“Missing mutations” in MPS I: Identification of two novel copy number variations by an IDUA‐specific in house MLPA assay

Abstract Background Mucopolysaccharidosis type I (MPS I) is a rare, recessively inherited lysosomal storage disorder, characterized by progressive multi‐systemic disease. It is caused by a reduced or absent alpha‐l iduronidase (IDUA) enzyme activity secondary to biallelic loss‐of‐function variants in the IDUA. Over 200 causative variants in IDUA have been identified. Nevertheless, there is a fraction of MPS I patients with only a single mutated IDUA allele detectable. Methods As genetic testing of MPS I is usually based on sequencing methods, copy number variations (CNVs) in IDUA can be missed and therefore presumably remain underdiagnosed. The aim of this study was the detection of CNVs using an IDUA‐specific in house multiplex ligation‐dependent probe amplification (MLPA) assay. Results A total of five unrelated MPS I patient samples were re‐analyzed after only a single heterozygous IDUA mutation c.979G>C (p.A327P), c.1469T>C (p.L490P), c.1598C>G (p.P533R), c.1205G>A (p.W402X), c.973‐7C>G (p.?) could be identified. We detected a novel splice site variant c.973‐7C>G (p.?), as well as two novel CNVs, a large deletion of IDUA exon 14 and 3’UTR c.(1828 + 1_1829‐1)_(*1963_?)del, and a large duplication extending from IDUA exon 2 to intron 12 c.(157 + 1_158‐1)_(1727 + 1_1728‐1)dup. Conclusion Together with the CNVs we previously identified, a total of four pathogenic IDUA CNVs have now been reported.

According to the severity of the disease, clinical signs and symptoms of MPS I are present in the first decade of life, including organomegaly, obstructive airway disease, heart disease, skeletal deformities, growth retardation, neurological complications, and severe mental retardation (Beck et al., ; Cleary & Wraith, ). However, the phenotypical spectrum is continuous and the classification is frequently complex (Poletto et al., ). The clinical diagnosis of MPS I is confirmed biochemically based on elevated GAGs in urine, and reduced or absent IDUA activity in leucocytes or skin fibroblasts (Beesley et al., ; Scott et al., ). While clinical and biochemical abnormalities are present, genetic background still remains unresolved in a fraction of MPS I patients (Bunge et al., ; Ghosh et al., ; Scott et al., ).
The spectrum of causative genetic variants in IDUA is highly heterogeneous as more than 200 single-gene defects have been reported, including missense, nonsense, splice site, insertions, as well as small deletions and duplications (Bertola et al., ). Recently we described a novel mutational mechanism for MPS I: two distinct large IDUA-deleterious copy number variations (CNVs) detected by an in house multiplex ligation-dependent probe amplification (MLPA) approach (Breen et al., ).
In this study we used the same copy number screening tool as well as the Sanger-based sequencing to re-analyze five unrelated MPS I DNA samples, which were initially tested positive for a pathogenic variant on a single IDUA allele.

| Ethical compliance
The study was approved by the Ethics Committee of the Faculty of Medicine of Friedrich-Schiller-University Jena (reference number: 2018-1107).

| Patients
Following publication of the first two IDUA deletions and introduction of an IDUA MLPA assay (Breen et al., ), the corresponding author was contacted by several clinicians who had MPS I patients in whom only a single mutated IDUA allele could be identified. DNA samples with consent for extended genetic workup were eventually enrolled. Table 1 summarizes the geographic, clinical, biochemical, and genetic backgrounds of the patients investigated.

| Genetic analyses
The IDUA coding sequence plus >50 nucleotides of neighboring UTR or intronic sequence was amplified from genomic DNA (primers available upon request). PCR products were gel-purified and Sanger-sequenced from both directions using a commercial service (Macrogen Europe, Amsterdam, The Netherlands).
IDUA-specific MLPA was based on the synthetic kit presented by us previously (Breen et al., ). In the frame of the study, this assay was extended with additional MLPA probes; corresponding oligonucleotides were purchased from Biolegio (Nijmegen, The Netherlands). Supporting Information Table S1 AB provides all target sequences and product sizes. Variants were described at cDNA and protein level using reference sequences NM_000203.4 and NP_000194.2, respectively.

| Sanger-sequencing detects a single heterozygous IDUA variant in five unrelated patients with a clinical diagnosis of MPS I
All patients enrolled in this study had a clinical diagnosis of MPS I, but Sanger-sequencing performed in a routine diagnostic setting had only detected a single heterozygous IDUA mutation. By repeating Sanger-sequencing of the whole gene, we confirmed the presence of the previously reported heterozygous variants (Figure 1), and the lack of additional, potentially pathogenic variants (data not shown  Information Table S2).

IDUA-specific MLPA assay suggests the presence of a heterozygous deletion in one patient and the presence of a heterozygous duplication in another
An IDUA-specific MLPA assay as presented by us previously (Breen et al., ) suggested normal diploid IDUA copy number in three of the five patients. In the sample with the heterozygous c.1469T>C variant, the signal for a 3'UTRspecific probe was reduced by ~50%, while in the sample with the heterozygous c.1598C>G variant, the signals for exons 2, and 7 were increased by ~50% (data not shown). These observations suggested the heterozygous presence of one presumably large deletion and one presumably large duplication, respectively.

| Additional MLPA probes confirm the two suspected IDUA CNVs, define their extent, and characterize the remaining three samples as definitely CNV-negative
Based on the above described suggestive observations, we designed additional MLPA probes for exons not covered by our initial assay, and applied them to the two samples of interest. A second 3'UTR-specific probe confirmed the deletion in sample with the heterozygous c.1469T>C variant. Two novel probes which target the penultimate exon 13 and a sequence near the stop codon in exon 14, respectively, revealed that the deletion was restricted to exon 14, and that it overlapped with the coding sequence. Its 3'-extend, however, could not be delineated by the set of MLPA probes used (Figure 2a). The duplication as suggested by the increased MLPA signals for exons 2 and 7 was first confirmed by a novel MLPA probe against exon 5. Based on additional probes, its 3'-extend was subsequently mapped to intron 12 (Figure 2b). The application of all novel MLPA probes to the three samples that had remained negative upon analysis with the original MLPA assay did still not reveal evidence for IDUA CNVs (data not shown).

| DISCUSSION
This study reports on a heterozygous deletion and a heterozygous duplication which partially affect the IDUA. Several lines of evidence support pathogenicity of these CNVs. First, they are not listed by variation databases (ExAC, database F I G U R E 1 Molecular findings based on Sanger-sequencing in five unrelated MPS-I patients. Exemplary sequence traces showing the hereozygous presence of five pathogenic IDUA variants (red arrow) in DNA from index patients (I-V) as well as in DNA from mother of patient IV. Father of patient IV carries a benign single nucleotide variant rs115929690 (green arrow) sugesting that patient IV has inherited a pathogenic variant from his mother and a benign variant from his father, respectively of genomic variants, Decipher). Second, they were identified in MPS I patients in whom conventional Sanger-sequencing had only identified a pathogenic IDUA variant on one allele. Third, they are predicted to result in reduced IDUA mRNA levels. For the duplication, this is based on the out-of-frame nature of the duplicated region, and nonsense-mediated mRNA decay as the likely consequence (Nagy & Maquat, ). The deletion, which involves the gene's 3'-end including some coding nucleotides and the stop codon, should entail nonstop-mediated mRNA decay (Hamby, Thomas, Cooper, & Chuzhanova, ; Rebelo et al., ). By thereby representing bona fide loss-of-function alleles, both CNVs thus resemble the majority of already known pathogenic variants in IDUA (Bertola et al., ; Poletto et al., ). We conclude that the above arguments represent strong cumulative evidence for pathogenicity of both the deletion and the duplication. We previously presented an MPS I patient who carried two distinct large IDUA deletions (Breen et al., ). Together with the findings presented here, a total of four pathogenic IDUA CNVs have now been identified. This number is small compared to the long list of known "small" aberrations [HGMD]. However, a fraction of patients for whom homozygous variants have been reported may in fact be compound heterozygous for this variant and a large overlapping deletion. In addition, MPS I patients with only one identified mutant IDUA allele are frequently encountered (Bertola et al., ; Scott et al., ; Uttarilli et al., ). The existence of hitherto undetected IDUA CNVs can therefore be expected. The availability of MLPA assays has greatly facilitated the detection of deletions as well as duplications in many other inherited disorders (Günther et al., ). Our corresponding tool will be made available to researchers interested in more accurately defining the prevalence of IDUA CNVs.
We compiled a total of five MPS I samples in which a second pathogenic variant could not be identified. Two of these could eventually be explained by pathogenic CNVs, while for three samples a second mutation remains to be discovered. One may therefore hypothesize that undetected/unscreened variants in the promotor, in other regulatory elements or in deep intronic regions (Beesley et al., ; Vazna et al., ) and/ or large genomic inversions (Scott, Litjens, Nelson et al., ;