DMD pseudoexon mutations: splicing efficiency, phenotype, and potential therapy




The degenerative muscle diseases Duchenne (DMD) and Becker muscular dystrophy result from mutations in the DMD gene, which encodes the dystrophin protein. Recent improvements in mutational analysis techniques have resulted in the increasing identification of deep intronic point mutations, which alter splicing such that intronic sequences are included in the messenger RNA as “pseudoexons.” We sought to test the hypothesis that the clinical phenotype correlates with splicing efficiency of these mutations, and to test the feasibility of antisense oligonucleotide (AON)–mediated pseudoexon skipping.


We identified three pseudoexon insertion mutations in dystrophinopathy patients, two of whom had tissue available for further analysis. For these two out-of-frame pseudoexon mutations (one associated with Becker muscular dystrophy and one with DMD), mutation-induced splicing was tested by quantitative reverse transcription polymerase chain reaction; pseudoexon skipping was tested using AONs composed of 2′-O-methyl–modified bases on a phosphorothioate backbone to treat cultured primary myoblasts.


Variable amounts of pseudoexon inclusion correlates with the severity of the dystrophinopathy phenotype in these two patients. AON treatment directed at the pseudoexon results in the expression of full-length dystrophin in a DMD myoblast line.


Both DMD and Becker muscular dystrophy can result from out-of-frame pseudoexons, with the difference in phenotype being due to variable efficiency of the newly generated splicing signal. AON-mediated pseudoexon skipping therapy is a viable approach to these patients and would be predicted to result in increased expression of wild-type dystrophin protein. Ann Neurol 2007

Mutations in the DMD gene cause the dystrophinopathies, a collective term for Duchenne muscular dystrophy (DMD), Becker muscular dystrophy (BMD), and the relatively rare X-linked dilated cardiomyopathy. In approximately 90% of cases, the more severe DMD is associated with mutations that interrupt the messenger RNA (mRNA) open reading frame, either by frameshifting deletions or insertions, or by premature stop codon mutations. In contrast, the less severe BMD is typically associated with mutations resulting in an internally altered but partially functional dystrophin protein, with an intact C-terminal domain.1 This principle forms the basis of a promising line of molecular therapy, now reaching clinical trials. Antisense oligonucleotide (AON)–mediated exon skipping therapy targets motifs at intron-exon junctions or within exonic splice enhancers (ESEs) of the pre-mRNA, resulting in the exclusion of the targeted exon or exons and restoration of an open reading frame (reviewed elsewhere2).

Mutation analysis of DMD was originally problematic due to the large size of the gene (2.4 million bases, with 79 exons and 8 promoters). For many years, mutation analysis was limited to methods that detect deletions or duplications of one or more exons. The most widely used was the multiplex polymerase chain reaction (PCR) method for deletion detection, which sampled a subset of commonly deleted exons.3 Only with the recent implementation of high-throughput screening4 or direct sequencing methodologies5 has rapid diagnosis of point mutations within the gene become feasible on a routine basis. Similarly, improvements in methods of dosage analysis have allowed the rapid identification of deletions and duplications of all exons across the entire gene.6, 7 Combining these modern methods of molecular analysis results in detection of about 93 to 96% of mutations via genomic DNA derived from blood samples.8, 9

The availability of detailed genomic mutation analysis has resulted in the increasingly frequent identification of individuals with DMD but no coding region mutations in the DMD gene, and thus to the increasing recognition of a previously difficult to identify category of DMD mutations: intronic point mutations that create novel splice sites, resulting in the inclusion of intronic sequence as a “pseudoexon” within the mRNA.10–12 The large size of the DMD introns still precludes direct intronic sequencing as a practical assay. Consequently, the presence of pseudoexons must be determined from analysis of mRNA, generally obtained from muscle biopsies.

Here, we report three such intronic point mutations resulting in pseudoexon insertion. One patient has an in-frame pseudoexon insertion containing several in-frame stop codons; his phenotype is consistent with DMD responsive to steroid treatment. The two other patients have mutations causing the inclusion of out-of-frame pseudoexons, but they have different phenotypes: one patient has relatively mild BMD, whereas the other has DMD. Using quantitative reverse transcription polymerase chain reaction (qRT-PCR), we demonstrate the molecular basis of the phenotypic discrepancy between these two patients. Lastly, in primary myoblast cultures from these two out-of-frame patients, we demonstrate AON-induced pseudoexon skipping, resulting in the restoration of a full-length, in-frame DMD transcript. As a therapy, pseudoexon skipping is predicted to be most beneficial to patients with this class of mutations, as the resultant rescued dystrophin should be wild type. In contrast, applying targeted exon skipping to the more common frameshifting rearrangements or nonsense mutations generates a BMD-like protein, which will be of variable functional activity depending on the extent and nature of the primary gene lesion.

Subjects and Methods

Patient Ascertainment

Patients were ascertained from among those enrolled in the United Dystrophinopathy Project, an ongoing natural history and genotype-phenotype database consortium. All three patients were diagnosed with dystrophinopathy utilizing criteria for enrollment in the United Dystrophinopathy Project, having clinical features consistent with a dystrophinopathy and any of the following characteristics: (1) an X-linked family history; (2) altered or absent dystrophin expression with immunohistochemical, immunofluorescent, or immunoblot analysis of muscle biopsy; or (3) clinical testing that demonstrated a mutation in the DMD gene.

Genetic Analysis

Under an institutional review board–approved protocol, and following parental and/or patient consent, genomic blood samples were obtained for DNA extraction and mutation analysis. A sufficient quantity of archived muscle tissue was available from all three patients to perform diagnostic mRNA extraction and RT-PCR–based sequencing. Total RNA was extracted using Trizol (Invitrogen, La Jolla, CA) according to the manufacturer's recommended protocol, and complementary DNA (cDNA) was synthesized using random hexamers and SuperScript III reverse transcriptase (Invitrogen). PCR amplification was conducted using published primer sets13 and using Expand High Fidelity Taq polymerase (Roche, Indianapolis, IN). Amplicons were sequenced using an internal set of sequencing primers and were electrophoresed on an ABI 3700 DNA analyzer (Applied Biosystems, Foster City, CA) prepared with POP-5 capillary gel matrix. The entire DMD cDNA was sequenced in each patient. Sequence files were analyzed using Consed14, 15 or Sequencher (Gene Codes, Ann Arbor, MI) and compared against the wild-type dystrophin cDNA sequence (GenBank NM_004006.1). If the sequencing trace for a given fragment appeared to contain multiple sequences, the PCR product for that amplicon was TA cloned (Invitrogen), and individual clones were sequenced and analyzed, allowing resolution of all mutations.

For each pseudoexon detected, the sequence was run through BLAST to identify the corresponding intronic fragment. Primers complementary to flanking intronic sequence were designed for both PCR amplification and sequencing.

Quantitative Real-time Polymerase Chain Reaction

Real-time quantitative PCR of DMD wild-type and mutant transcripts was performed in a RotorGene capillary thermocycler (Corbett Research, Mortlake, NSW, Australia), and data analyzed using the RotorGene software package. For each case, we designed forward amplification primers that spanned either the wild-type exon junctions (exons 11/12 or exons 45/46) or the mutant splice junctions (exon 11/pseudoexon 11A and pseudoexon 45A/exon 46) (Fig 1). The reverse primer was placed within exonic sequence in the distal exon (either exon 12 or 46).

Figure 1.

Graphic representation of the pseudoexon mutations. (A) Patient 42273 (Becker muscular dystrophy [BMD]). (B) Patient 43012 (Duchenne muscular dystrophy [DMD]). (C) Patient DC0160 (DMD). Dystrophin exons are represented as black boxes; pseudoexons are represented as gray boxes. Arrows indicate positions of the mutations relative to the pseudoexons. The positions of the antisense oligonucleotides (AONs) used for exon-skipping studies are indicated as A11 and D11 (A) and as 45 and 45D (B).

Total muscle RNA was obtained from archived clinical muscle biopsy specimens that had been snap-frozen in isopentane cooled in liquid nitrogen. cDNA synthesis from total muscle RNA was performed using random hexamer primers.

For each patient sample, wild-type DMD, pseudoexon-containing DMD, and a control housekeeping gene (eukaryotic translation elongation factor 1, α-1 [EEF1A1]) were amplified in quadruplicate PCRs. In addition, within the same thermocycler run, the same amplifications were performed using normal control muscle template. Within each patient sample, the amount of DMD message (wild type or pseudoexon) was normalized against EEF1A1; the resulting ratio was compared with the ratio found in normal control muscle tissue, which was normalized to a value of 1.

Splice-Site Strength Scoring Methods

The strength of the pseudoexon splice sites was analyzed using the following methods: (1) a neural network approach, which uses machine-learning algorithms for recognition of mammalian splice-site sequence patterns16, 17 (; (2) first-order Markov model, which takes into account only dependencies between adjacent positions within the splice-site sequence, and the Maximum Entropy Model, which takes into account dependencies between adjacent and nonadjacent positions18 (; and (3) the Shapiro and Senapathy position-weight matrix, which reflects the degree of conservation at each position of the consensus 5′ and 3′ splice motifs.19, 20 For each method, a higher score indicates a greater probability of the corresponding sequence being used as a splice site, but does not represent absolute predictive values. Upper limits are defined only for the Shapiro and Senapathy (upper limit = 100) and neural network (upper limit = 1) predictive methods.

Pseudoexon Skipping

Two of the patients (43012 and 42273) agreed to undergo muscle biopsy to establish myoblast cell lines. Under an institutional review board–approved protocol, and after parental and/or patient consent was obtained, a needle biopsy of the quadriceps muscle was performed. Tissue was macerated, enzymatically digested, and plated in SKGM (Clonetics, San Diego, CA). Primary outgrowth was subcultured at 70% confluence.

For pseudoexon skipping studies, myoblasts were cultured in SKGM2 complete medium for skeletal muscle cells (Cambrex, East Rutherford, NJ) on poly-D-lysine/laminin–coated plates (VWR International, West Chester, PA). When cells were confluent, the media were switched to differentiating media (Dulbecco's minimum essential medium supplemented with glutamine, 2% fetal bovine serum, 1% glucose, 1X penicillin/streptomycin) to induce myogenesis. Myotubes were transfected with AONs 7 to 10 days after switching to differentiation media using ExGen500 (Fermentas, Glen Burnie, MD), according to manufacturer's protocol for 4 hours. For RNA studies, myotubes were harvested using Trizol (Invitrogen) at the indicated time points. RT-PCR was conducted using SuperScriptIII RT-PCR system (Invitrogen) with amplification primers in exons 11 and 12 for Patient 42273 and exons 45 and 46 for Patient 43012 for 35 cycles of amplification. A second round of PCR (35 cycles of amplification) was performed with the nested primers in the same exons using AccuPrime Taq (Invitrogen). Primer sequences are included in the Supplemental Table, and conditions are available on request. For protein studies, cells were further differentiated for 5 days after transfection and lysed in Radioimmuno Precipitation Assay (RIPA) buffer (150mM NaCl, 10mM tris[hydroxymethyl]aminomethane pH7.2, 0.1% sodium dodecyl sulfate, 1% Triton X-100 [Sigma, St Louis, MO], 1% deoxycholate, 5mM EDTA, complete protease inhibitors [Roche, Indianapolis, IN]). Lysates were loaded directly on 3 to 8% tris[hydroxymethyl]aminomethane-acetate polyacrylamide gels (Invitrogen). Expression of full-length dystrophin was analyzed by Western blotting using Mandra I (Sigma) antibody to the C-terminal part of dystrophin protein. A commercial skeletal muscle cell line (Cambrex, East Rutherford, NJ) was used as a wild-type control.

Antisense Oligonucleotide Synthesis

The AONs were prepared on an Expedite 8909 Nucleic Acid Synthesizer using the 1μmol thioate synthesis protocol with 2′-O-methyl cyanoethyl phosphoramidites and support columns supplied by Glen Research (Sterling, VA). The AONs were deprotected and cleaved from the support column using NH4OH and desalted in NAP-10 columns under sterile conditions. All bases carried a 2′-O-methyl ribose modification on a phosphorothioate backbone. The nucleotide sequences of the AONs are shown: 5′-3′ A11-ggg aca gag guu gca gug agc uga gau; D11-cuc acg agg cug agg cag gag aau; 45-uug uca gca auc cau ugc uug aag gc; and 45D-uac cac ugc cuu gcu ucc guc ucc ca.


All three probands had clinical features consistent with a dystrophinopathy. Patient 42273 is currently 23 years old and remains tenuously ambulant. He first noted symptoms of calf hypertrophy and leg weakness at age 9 years, followed by the development of pelvic girdle weakness. He has an X-linked family history of similar symptoms, with ambulation into young adulthood, consistent with a clinical diagnosis of BMD.

Patient DC0160 is currently 15 years old and remains ambulant at home and in the classroom, although he uses a power scooter otherwise. He was seen by a neurologist at age 5 for an unusual gait, without a diagnosis being made. At age 8, he was having difficulty climbing stairs. Serum creatine kinase concentration was reportedly greater than 20,000 IU/L, and a muscle biopsy was performed that was reported to demonstrate absence of dystrophin staining. Based on this biopsy, he was diagnosed with DMD and has been treated with daily prednisone since age 9.

Patient 43012, currently 11 years old and ambulant, presented at age 20 months with delayed ambulation and slow motor and language skills. At age 5, he was diagnosed with DMD based on calf hypertrophy, proximal weakness, increased creatine kinase concentration, and positive X-linked history consistent with DMD. Muscle biopsy showed absent staining for dystrophin, although some revertant fibers were seen. He has a 9-year-old brother who is similarly affected and has had the same intronic point mutation confirmed from a genomic blood sample. However, the brother has never undergone a muscle biopsy.

Muscle biopsies from the DMD Patient 43012 demonstrated absent staining for dystrophin using antibodies directed to the C-terminal, N-terminal, and rod domain of the dystrophin protein (Fig 2A and data not shown). Consistent with this, immunoblot demonstrated absent dystrophin (see Fig 2B). In contrast, immunofluorescent analysis of muscle from the BMD patient (Patient 42273) demonstrated a diffuse reduction of staining at the muscle membrane with all three antibodies (see Fig 2A; also, additional data not shown), indicating the presence of full-length dystrophin protein. This is corroborated by the results of his sample's immunoblot analysis, which shows the presence of low amounts of an apparently full-length protein of 427kDa (see Fig 2B). Patient DC0160, with clinical features suggesting BMD but consistent with steroid-responsive DMD, was reported to have absent staining for dystrophin by immunohistochemistry, according to his clinical report. A tiny remnant fragment was used for generation of mRNA for cDNA sequencing (as discussed later); however, insufficient archived clinical tissue was available for immunofluorescent and immunoblot analysis for this patient, and he refused a research muscle biopsy.

Figure 2.

Analysis of dystrophin expression in patient muscle samples. (A) Immunofluorescence analysis of Patients 42273 and 43012 with Mandra I antibody. (B) Western blot analysis of Patients 42273 and 43012. Molecular weight marker is MagicMark XP (Invitrogen). GAPDH antibody (AbCam, Cambridge, MA) is present as a loading control. WT = wild type.

For each patient, analysis of genomic DNA using both the single condition amplification/internal primer (SCAIP) sequencing5 and either MLPA or MAPH methods6, 7 was negative for mutations affecting any of the 79 exons of the DMD gene. cDNA sequencing from muscle biopsy–derived mRNA resulted in the identification of out-of-frame pseudoexon insertions in all three patients (see Fig 1). The details of each mutation (including the position of each intronic fragment relative to the flanking exons) are listed in Table 1. In each case, sequencing of the appropriate intron showed point mutations creating novel splice sites.

Table 1. Pseudoexon Insertion Mutations in Three Dystrophinopathy Patients
inline image

To evaluate the relative efficiency of the splice sites created by the intronic point mutations, we undertook qRT-PCR analysis of the relative levels of each pseudoexon-containing transcript compared with wild-type DMD mRNA. Biopsies from one BMD (42273) and one DMD (43012) patient yielded sufficient RNA for this analysis. qRT-PCR values for the respective pseudoexon inclusion (11A or 45A) and wild-type dystrophin transcripts were normalized against EEF1A1 as a housekeeping standard. The wild-type DMD:EEF1A1 ratio was normalized to a value of 1.0 in a wild-type sample run at the same time. The results of qRT-PCR are shown in Figure 3. Both the intron 11 BMD-associated mutation and the intron 45 DMD-associated mutation result in pseudoexon-containing DMD message, present at levels of approximately 40% of the amount of wild-type DMD transcript found in normal control muscle. However, in the muscle from the BMD patient, a significant level of wild-type transcript is also present (approximately 13% of that found in control muscle), whereas in DMD muscle, no wild-type transcript is seen.

Figure 3.

Quantitative reverse transcription polymerase chain reaction (RT-PCR) analysis of pseudoexon and wild-type transcripts. BMD = Becker muscular dystrophy; DMD = Duchenne muscular dystrophy; mRNA = messenger RNA.

To investigate the feasibility of pseudoexon skipping induced by AONs directed to the intronic sequence, we established primary muscle cultures from needle biopsies of the BMD (42273) and DMD (43012) patients. Two AONs were designed for each of the pseudoexons: one was directed to a cluster of ESEs within the pseudoexons, as predicted using ESEFinder2 (,21 and the other was directed to the donor or 5′ splice-site junction created by the point mutation itself (see Fig 1).

First, patient myotubes were treated with increasing amounts (from 10–300nM) of AONs directed at what was considered the major cluster of ESEs, the cryptic 5′ splice site, or both in combination. In the BMD myoblasts containing the intron 11 pseudoexon (11A), when used individually, the AON directed at the ESE (A11) partially corrected splicing only at the greatest concentration (300nM), and the AON directed to the pseudoexon-intron junction partially corrected splicing at a lower concentration (100nM) (Fig 4A). In contrast, in the DMD myoblasts containing the intron 45 pseudoexon (45A), AON treatment increased levels of normal splicing even at the lowest concentration (10nM), with a dose-dependent response at 30 and 100nM (although there was no obvious further increase of normal splicing at 300nM) (see Fig 4B). For either pseudoexon mutation, a combination of both AONs (each one at the concentration stated in Fig 4) worked better than a single AON separately at the same concentration. However, a greater concentration of each single AON could apparently induce the same level of correct splicing as could a lower concentration of the primers used in combination.

Figure 4.

Reverse transcription polymerase chain reaction (RT-PCR) analysis of titration of antisense oligonucleotide (AON) treatment. (A) Myotube cultures from Patient 42273 (BMD) were treated with two AONs to pseudoexon 11a: A11 and D11 in combination and separately at 10, 30, 100, and 300nM. (B) Myotube cultures from Patient 43012 (DMD) were treated with two AONs to pseudoexon 45a (AONs 45 and 45D). These were used either separately at 10, 30, 100, and 300nM or in combination (eg, 300nM of each, dosed simultaneously). RNA was harvested 24 hours after transfection. WT = wild type.

Next, time-course experiments were undertaken to determine the persistence of splicing correction. Cultures were treated with AONs for 4 hours at the optimal concentration for each AON as determined in the first experiment. On day 7 after serum withdrawal (to induce myoblast differentiation), myoblasts were treated with AONs at the concentrations shown in Figure 5. RNA was extracted at 0, 1, 2, 3, and 5 days after transfection. For both of the mutations, using either of the AONs resulted in a detectable increase of correctly spliced mRNA on the day of transfection. Corrected splicing reached an apparent maximum on day 1 and persisted at apparently the same level until day 5 (see Fig 5). A longer time course was not tested because each patient's primary myotube cultures started to die after 12 days of differentiation (5 days after transfection), regardless of transfection status.

Figure 5.

Reverse transcription polymerase chain reaction analysis of time course of induction of pseudoexon skipping. (A) Myotube cultures from Patient 42273 (Becker muscular dystrophy [BMD]) were treated with 300nM of either antisense oligonucleotide (AON) A11 or D11 and assayed 0, 1, 2, 3, and 5 days later. (B) Myotube cultures from Patient 43012 (BMD) were treated with 100nM of either AON 45 or 45D and assayed 0, 1, 2, 3, and 5 days later. WT = wild type.

The expression of dystrophin protein was assessed only in the myotubes from Patient 43012. This DMD patient has no convincing dystrophin expression by immunofluorescent analysis of muscle sections. However, correlating with the RNA studies showing increased wild-type mRNA, a significant amount of dystrophin was present in cultured myotubes at 5 days after treatment, as detected by immunoblot analysis, which shows expression of apparently full-length dystrophin (Fig 6).

Figure 6.

Expression of full-length dystrophin after antisense oligonucleotide (AON) 45 treatment of 43012 (Duchenne muscular dystrophy) myotubes. GAPDH was used as a control for protein loading. WT = wild type.


The application of the most current molecular diagnostic techniques to the dystrophinopathies allows the identification of the majority of DMD mutations via blood sample analysis alone. However, the usage of these techniques has also led to the increasing recognition of a novel class of mutations, consisting of deep intronic mutations resulting in pseudoexon inclusion, as causative of DMD and BMD. Here, we discuss intronic point mutations resulting in the creation of novel splice sites (although both intronic deletions and classical exonic deletions could also result in the creation of novel splice sites, with pseudoexon incorporation into mRNA). At a practical level, this class is not currently readily detectable from genomic DNA because of both the cost and effort required for direct sequence analysis of the large introns in the DMD gene and the need to establish definitively the RNA splicing implications of putative splice-affecting mutations. mRNA can be obtained from lymphocytes for performance of the protein truncation test22; however, because no significant dystrophin expression is thought to occur in lymphocytes, abnormal results risk reflecting nonpathogenic splice variation occurring in the setting of nonphysiological transcription. Similarly, diagnosis of splice variation can be made from primary fibroblast cultures induced to differentiate into the myogenic pathway via MyoD transfection,23 although this is more time and labor intensive than analysis of muscle-derived mRNA. For these reasons, a role remains for muscle biopsy in the diagnosis of pseudoexon mutations. The exact prevalence of these mutations is unknown, but they may account for many of the 4 to 7% of mutations not detected by genomic mutation analysis methods.8, 9

All three patients we describe carry mutations that result in aberrant mRNA that would be predicted by the reading-frame rule1 alone to cause DMD. Patient DC0160 has an in-frame pseudoexon in which a premature stop signal is encoded. He is an informative case regarding the difficulty presented to the clinician in offering prognosis in the dystrophinopathies, particularly in the absence of a family history to guide prognostication. Based on serum creatine kinase levels, and the absent dystrophin staining on a clinical biopsy, he was given a clinical diagnosis of DMD. However, he remains relatively strong and ambulant at 15 years old; this pattern may be suggestive of BMD, but the consensus is that he has been responsive to steroid therapy. (Notably, he has no family history from which the phenotype in nonsteroid-treated relatives can be determined.) Although insufficient tissue was available from the clinical biopsy for qRT-PCR studies, mRNA sequencing showed no evidence of significant amounts of a second mRNA species, such as may be seen with stop codon–encoding point mutations that induce exon skipping via alteration of ESE elements.24 This raises the possibility that the premature stop codon signal may be subject to ribosomal readthrough, which is sequence context dependent.25 However, readthrough does not appear to correlate well with phenotype in DMD patients,26 and in only a single case has readthrough been implicated in ameliorating a disease phenotype.27

In the other two patients, with out-of-frame pseudoexons, we hypothesized that the difference in disease severity between BMD and DMD would be explained by a differential efficiency in the function of the mutation-induced splice site, and hence a differential amount of pseudoexon inclusion. qRT-PCR analysis of the ratio of wild-type and mutant DMD splice products confirmed that the intronic splice mutations are, in fact, disease-causing.

As expected, in the DMD Patient 43012, no dystrophin mRNA corresponding to the normal transcript (splicing exon 45 to exon 46 was detected). The lack of normally spliced dystrophin mRNA predicts the absence of dystrophin protein and correlates with the DMD phenotype observed for this patient. The altered splice product 45A-46 containing the 139 nt cryptic exon was found at 40% the level of correctly spliced dystrophin mRNA in unaffected skeletal muscle, indicating at least partial resistance to NMD and the potential expression of a severely C-terminally truncated protein (although none was seen on immunoblot of biopsy tissue, suggesting that such a product may be subject to rapid degradation). In contrast, in the patient with BMD, inefficient splicing at the intronic cryptic splice site results in the presence of wild-type exon 11/12 splicing at 13% of the levels found in unaffected control skeletal muscle. This residual amount of normally spliced dystrophin mRNA produces detectable amounts of apparently full-length dystrophin. Even this low level of expression is apparently sufficient to maintain partial dystrophin function, as evidenced by dystrophin membrane localization and amelioration of the dystrophinopathy phenotype. The results demonstrate that our hypothesis was correct: dystrophin expression and a less severe BMD phenotype are due to residual levels of wild-type dystrophin transcript produced by normal splicing and to relatively inefficient altered splicing to include the pseudoexon.

The difference in splicing efficiency between these two patients is likely to be attributable to the strength of the donor splice sites created by the point mutations. Alternatively, the primary determinant could be the strength of the cryptic acceptor splice sites that are present in the wild-type sequence and activated by the point mutations. In addition, genetic backgrounds may also play a substantial role, as suggested by the description of a family with variable severity arising from a nonsense mutation in exon 29 that compromised recognition, and hence inclusion of that exon in the mature dystrophin mRNA.28 Our data favor the primary importance of the mutation-created splice site. The consensus sequence for 5′ splice donor sites (Table 2) corresponds to perfect Watson–Crick base pairing of the U1 small nuclear RNA 5′ terminus29; this base pairing plays a critical role in 5′ splice-site selection and spliceosome assembly. In Patients 42273 and 43012, the intronic point mutations create 5′ donor splice sites. Notably, the 5′ splice donor created by the mutation in Patient 43012 displays better complementarity to the U1 small nuclear RNA consensus sequence than does the 5′ splice donor created by the mutation in Patient 42273; this may explain the higher level of abnormal splicing found in Patient 43012 compared with Patient 42273. The strength of the 43012 splice donor site is also predicted to be greater by several independent algorithms (see Table 2).

Table 2. Pseudoexon 5′ and 3′ Splice Sites and Predicted Scores
inline image

In contrast, and arguing against the importance of the cryptic acceptor site, all algorithms predict the 3′ acceptor site in the DMD patient (43012) to be the weakest of the three in our set. Potentially, the stronger 5′ splice site can activate splicing of the downstream intron more efficiently.30, 31 In addition, a complex interplay of other factors, such as branch point position and strength, ESE and silencer positions, and intron lengths,30, 32 may play a role in the more efficient splicing of the pseudoexon in Patient 43012.

Significant rescue of the DMD phenotype by low levels of wild-type splicing suggests that this class of mutations may benefit from exon-skipping therapies directed toward the pseudoexon splice junctions. When tested in cultured cells, splicing correction was observed for both the BMD and DMD samples whether treated with AON to the corresponding ESE or to the 5′ splice donor site. Interestingly, treatment with both AONs to the pseudoexon 45A sequence corrected splicing more efficiently than with the ones to the pseudoexon 11A. Theoretically, one might predict the pseudoexon 11A to be more responsive to the AON treatment because it inherently has weaker splicing capabilities as shown by RT-PCR. Observation of the reverse effect suggests that both AONs directed toward the pseudoexon 45Aa were better at inhibiting splicing than those selected for exon 11A.

Pseudoexon skipping may, in fact, be less problematic than “constitutive” exon skipping, in which the patient's mutated transcript is further altered by deletion of one or more exons to restore an open reading frame. In both exon and pseudoexon skipping, one of the theoretical problems is that the new exon junctions may result in novel epitopes, which may be immunogenic. However, low levels of exon skipping are common in DMD patients33 (and in the mdx mouse34) where they result in revertant fibers, which express dystrophin proteins with a restored open reading frame. Although we detected no measurable levels of the wild-type mRNA in our DMD patient sample, even exceedingly low levels of wild-type transcript may result in epitope tolerization to the wild-type protein. More importantly, pseudoexon skipping resulting in wild-type protein can reasonably be expected to have a more significant benefit than exon skipping resulting in a BMD-like internally deleted protein.

Development of pseudoexon-skipping therapies represents a type of personalized medicine, directed at individual patients with private mutations. In contrast, exon-skipping therapies have entered trials for regions of the gene likely to benefit a large number of patients.35 Nevertheless, our results suggest that pseudoexon mutations may be highly amenable to such therapies and point out the continued utility of muscle biopsy in the diagnosis of this novel class of mutations.


This work was supported by the NIH (National Institute of Neurologic Diseases and Stroke, R01 NS043264, K.M.F., M.T.H., R.B.W.; T32 NS07493, O.G.), the National Center for Research Resources (M01-RR00064), and the Association Francaise Contre les Myopathies (K.M.F.).

We acknowledge the study coordinator assistance of K. Hart, the technical assistance of L. Zhao, and the University of Utah Core Imaging Facility (C. Rodesch).