DISCUSSION
DMD is one of the most prevalent among the rare diseases (Mendell et al., 2012). It is caused by sequence variants in DMD , the largest gene of the human genome, which is located in a genomic region with high rates of recombination. These characteristics make DMD highly susceptible to mutation (Nguyen & Yokota, 2019). The current work presents a familial case of DMD with a cxSV, a multi-technique and bioinformatic approach was implemented in pursuit of 2 major aims. The first one focus on the molecular diagnosis and genetic assessment of the family, including manifesting carrier and prenatal studies. While the second one, centered on finding possible explanations of the origin of SVs in DMD .
The design of the prenatal diagnosis molecular strategy was based on the previously detected DMD deletion of exons 45-54 in the affected alive male (III4). On the one hand, we performed a CVS maternal cell contamination study to determine the usefulness of the fetal material. On the other hand, we decided to evaluate the presence of the deletion by 2 different means, MLPA and the co-segregation of the deletion by STRs. Therefore, 2 STRs mapping within the deletion and 2 flanking it were selected for the haplotype study.
A priori, we were expecting that II1 and II3 shared III4´s deletion, so it was a surprise when studies revealed the existence of 2 SVs in the family. A duplication of exons 38-43, shared by all family members, probing to be the ancestral familial causing mutation. Also, the well-known deletion, found only in III4, suggesting being the second mutational event i.e. a de novo alteration. Therefore, this case works as an example of the importance of retesting historic patients diagnosed by multiplex-PCR, where deletion´s boundaries might not have been precisely determined or other SVs could have been missed out, affecting eligibility and effectiveness of mutation-dependent therapies.
We would like to hallmark the usefulness of the old-fashioned STRs segregation studies as a tool to identify homologous recombination and STRs retraction/expansion events, determine parental origin of the haplotypes, corroborate sample´s gender and discard CVS maternal cell contamination. In this case, STRs revealed a recombination between STR44 and STR62 in III4, a retraction of STR49 in III1, the existence of 2 different paternal X-chromosomes in II1 and II3, an estimation of female gender in the fetus and the exclusion of CVS contamination.
Regarding the AR-Assay, the skewed XIP observed in II1 suggests that the fully active X-chromosome should be the maternal one, carrying the duplication, giving the presence of DMD symptomatology. On the contrary, as expected, her asymptomatic sister (II3) had a random inactivation. Because of the STRs and MLPA results, II1 and II3 were already expected to present different paternal X-chromosomes but to share the maternal one, as both carried the duplication. Remarkably, they did not share anyAR allele, corroborating the double paternity and suggesting another recombination event that separated the DMD causative mutation from the AR allele.
Although, AR-assay has been recently validated on amniotic fluid for DMD symptomatology prediction, we could not test the XIP in the fetus, as their parents did not agree on performing a second puncture (He et al., 2019). There are conflicting reports in the literature about the heritability of the XIP (Renault et al., 2007; Viggiano et al., 2017). Therefore, future XIP studies may be needed in case any clinical symptoms arise in III1.
Several authors reported the occurrence of non-contiguous rearrangements, within the same DMD allele, with frequencies up to 2% (Kerr et al., 2013). In our studied cohort, we estimated a cxSV rate of 1.4% (6/437 DMD patients), encompassing deletions-duplications, non-contiguous duplications and large deletion plus a 20pb insertion. The present case offered an extra advantage, the possibility of establishing a mutational timeline thanks to familial segregation study of the detected cxSV.
Molecular diagnosis encompasses the study of index cases, usually lacking a complete familial analysis. The obtained molecular alterations have a “photo effect”, like a picture of the resulting rearrangement at a specific moment careless of timeline determination. This lack of information is more obvious when dealing with cxSVs, where is impossible to differentiate between 2 concurrent alterations and 2 different mutational events that take place in different times/generations. Studying chronologically these cxSVs could be useful to unravel if the first alteration acts as a predisposing factor to the latter.
Focusing on our second aim, centered on finding the molecular mechanisms underlying the origin of the observed SVs, we decided to precisely determine their breakpoints. For this purpose, we used several practical approaches, including SNP-array, WGS and customized PCRs. Followed by a bioinformatic approach, screening the surroundings of the SVs’ breakpoints for the identification of DNA DSB stimulator motifs, which have been found statistically significantly more frequent at breakpoint junctions than expected by chance according to Abelleyro et al(2020) (Abelleyro et al., 2020).
While deletion breakpoints were neatly and accurately identified by SNP-array and LR-PCR, the characterization of duplication breakpoints was particularly challenging, as the SNP-array could not certainly delimitate the single/double-doses transitions. It was only after WGS analysis that the identification of the reads mapping on the tandem duplication breakpoint junction permitted its precise delimitation. Therefore, these results remark the importance of revalidating the SV´s boundaries detected by SNP-array.
Regarding the tandem segmental duplication, this event may be adequately explained either by MMBIR or FoSTeS models, both involving de novo DNA synthesis. However, considering the principle of maximum parsimony and the complexity of the event, characterized by at least 3 strand invasions preceded by replication fork collapses and facilitated by microhomologies, supports a FoSTeS mechanism over MMBIR (Figure 5A). In addition, the necessary DNA strand break collapses may have been stimulated by the presence of Jurka hexa-nucleotides, secondary structures as well as repetitive elements at relevant locations (Vissers et al., 2009).
Concerning the deletion, it seems to occur de novo concurrently with a meiotic recombination event switching STRs alleles between maternal homologous, given the coincident X-chromosome location of both molecular events. The most likely scenario suggests that these two concomitant molecular events may actually represent a single inter-chromosomal recombination event, which could have taken place either in II3’s oocyte meiosis or in II3´s oogonia mitosis, making the deletion inheritable. The above mentioned evidence indicates that II3 carries the DMD duplication, but does not have the deletion in peripheral blood, suggesting that the complex series of events took place in her germline resulting in a chimeric X-chromosome with STRs alleles from both progenitors (I1 and I2) and both SVs in phase. Furthermore, our recombination hypothesis relies on the presence of repetitive elements relatively near, both, 5’ and 3’ deletion breakpoints, supporting a putative inter-chromosome non-allelic pairing structure or ectopic synapsis suitable for a localized recombination (Figure 4) (Liu et al., 2012).
At the molecular level, the breakpoint junctions of the deletion point out the hypothesis of a classical NHEJ, given that the event does not present any molecular characteristic that could imply an association with a DNA replication dependent mechanism. We detected just a single-thymine of microhomology, which does not provide sufficient evidence to sub-classify the event as MMEJ (Figure 5B). Moreover, supporting the classical NHEJ model DNA DSB stimulating motifs were found at both deletion breakpoints, such as Jurka hexa-nucleotides and SAR on the 5’ breakpoint and Ig heavy chain switch region on the 3’ deletion breakpoint (Gale et al., 1992; Jurka, 1997; Rabbitts et al., 1981). Other possible stimulators for DBS, possibly exerting its action, could be LTRs, STRs and highly stable secondary structures, which have also been detected at both deletion breakpoints. Furthermore, 2 of these motifs might be related, as long STRs have the ability to form non-B-DNA structures.
Resuming the chronology of the molecular events that gave rise to the cxSV, the duplication could have favored the approach of the regions involved in the deletion and stabilize an unequal pairing between the duplicated X-chromosome and its unduplicated counterpart. This theory supports both previously mentioned hypothesis for the origin of the deletion. On the one hand, the duplication could have helped the previously mentioned repetitive elements in the formation of the ectopic synapsis during female gametogenesis. While, it could have also been implicated in the unequal pairing that resulted in NHEJ mechanism.
On the other hand, we want to highlight the important role that secondary structures must be exerting in the origin of SVs, given that we found them in 7/8 breakpoints. Its implication for DSB formation is suggested by the fact that if we randomly select 50pb throughout the genome it is very likely that they form secondary structures, but if we simulate random breakpoints and analyze 25bp at each side of it, the chances of finding secondary structures involving them are greatly reduced (Abelleyro et al., 2020).
In conclusion, the strategy implemented in the present work allowed us to provide an accurate diagnosis and genetic assessment, enabling early diagnosis and the selection of the appropriate standard of care and mutation-specific treatments. Finally, the thorough characterization of the SV´s breakpoints and the analysis of their molecular scars not only widen the understanding of the molecular mechanisms involved in their generation, but also may help the development of new therapeutic approaches.