mRNA levels in myotubes derived from DMD patients compared with myotubes derived from healthy control participants (n = 3 biologically independent samples). c, Schematic of the binding site of the i51-5ss ASO and RT–PCR showing DMD E51 skipping upon transfection in DMDΔE52 myotubes. Bars display means ± s.d. of the percentage of the intensity of the E51-containing band over the sum of intensity of the bands containing E51 and the skipped isoform (E51−) (n = 3 biologically independent samples). The asterisk * denotes a cryptic splice site. d, Restoring the reading frame in DMDΔE52 myotubes results in a 175% increase in DMD mRNA levels. qPCR analysis of DMD mRNA levels in WT and DMDΔE52 myotubes after transfection with a scrambled (Sc) control ASO or i51-5ss (n = 4 biologically independent samples). e, Western blot analysis and protein quantification showing dystrophin levels in WT and DMDΔE5 myotubes after transfection with the Sc ASO or i51-5ss (n = 2 biologically independent samples). f, Restoring the reading frame in DMDΔE52 myotubes results in a 33.85% decrease in UTRN mRNA levels. qPCR analysis of UTRN mRNA levels in WT and DMDΔE52 myotubes after transfection with the Sc ASO or i51-5ss (n = 6 biologically independent samples). g, Western blot analysis and protein quantification showing utrophin levels in WT and DMDΔE52 myotubes after transfection with the Sc ASO or i51-5ss (n = 2 biologically independent samples). h, qPCR analysis of UTRN mRNA levels in WT and DMDΔE52 myotubes transfected with the DMDT3H38-iHHR or DMDT3H38-aHHR minigene (n = 3 biologically independent samples). Data are normalized to WT or to the Sc ASO samples, and are mean ± s.d.; two-tailed Student’s t-test used to calculate P values. Ct values are in Supplementary Table 5.

Source Data

Full size image

UTRN expression in DMD ΔE52 myotubes

We then sought to restore the DMD reading frame in DMDΔE52 myotubes through exon skipping, a current treatment for certain DMD patients with specific mutations in DMD. For example, eteplirsen is an FDA-approved ASO designed to treat DMD patients with a disrupted reading frame that can be restored by the skipping of exon 51 (refs. 44,45), which allows the production of an internally deleted but partially functional dystrophin protein. We therefore designed an ASO similar to eteplirsen that promotes exon 51 (i51-5ss) skipping similar to eteplirsen and introduced it in DMDΔE52 myotubes (Fig. 4c). Notably, we found that this ASO not only led to increased DMD mRNA levels in these DMD mutant myotubes (Fig. 4d,e), but that it also reduced utrophin upregulation (Fig. 4f,g), thereby potentially counteracting some of the beneficial effects of restoring the reading frame. We then decided to try to enhance UTRN upregulation rather than block it in these DMDΔE52 myotubes. Therefore, we took advantage of the DMDWT minigene containing the self-cleaving ribozyme (DMDT3H38-HHR) to determine whether UTRN mRNA levels could be further upregulated in DMDΔE52 myotubes. We indeed observed higher UTRN mRNA levels in DMDΔE52 myotubes transfected with the active ribozyme compared with those transfected with the inactive ribozyme (Fig. 4h). We also observed increased UTRN mRNA levels in other myotubes derived from DMD patients transfected with the active ribozyme compared with those transfected with the inactive ribozyme (Extended Data Fig. 8c,d).

Splice-switching ASOs can trigger UTRN upregulation

In a converse approach to restoring the reading frame, we transfected WT myotubes with splice-switching ASOs designed to target the 5′ splice sites of two out-of-frame DMD exons (E6 (i6-5ss) and E52 (i52-5ss)), and thereby induce their skipping (Fig. 5a). We reasoned that skipping E6 or E52 should lead to the introduction of PTCs in the DMD transcript by disrupting the reading frame, and thereby trigger UTRN upregulation (Fig. 5a). We tested these ASOs in WT myotubes and observed an almost 100% efficient skipping of E6 and E52 (Fig. 5b) when compared with a scrambled control ASO (Sc). Skipping E6 or E52 triggered DMD mutant mRNA decay as well as UTRN upregulation (Fig. 5c,d). Next, we reasoned that if UTRN upregulation by means of splice-switching ASOs was mediated by TA and not protein loss, then partial exon skipping (that is, 25–50%) should also lead to UTRN upregulation without causing a severe reduction in DMD mRNA or protein levels. Partial exon skipping indeed led to UTRN upregulation (Fig. 5g,h) without causing a significant reduction in DMD mRNA (Fig. 5f) or protein (Fig. 5h) levels, probably due to the higher prevalence of the full-length isoform compared with the isoforms in which E6 or E52 is skipped (Fig. 5e). Furthermore, we also used the i51-5ss ASO, which we had used in Fig. 4c–g to restore the reading frame in DMDΔE52 myotubes, to disrupt the reading frame in WT myotubes, as E51 is also an out-of-frame exon (Extended Data Fig. 9a). As expected, we observed UTRN upregulation (Extended Data Fig. 9b). We also observed an increase in DMD mRNA levels in WT myotubes transfected with the i51-5ss ASO compared with those transfected with the Sc ASO (Extended Data Fig. 9c), which could be explained by self-TA8 (that is, DMD mutant mRNA decay leads to the transcriptional upregulation of DMD itself). To further test whether self-TA could be occurring with this ASO, we measured DMD pre-mRNA levels and indeed observed their increase in WT myotubes transfected with the i51-5ss ASO compared with the Sc ASO (Extended Data Fig. 9d). We then decided to use this approach (that is, splice-switching ASOs that disrupt the reading frame) in the C2C12 mouse skeletal muscle cell line to determine whether Utrn upregulation could be observed upon inducing Dmd mRNA decay in mouse as in human. We used a splice-switching ASO targeting the out-of-frame Dmd E22, leading to PTCs in E23 (which incidentally is also the PTC-containing exon in the mdx mutant mouse12), as well as an ASO targeting the out-of-frame E52 (Extended Data Fig. 9e,g). We observed Utrn upregulation, despite no significant changes in Dmd mRNA levels, with both of these splice-switching ASOs compared with scrambled control (Extended Data Fig. 9f,g), indicating that in mouse as in human, TA can trigger Utrn/UTRN upregulation in Dmd/DMDPTC/+ cells.

Fig. 5: Introducing PTCs in DMD using splice-switching ASOs triggers UTRN upregulation.
figure 5

a, Schematic illustration of splice-switching ASOs disrupting the DMD reading frame by inducing the skipping of the out-of-frame exons E52 and E6, thereby leading to the introduction of PTCs and TA. b, RT–PCR showing DMD E52 or E6 full skipping upon i52-5ss or i6-5ss transfection in WT myotubes. Bars display means ± s.d. of the percentage of the intensity of the bands containing E52 or E6 over the sum of intensity of the bands containing E52 or E6 and the skipped isoform (E52− or E6−) (n = 3 biologically independent samples). c,d, Skipping the out-of-frame DMD exons E52 or E6 triggers UTRN upregulation. qPCR analysis of DMD (c) and UTRN (d) mRNA levels in WT myotubes after transfection with the Sc ASO, i52-5ss or i6-5ss (n = 3 biologically independent samples). e–g, Graphs show the results from the same experiments as in b (e), c (f) and d (g) but with the ASO transfections inducing partial exon skipping (n = 3 biologically independent samples). Data are normalized to the Sc ASO samples and are mean ± s.d.; a two-tailed Student’s t-test was used to calculate P values. h, Western blot analysis and protein quantification showing dystrophin and utrophin levels in WT myotubes treated with the Sc ASO, i52-5ss or i6-5ss at concentrations inducing partial exon skipping (n = 2 biologically independent samples). Ct values are included in Supplementary Table 5.

Source Data

Full size image

Discussion

Since the identification and analysis of the mdx mouse model to study DMD, when utrophin upregulation was observed and shown to compensate for the loss of dystrophin12, upregulation of utrophin has been one of the main strategies to treat this disease46,47. Nevertheless, the mechanisms underlying utrophin upregulation are poorly understood, and thought to be caused by the loss of dystrophin protein. By using four different approaches, we provide evidence that DMD mRNA/pre-mRNA decay is sufficient to upregulate UTRN, and also provide evidence that NMD is necessary for UTRN upregulation in DMDPTC/+ cells (Extended Data Fig. 10). TA as a mode of genetic compensation was first identified in zebrafish (Danio rerio) when investigating differences between mutation-induced phenotypes and antisense (morpholino)-induced phenotypes6. Further investigations revealed the presence of TA in C. elegans19, and identified mutant mRNA decay as playing a key role in triggering TA in zebrafish8 and C. elegans19 as well as in mouse cells in culture8. Nevertheless, TA has not been reported in humans thus far. Here we present an example of TA in humans and its potential role in a hereditary disease. Our data show that myotubes from a DMD patient with an E52 deletion display UTRN upregulation. Skipping of E51 in these myotubes using an ASO similar to eteplirsen reduced UTRN upregulation. For these DMD patients, using DMD and/or UTRN minigenes with self-cleaving ribozymes might lead to a level of UTRN upregulation that could synergize with the benefits of the eteplirsen treatment. Of course, most, if not all, DMD patients might benefit from these self-cleaving ribozyme-containing minigenes. Missense mutations are reportedly more prevalent than frameshift or nonsense mutations, which often lead to mRNA decay, in many human genetic diseases48,49,50. In some of these diseases (for example, sickle cell disease, Marfan syndrome, and hypertrophic cardiomyopathy caused by mutations in MYH7), it has been reported that missense mutations have more detrimental effects than nonsense mutations48,49,50. Functional compensation through TA could explain why nonsense mutations are less frequently reported than missense mutations, as they may cause a milder phenotype. ASOs have been approved by the FDA for the treatment of multiple diseases; here, we show a new application for splice-switching ASOs to trigger mutant mRNA decay, and thereby induce functional compensation through TA. Furthermore, minigenes with self-cleaving ribozymes provide another approach to induce TA. This method could be used in patients with missense mutations as well as in those with frameshift or nonsense mutations as increased RNA degradation can further upregulate compensating genes, in line with our previous data8. ASOs could similarly be effective in some patients with frameshift or nonsense mutations. We took advantage of DMD E37 alternative splicing to create an inducible DMD mutant mRNA decay model. This inducible splicing-dependent system constitutes a new tool to induce endogenous mRNA decay. In the last few years, it has been shown that alternative splicing is not an exception but rather more of a rule in multicellular organisms, occurring in roughly 95% of human genes51, and genome-wide analyses have revealed that around 20% of a cell’s alternative splicing events are elongation-sensitive51. Therefore, this inducible splicing-dependent system holds promise to trigger mRNA decay across a wide range of genes and observing changes over time. Furthermore, in the 2019 Ensembl database of the human genome, about 15,000 alternative splicing variants are annotated as NMD-sensitive isoforms51. Inducing the inclusion of elongation-sensitive exons bearing PTCs (poison exons), or inducing PTCs by splice-switching ASOs, also emerge as promising approaches to trigger mRNA decay, without the need to modify the genome. Another potentially promising approach to trigger TA is to use the recently discovered strategy named the proximity-induced nucleic acid degrader in which a small molecule degrades RNA in a targeted manner by being in close proximity to its target52. It will also be interesting to test whether RNA interference and/or yet other forms of RNA degradation, can trigger TA or a TA-like response. Altogether, these findings highlight the importance of TA as a mechanism underlying genetic robustness in the human population and its relevance to hereditary diseases, helping in the design of new therapeutic approaches that take advantage of, or do not interfere with, TA-triggered functional compensation.

Methods

Dystrophin expression plasmid

The plasmid p37-2iDMD-LR containing the coding sequence for the WT DMD gene was a gift from M. Calos (Addgene no. 88892)53. DMD exons 29 to 34 from p37-2iDMD-LR as well as intron 31 from HEK293T WT genomic DNA were cloned at the 3′ end of the Kozak sequence of the pSBbi-GP plasmid using a Gibson assembly cloning kit (New England Biolabs), resulting in the DMDWT minigene. A PTC at amino acid position 1421 (E1421X) was introduced in the DMDWT minigene through site directed mutagenesis to generate the DMDPTC minigene. DMD exons 9 to 11 including intron 10, or exons 34 to 36 including intron 35, from HEK293T WT genomic DNA or complementary DNA (cDNA) were cloned at the 3′ end of the Kozak sequence of the pSBbi-GP plasmid using a HiFi cloning kit (New England Biolabs). The PTCs at amino acid position 338 (E338X) or 1624 (Q1624X) were introduced in the DMDWT minigenes through site directed mutagenesis to generate the DMDPTC minigenes. The pSBbi-GP plasmid was a gift from E. Kowarz (Addgene plasmid no. 60511)54. The active and inactive versions of the T3H38-HHR ribozyme were inserted in the DMDWT minigene consisting of exons 29 to 34 and intron 31 by In Vivo Assembly55.

Cell culture and pharmacological treatments

HEK293T cells (DSMZ) and HeLa cells (Enzo Life Sciences) were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 4.5 g l−1 of glucose (Gibco), 10% fetal bovine serum (Sigma) and 1% penicillin streptomycin (Gibco) at 37 °C. HAP1 cells (Horizon Discovery) were grown in Iscove’s modified Dulbecco’s medium containing 10% fetal bovine serum and 1% penicillin streptomycin at 37 °C. Cells were plated at a density of 50,000 cells or 400,000 cells per well in six-well plates 24 h before small-interfering RNA (siRNA) or plasmid transfection, respectively. siRNA (20 nM final concentration), plasmid (1 µg) transfections were performed 24 h after cells were plated using Lipofectamine 3000 (Thermo Fisher Scientific) according to the manufacturer’s instructions. After 24 h, cells were treated with 1 µM TSA (Sigma, T8552), 3 µM CPT (Sigma, C9911) or vehicle for the indicated time and collected for downstream procedures. It should be noted that Trischostatin A and CPT have been reported to have broad effects56,57 and that CPT decreases overall transcription57 (Supplementary Fig. 2). siRNAs used are listed in Supplementary Table 1. To block translation, WT and DMDPTC/+ cells were treated with 100 µg of cycloheximide (Sigma) or DMSO for the indicated times. For minigene transfections, cells were selected using puromycin (4 µg ml−1 for HEK293T cells and 1 µg ml−1 for HAP1 and HeLa cells). All cell lines tested negative for mycoplasma contamination.

Myoblast culture and differentiation

The biopsies used for generating the myoblast lines were supplied by MyoBank, the tissue bank associated with the Institut de Myologie in Paris and affiliated with EuroBioBank. MyoBank is authorized by the French Ministry of Higher Education, Research and Innovation to distribute human samples for research purposes (authorization code AC-2019-3502). Human myoblasts were cultured in Skeletal Muscle Cell Growth Medium (Promocell), adjusted to contain 20% fetal bovine serum and 1% penicillin streptomycin, at 37 °C. They were differentiated by replacing the growth medium with Skeletal Muscle Cell Diffentiation Medium (Promocell), supplemented with 1% penicillin streptomycin, and incubation at 37 °C for 4 to 6 days (ref. 58). The characterization of the DMD human myoblast lines (including information about the tissue from which they are derived) is included in Supplementary Table 2.

C2C12 cell culture and differentiation

C2C12 cells (American Type Culture Collection) were cultured in DMEM containing 4.5 g l−1 of glucose, 10% fetal bovine serum and 1% penicillin streptomycin at 37 °C. C2C12 cells were differentiated by replacing the growth medium with DMEM containing 4.5 g l−1 of glucose and 2% horse serum and incubation at 37 °C for 3 days.

Generation of DMD PTC/+ HEK293T cells

HEK293T cells were transfected with the single-guide RNA (5′-GAATGACATACGCCCAAAGG-3′) cloned into the pSpCas9(BB)-2A-Puro plasmid, a gift from F. Zhang (Addgene no. 62988)59. HEK293T cells were seeded at a density of 600,000 cells per well in a six-well plate 24 h before plasmid transfection. Plasmid (2.5 µg) transfections were performed using Lipofectamine 3000 according to the manufacturer’s instructions. Cells were incubated for 24 h after transfection, then the transfected cells were selected with medium containing puromycin (4 μg ml−1) (Gibco) for 1 week. Puromycin resistant cells were diluted in 10-cm dishes and incubated in fresh medium until single clones formed colonies. Colonies were transferred to a 96-well plate and genomic DNA was isolated from each colony for genotyping with the primer sequences listed in Supplementary Table 3.

ASO and minigene transfections in skeletal muscle cells

All ASOs used in this study are 18mers uniformly modified with 2′-O-MOE ribose, PS linkages and 5′-methylcytosine (sequences in Supplementary Table 4). They were obtained from Integrated DNA Technologies. Human myoblasts were seeded at a density of 100.000 cells per well in a 24-well plate 24 h before ASO or minigene transfection. The myoblasts were transfected with 2.5 or 5 μM ASOs or 1.2 µg of minigene using Lipofectamine 3000 according to the manufacturer’s instructions. For the ASO transfections, the growth medium was replaced with fresh differentiation medium without ASOs 4 h post-transfection and cells were harvested 5 days after transfection. For the minigene transfections, the growth medium was replaced with fresh growth medium without the minigene 4 h post-transfection; fresh differentiation medium was added 24 h later, and cells were harvested 5 days after the differentiation medium was added. In the case of the i51-5ss ASO, where we observed an extra isoform besides the full-length and the E51-skipped isoform, we sequenced the band and found that it was generated by the use of a cryptic splice donor within E51, activated by blocking the splice site at the 5′ end of intron 51. Mouse C2C12 cells were seeded at a density of 60,000 cells per well in a 12-well plate and transfected with 5 μM ASOs using Lipofectamine 3000, according to the manufacturer’s instructions. Then 24 h after transfection, the growth medium was replaced with fresh differentiation medium without ASOs, and cells were harvested 3 days post-transfection. For the C2C12 cells, the differentiation medium was replaced daily until collection.

RNA extraction and RT–(q)PCR

Cells were collected with 1 ml of Trizol (Invitrogen). Total RNA was isolated according to the manufacturer’s instructions. One microgram of total RNA was reverse transcribed using Superscript III (Thermo Fisher Scientific) reverse transcriptase and oligo-dT primers. The resulting cDNA was amplified using Gotaq (Promega) and DMD primers surrounding E6, E22, E37 or E52. After amplification, products were loaded in a 2% agarose gel and stained with SYBR Safe (Invitrogen) for visualization. Primer sequences and product sizes are listed in Supplementary Table 3. Non-consecutive lanes are denoted by a vertical white line of separation in Fig. 1a and Extended Data Figs. 2a, 2d and 9a. For quantitative PCR (qPCR), 1 μg of total RNA was reverse transcribed using a MAXIMA cDNA Synthesis Kit (ThermoFisher Scientific, K1671). qPCR reactions were prepared using DyNAmo ColorFlash SYBR Green PCR mix (Thermo Fisher Scientific, F-416). Primer sequences and products sizes are listed in Supplementary Table 3. A standard program was run on QuantStudio 7 Pro Real-Time PCR System (ThermoFisher Scientific, A43185) and data analysis was performed using Design & Analysis Software v.2.7.0 from ThermoFisher Scientific.

Western blots

Cells were lysed in RIPA buffer (Sigma). Protein samples were separated by 3–8% Tris-Acetate Protein Gel or 4–20% precast polyacrylamide gel and electroblotted onto PVDF membranes (Bio-Rad). The blots were probed with 1:5,000 anti-Dystrophin (Proteintech, 68120-1-Ig), 1:2,000 anti-Utrophin (Proteintech, 29133-1-AP), 1:2,000 anti-H3K9ac (Proteintech, 29133-1-AP) or 1:2,000 anti-tubulin (Sigma, T6557) antibodies.

RNA-seq data acquisition and processing

PolyA-enriched RNA-seq was performed for DMDPTC/+ and WT HEK293T cells treated with 1 µM TSA for 24 h. RNA-seq data from myotubes differentiated from iPS cells derived from DMD patients, carrying deletions of exons 45 (DMD1) or 51 (DMD2), both of which lead to PTCs, were obtained from PRJNA1095368 (ref. 41). Reads were quality-trimmed using Skewer. The processed reads were aligned to the GRCh38/Gencode v46 genome using STAR, and transcript abundance was estimated using HT-Seq, followed by DESeq2 for differential expression analysis in patient myotubes or IsoDE2 for HEK293T cells. When both experiments were analysed together, batch effects were removed using RUVSeq60. DMD1 cells were excluded from the final intersection with our experimental data due to the lack of DMD mRNA reduction, suggesting the absence of NMD when compared with control (that is, WT) iPS cells (Extended Data Fig. 6a).

Statistical analysis

GraphPad Prism v.9 was used for statistical analysis. Two-tailed Student’s t-tests were used to compare two pairs of conditions. Unless otherwise stated in the figure legends, we analysed three biologically independent samples with a significance level of P < 0.05.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.