Newborn screening of duchenne muscular dystrophy specifically targeting deletions amenable to exon-skipping therapy

Duchenne Muscular Dystrophy (DMD) is a lethal progressive muscle-wasting disease. New treatment strategies relying on DMD gene exon-skipping therapy have recently been approved and about 30% of patients could be amenable to exon 51, 53 or 45 skipping. We evaluated the spectrum of deletions reported in DMD registries, and designed a method to screen newborns and identify DMD deletions amenable to exon 51, 53 and 45 skipping. We developed a multiplex qPCR assay identifying hemi(homo)-zygotic deletions of the flanking exons of these therapeutic targets in DMD exons (i.e. exons 44, 46, 50, 52 and 54). We conducted an evaluation of our new method in 51 male patients with a DMD phenotype, 50 female carriers of a DMD deletion and 19 controls. Studies were performed on dried blood spots with patient’s consent. We analyzed qPCR amplification curves of controls, carriers, and DMD patients to discern the presence or the absence of the target exons. Analysis of the exons flanking the exon-skipping targets permitted the identification of patients that could benefit from exon-skipping. All samples were correctly genotyped, with either presence or absence of amplification of the target exon. This proof-of-concept study demonstrates that this new assay is a highly sensitive method to identify DMD patients carrying deletions that are rescuable by exon-skipping treatment. The method is easily scalable to population-based screening. This targeted screening approach could address the new management paradigm in DMD, and could help to optimize the beneficial therapeutic effect of DMD therapies by permitting pre-symptomatic care.


Results
Database review. Our approach was to identify patients amenable to exon-skipping therapy by screening for individuals with hemi(homo)-zygotic deletions of the flanking exons of exon-skipping targets. Identification of a deletion of any of these flanking exons does not, however, invariably involve a genotype that could benefit from these treatments. For example, a deletion of flanking exon 46 may indicate a deletion of exons 45-47, or a deletion of exons 46-49. The in-frame deletion of exons 45-47 would not benefit from exon-45 skipping, while the out-of-frame deletion of exons 46-49 can be rescued by skipping of exon-45 (Fig. 1A).
To determine the efficiency of our approach, we assessed the percentage of DMD patients reported in the LOVD-DMD registry that would be identified by our assay and would be rescuable by either exon-51, exon-53 or exon-45 skipping therapy (Fig. 1B). This analysis also shows that assaying the flanking exons is an effective approach to identify many patients that could benefit from these therapies (Table 1). qPCR results. We analyzed 120 samples with known genotypes. Deletion of any of the five target exons (exons 44, 46, 50, 52 and 54) of the DMD gene was characterized by the absence of amplification of the corresponding target. As shown in Table 2, the assay correctly identified all subjects, with either presence or absence of amplification of the target exon. To corroborate the absence of amplification, for each exon, an NFR below 0.2 was found as being discriminant between the deleted group and the carrier and control groups.
The 34 DMD patients of the "deleted group" with a deletion overlapping at least one of the five target exons were correctly characterized by an absence of fluorescence of the corresponding probes. The other 17 male patients of the "deleted group", with a deletion of a DMD exon not targeted by our assay, presented a clear amplified profile of all target exons. All subjects of the female "carrier group" and the "control group" also were characterized by a normal significant fluorescent signal of each probe. Our assay did not discriminate between

Discussion
We show the feasibility and reliability of a simple test to detect newborns with DMD that would be eligible for exon skipping therapy. Our study is a proof-of-concept approach to NBS for DMD using a qPCR method. We set up a high-throughput assay to target deletions that are amenable to exon-51, exon-53 or exon-45 skipping treatments (i.e. screening for deletion of exons 44, 46, 50, 52 and 54 of the DMD gene). We analyzed 51 DMD patients, 50 female DMD gene deletion carriers and 19 healthy individuals without any deletion of the DMD gene. Each of the 120 subjects was correctly categorized by the assay with either the presence or absence of  www.nature.com/scientificreports/ amplification of the target exon corresponding to their known genotype. Should a very rare variant affect the specificity of our primers and probes, we expect scoring a false positive result due to absence of amplification. DMD is devastating muscle disease with an early onset. However, due to a 1.3-2.5 year delay in clinical diagnosis after the appearance of first symptoms, the mean age at diagnosis of DMD patients is 4.4 years [25][26][27] . Since early treatment is recommended [16][17][18] , reducing the delay before diagnosis might be aided via the inclusion of DMD in newborn screening performed under existing public health screening protocols. To the best of our knowledge, only exon-skipping is being studied to date in children as young as six months of age (ClinicalTrials identifier: NCT03218995). Pre-treating patients before gene therapy using an exon-skipping approach could potentiate the effect of gene therapy. Such pre-treatment would allow the use of lower-and therefore safer-doses of vector to bring about a higher level of dystrophin expression in the long term 28,29 . Table 2. Amplification results of the different subject groups: "deleted group", "carrier group" and "control group". a = male patients with a DMD deletion not overlapping any of the 5 target exons. b = female subjects carrying a DMD deletion overlapping at least one of the 5 target exons. c = female subjects carrying a DMD deletion not overlapping any of the 5 target exons. d = Deleted exons were identified through MLPA. Intronic breakpoints have not been sequenced; genotypes are reported according to the LOVD-DMD database. *Results show if the target exon is deleted (Del) or amplified (Ampli). N/A Not Applicable. www.nature.com/scientificreports/ The availability of personalized treatments represents a paradigm shift in the management of the subset of DMD patients that are amenable to their use, particularly as treatment can now be initiated pre-symptomatically. Deployment of our test raises some challenges as it addresses a subset of patients in which the use of an exon skipping therapy is possible. For those patients and their families, the benefit of early diagnosis of DMD is clear. For the large proportion of patients that are currently not suitable for these therapies, the ongoing development of gene therapy offers hope, albeit that significant clinical hurdles need to be overcome. For these patients, early diagnosis is still relevant to allow for the timely introduction of physiotherapy and corticosteroids. As part of evolving improvements in the treatment of DMD, we believe that our approach represents an effective and practical optimization of diagnosis. As therapeutic options expand, this test could also evolve as part of a more comprehensive DMD testing strategy when combined with other techniques.
Initiatives have previously been undertaken in various countries to explore the feasibility and efficacy of NBS for DMD [19][20][21][22][23] . Those studies, relying either on CK or CKMM quantitation, achieved modest sensitivity and specificity. The Wales DMD newborn screening program, for example, has carried out CK quantitation as a first-tier test on more than 340,000 neonates. The CK assay has shown a sensitivity of 81.6% with a positive predictive value of 38.6% 20 . More recently, a CKMM assay was evaluated in two independent populations in a limited number of newborns 23 . That study reported one false negative result on a premature infant. Authors also showed that both the gestational age and the age of the newborn at time of sampling affecting CKMM levels. This can become an issue when screening large populations with significant subgroups of preterm neonates, and could greatly impact the number of second-tier molecular tests required to confirm DMD diagnosis. Comparatively, while restricted to a small number of samples, our assay presented a sensitivity of 100%. Experience shows that qPCR assays dedicated to the NBS of spinal muscular atrophy (i.e. a technically similar assay) have very high sensitivities and specificities [30][31][32] . We suggest that there is a reasonable expectation that the DMD assay we describe could perform similarly in larger cohorts. www.nature.com/scientificreports/ This study acts as a feasibility study for designing a subsequent wider NBS project for DMD. Technically, since our real-time PCR reader is equipped with five fluorescence channels, a two-step validation protocol is necessary. Results showing an amplification of either one or more exons are acceptable, as any amplified exon can act as a reference sequence. Samples showing no amplification for any of the five exons (i.e. any DMD deletion covering exons 44 to 54) would be re-run in a 2-plex qPCR reaction, combining each individual exon with the RPP30 reference gene to confirm the deletion. In the context of neonatal screening, positive results would be confirmed on a fresh-EDTA sample using the MLPA assay, as the test is already commercially available. MLPA testing, combined with an analysis of the RNA transcript, is also mandatory to establish precisely the individual genotype, confirming whether a patient is eligible for exon-skipping therapy.
Regarding the experimental conditions, the qPCR assay we describe is highly practical, robust and easily scalable to population-based screening. It is also not economically burdensome (less than 5 euros per sample), considering that the DNA is already extracted for NBS of other molecular genetic diseases. These other targets include spinal muscular atrophy 33,34 another devastating genetic condition without circulating diagnostic biomarkers for which innovative medications have dramatically changed the prognosis when administered early, and that has prompted several NBS programs [30][31][32] . In the future, societal and medical cost-effectiveness evaluations could be carried out on large population to compare the "qPCR/MLPA" protocol we describe with the two-tier "CKMM DMD genotyping" strategy proposed by Mendell et al. 24 . In the context of developing an NBS algorithm for DMD, our qPCR assay could also be considered as a second-line fast-track procedure to rapidly identify whether CKMM positive samples would be eligible for exon-skipping therapies.
Although DMD is expected to affect only male patients, and the costs of processing the sorting of samples by gender greatly exceed the low costs of analysis, we envisage performing screening on both male and female neonates. We could thereby identify the very rare cases of DMD in women with co-existing Turner syndrome 35 . On the other hand, a few cases of female carriers expressing a DMD phenotype have been described in situation of X-chromosome inactivation (XCI) skewing 36 . Our assay is not intended to identify carriers of a DMD deletion, and females with a DMD phenotype related to non-random patterns of XCI would be missed.
DMD patients eligible for either exon-51, exon-53 or exon-45 skipping treatments account for approximately 30% of DMD cases 37 . With an incidence evaluated at about of 1/5,000 males in the general population 38 , approximately 1/15,000 males could be identified using our approach, which is far above other commonly accepted disorders in NBS programs. Our multiplex PCR is not intended to detect all deletions in the DMD gene nor can it identify duplications or point mutations. A negative result on the current assay does therefore not definitively rule out a diagnosis of DMD, since about 70% of DMD patients are not eligible for either exon-51, exon-53, or exon-45 skipping therapies and would not be identified by our assay. Nevertheless, the spectrum of DMD genotypes potentially rescuable by single or double antisense-mediated exon skipping is constantly growing. The actual applicability of exon skipping approach, assessing which exon(s) should be skipped to restore the open reading frame in DMD gene, could theoretically benefit 83% of patients with DMD 37 . Should any future DMD therapies targeting other exons or specific variants be developed, our method could be easily scaled-up to target DMD exons other than 51, 53 and 45, thereby extending its applicability to a wider panel of DMD genotypes. qPCR technology has some limitations and cannot identify the entire panel of variants described in the DMD gene. Next-generation sequencing (NGS) represents the ultimate approach to screen newborns for DMD. Notwithstanding, NGS is currently too expensive to be used for full population screening for the foreseeable future.
Our methodology could raise some ethical concerns regarding DMD patients who would be left out by our assay. Our approach is, however, not the first initiative of NBS prompted by innovative therapies that misses some cases affected by a particular disease. Indeed, NBS for spinal muscular atrophy is nowadays largely accepted amongst pediatricians, families and advocacy groups, although screening assays specifically focus only on the identification of homozygotic deletions of the SMN1 gene and miss SMN1 heterozygotic and SMN2 deletions. Screening for spinal muscular atrophy using this method was added to the Recommended Uniform Screening Panel (RUSP), the official list of disorders which US public health departments use to screen newborns, despite the availability of only one approved treatment. Type 0 spinal muscular atrophy patients were, therefore, identified by screening, but could not benefit from any specific treatment available at that time.
The DMD population diagnosed at birth with our test accounts for approximately 9532/26078 (36.5%) in the LOVD database. Overall, 5144/26078 (19.7%) could be offered gene therapy. The remaining 4388/26078 (16.8%) that are diagnosed with exon deletions but are not amenable to current therapy, also have significant potential benefits stemming from early diagnosis. In this latter group, physiotherapy and steroids could be implemented years earlier than the current situation in which they are simply waiting for DMD to manifest clinically. Currently, physiotherapy and corticosteroids are typically initiated about the age of 4 years, even in patients diagnosed earlier. At that time, more than half of patients will have been diagnosed 39 . NBS of all DMD patients could reduce diagnosis journey and anticipate familial recurrence, but since the age at diagnosis is not the main driving factor for initiating disease-specific treatment like steroids, there is no more rationale to screen today all DMD patients than any recessive rare disorders, which explain why DMD is not usually part of NBS programs. The approval of new disease-modifying treatment brings a new insight for a subset of patients with a deletion amenable to treatment. The introduction of our novel screening test would thus not negatively impact the current management of DMD cases, while adding an easy, non-expensive, and efficient way to detect 100% of the patients that are eligible for gene therapy.
One of the limitations of this approach could be the development of non-mutation specific treatments such as microdystrophin. Nevertheless, gene therapy as a one-shot injection will probably not be applicable before the age of two years, and is currently in trials after the age of 4. The risk of a dilution effect through growth could limit the treatment of newborns. In this context, the pre-treatment of eligible exon skipping therapy patients from birth until microdystrophin injection could help to potentiate gene therapy treatment 40

Conclusion
In conclusion, we provide proof-of-concept of a test that can potentially detect DMD patients at birth that are eligible for currently approved exon skipping 51, 53 and 45 therapies. The development of this assay was driven by the approval of exon-skipping drugs in DMD management. Our test addresses a significant paradigm shift in NBS, by identifying newborns amenable to a specific available treatment, rather than detecting all newborns with a particular disease. This method meets a specific current -and possibly transitory-need in DMD care, focused on the identification of patients that could benefit from exon-skipping therapies. Early identification of DMD holds promise for preserving functional performance in affected children via the initiation of treatment at a young age. Should the current assay's performance be further validated in larger studies, it could act as a means to improve DMD patient identification and management in the clinical setting for countries where exon-skipping drugs are available.

Materials and methods
DMD registry analysis. We reviewed the Leiden Open Variation Database for the DMD gene (LOVD-DMD), that currently contains more than 26,000 patient entries 41 . We evaluated the spectrum of reported deletions to provide a theoretical rationale for the development of an assay identifying deletions that could benefit from either exon-51, exon-53 or exon-45 skipping therapies.
Subjects. Blood from DMD and control samples were collected as dried blood spots (DBS). DMD gene profile of all patients was performed using Multiplex Ligation-Dependent Probe Amplification (MLPA) assay in the course of diagnostic workup. In total, 120 samples were collected and were classified into three different subgroups: The "deleted group" consisted of 51 male patients with a clinical DMD phenotype. Amongst them, 34 had a confirmed deletion of one or more DMD exons amenable to exon-skipping targeted therapies (i.e. exons 44, 46, 50, 52 or 54). The other 17 male patients carried a deletion of another DMD exon.
The "carrier group" included 50 females with one deleted DMD allele, of whom 32 carried a deletion that overlapped at least one of the target exons 44, 46, 50, 52 or 54, while 18 had a deletion of any other DMD exon.
The "control group" consisted of 19 normal individuals with no DMD deletion/duplication or point mutations.
qPCR Technical design. The DMD genotyping assay was designed and validated to detect hemi-(homo)zygotic deletions of DMD exons 44, 46, 50, 52 and 54. The protocol uses a multiplex quantitative polymerase chain reaction (qPCR) assay. DNA was extracted from one 3.1-mm dried blood spot according to the protocol described previously in 33 . Isolated DNA was not quantitated, and 1 µL of freshly extracted DNA was mixed with 5 × Takyon master mix (Eurogentec), primers and probes in a total volume of 25 µL. The sequences and concentrations of primers and probes are shown in Table 3. Primers and probes have been designed to have the fewest possible described polymorphisms (Supplemental Table 1). Table 3. qPCR primers and probes used to amplify and identify DMD Exons 44, 46, 50, 52 and 54.