Molecular Modeling and Phenotypic Description of a Patient with a Novel Exonic Deletion of GALNS with Resultant Morquio Syndrome with Two Successful Pregnancies

Abstract

In this report, we describe phenotypic features of a patient with mucopolysaccharidosis type IVA (Morquio syndrome) harboring a novel exon 1 deletion in GALNS with enzymatic confirmation consistent with Morquio syndrome. To our knowledge, this is the first reported case of this variant. Additionally, we protein modelled wild-type GALNS and the pathogenic variant with an exon 1 deletion for comparative analysis using statistical mechanics methods described herein. We demonstrate that, even when the protein is translated, the mutation would affect protein stability and function via homodimer interaction modifications. Lastly, given the patient's 2 successful pregnancies, data about the management of pregnancies in mucopolysaccharidoses are reviewed, and we discuss the management of pregnancy in patients with Morquio syndrome.

Introduction

Mucopolysaccharidoses (MPS) are lysosomal storage disorders caused by the deficiency of enzymes required for the stepwise breakdown of glycosaminoglycans, previously known as mucopolysaccharides. The phenotypic severity of the disorder depends upon both the amount of residual enzyme activity and the distribution and turnover of the affected substrate. Morquio syndrome is a form of autosomal recessive MPS type IVA (MIM #253000), which results from pathogenic variants in GALNS (MIM *612222) encoding galactosamine-6-sulfatase (GALNS), on 16q24.3 [Baker et al., 1993]. The enzyme catalyzes the degradation of 2 glycosaminoglycans, chondroitin-6-sulfate and keratan sulfate. Pathogenic variants in GALNS may cause misfolding of the enzyme, resulting in the disruption of the catabolic process, and accumulation of the undegraded substrate in lysosomes of affected tissues. Additional pathomechanisms include lack of protein translation, and/or nonsense-mediated mRNA decay. MPS IVA is characterized phenotypically by skeletal dysplasia, mild corneal opacities, hepatosplenomegaly, valvular heart disease, hearing loss, and enamel hypoplasia. Most patients have identifiable biallelic missense pathogenic variants in GALNS as the molecular basis of their disease [Tomatsu et al., 2005; Morrone et al., 2014].

The heterogenicity of the pathogenic variants accounts for a degree of the variable age of onset and progression; patients with the severe form of this disease rarely survive beyond the third decade of life, and patients with the attenuated form may survive over 70 years. Here, we describe one novel pathologic variant, an exon 1 deletion in GALNS, including the diagnostic journey leading up to the late diagnosis of Morquio syndrome and discussion of the patient's mild clinical phenotype, specifically focusing on pregnancy outcomes.

There are 3 main objectives of this manuscript, to (1) describe the phenotype of this patient with a novel exon 1 deletion and discuss reasons why a second pathogenic allele was not found, (2) demonstrate that neoteric protein modeling techniques can be utilized as confirmation of pathogenicity of novel variants, and (3) detail the pregnancy outcomes in this patient and review current literature of pregnancy outcomes in Morquio syndrome.

Case Report

A 42-year-old female patient came to our clinic for evaluation of joint pain and stiffness that has been present from childhood. She self-referred after many years of failed medical work-up and herself believing there was something more to her apparently disparate medical symptoms. In childhood, she was diagnosed with spondyloepiphyseal dysplasia tarda but lost to medical follow-up. Per her report, her parents were not interested in pursuing a potential unifying diagnosis for her as a child via further genetic testing. Interestingly however, she did see a geneticist as a child (over 30 years ago), and the possibility of Morquio syndrome was raised. The patient reported that urine biochemical testing was performed and normal, although given the time passed since this testing no records were available to review. During her life, she has undergone 2 hip replacement surgeries at the ages of 25 and 32, respectively.

On physical exam, she is of significantly short stature, with the clear appearance of skeletal dysplasia (Fig. 1). There is a short broad barrel chest, marked hypermobility overall with stiffness in select joints such as the left elbow, ulnar deviation of wrists, waddling gait, genu valgum, and pes planus.

Fig. 1.

Proband's family and proband. The proband is in the back row, second from the right. The proband's sister, back row middle, presents a similar clinical phenotype to the proband and is suspected to have MPS IVA as well. The proband's mother and father are front row, first and second from the right, respectively. The remaining family members are the proband's sisters with no significant medical history.

Other medical history includes cholecystectomy, hearing difficulty (conductive hearing loss), depression, anxiety, osteopenia, and a C-section hernia. Except for an elder sister who has a similar clinical presentation, all other family members do not have any signs and symptoms of skeletal dysplasia (Fig. 1, 2). Testing performed on her sister with the similar clinical presentation at an outside facility confirmed the presence of Morquio syndrome per report.

Fig. 2.

A 3-generation pedigree. The proband is indicated by an arrow.

Although her obstetric history was complicated by 3 first trimester miscarriages, she has had 2 successful pregnancies with phenotypically normal children without any signs of skeletal dysplasia.

The patient followed up with our clinic around 4 years after initial presentation. She had a trial period of ∼18 months of enzyme replacement therapy (ERT) but was no longer receiving this treatment due to lack of perceived efficacy. She reported significant bone pain in her shoulders, elbows, hips, knees, and ankles requiring regular opioid medication. Additional history included 2 artificial hip and 2 artificial knee replacements, lower tone abnormalities on audiogram, and sleep apnea.

Sequence analysis and deletion/duplication testing for 52 genes associated with lysosomal storage disorders (Invitae Lysosomal Storage Panel) showed a novel heterozygous pathogenic variant, a deletion in exon 1 of GALNS (NM_000512.4). This pathogenic variant of GALNS is a gross deletion of the genomic region encompassing exon 1, which includes the initiator codon. Regarding coordinates, the laboratory report stated “The 5' end of this event is unknown as it extends beyond the assayed region for this gene and therefore may encompass additional genes. The 3' boundary is likely confined to intron 1 of the GALNS gene. This is expected to result in an absent or disrupted protein product.”

Subsequent enzyme analysis for MPS as part of BioMarin's “Simply Test For MPS” initiative via Greenwood Genetics commercial labs demonstrated a low N-acetyl galactosamine-6-sulfatase activity within the range consistent with a diagnosis of Morquio syndrome (3.78, normal 49–255 nmol/17 h/mg protein).

The genetic analysis was performed prior to the enzymatic analysis above using a commercial laboratory (Invitae Corporation). The reasons for this were the faster turnaround time of the genetic testing and ease of sample collection (saliva). Briefly, Invitae describes the methods of genetic analysis as follows: “Genomic DNA obtained from the submitted sample is enriched for targeted regions using a hybridization-based protocol and sequenced using Illumina technology. Unless otherwise indicated, all targeted regions are sequenced with ≥50× depth or are supplemented with additional analysis. Reads are aligned to a reference sequence (GRCh37), and sequence changes are identified and interpreted in the context of a single clinically relevant transcript. Enrichment and analysis focus on the coding sequence of the indicated transcripts, 10bp of flanking intronic sequence (20bp for BRCA1/2), and other specific genomic regions demonstrated to be causative of disease at the time of assay design. Exonic deletions and duplications are called using an in-house algorithm that determines copy number at each target by comparing the read depth for each target in the proband sequence with both mean read-depth and read-depth distribution, obtained from a set of clinical samples. All clinically significant observations are confirmed by orthogonal technologies, except individually validated variants and variants previously confirmed in a first-degree relative.”

Subsequent to the enzymatic confirmation of Morquio syndrome, Invitae was contacted to review their molecular findings for a second pathogenic allele with consideration of RNA analysis; however, they declined all additional analyses citing commercial restrictions.

Materials and Methods

Protein Informatics and Molecular Modeling

Our methodology has been documented previously in the literature [Macklin et al., 2018; Richter et al., 2018; von Roemeling et al., 2018; Cohen et al., 2019; Hines et al., 2019a, b]. The sequence of the human protein GALNS, a protein encoded by the GALNS gene, was taken from the NCBI Reference Accession Sequence NM_000512.5 and was used for computer-assisted modeling. Monte Carlo simulations were performed on the mutant to allow local regional changes for full-length 522 amino acids and when exon 1 deletion variant was introduced (only 482 amino acids). The GALNS protein forms a homodimer that localizes within the lysosome, as is shown to be the case from the X-ray structural data set. The Protein Data Bank (PDB) codes used for modeling and building the full-length structures and running simulations include the following: 4FDI and 4FDJ, which have excellent coverage (497 of 522 amino acids) within the 2.2-Å resolution X-ray crystal structures. Modeling allows us to determine the statistical mechanics differences that accumulate over time from doing simulations on how the variant may alter the core structure over time as a dynamical process (see www.karger.com/doi/10.1159/000519326 online suppl. Material 1).

Results

Molecular Modeling and Thermodynamic Measurements

Examinations of the GALNS homodimer reveal an interaction zone that consists of residues, shown, that form a pore-like circular interface (2 sheets of anti-parallel beta sheets layered onto 2 alpha helices). These residues are known to be critical for the formation of GALNS homodimerization contacts, which is crucial for the function in transcription regulation.

Wild-Type GALNS Homodimer Interactions

The N-terminus interaction set of residues around the homodimer interface that are critical for complex formation (Fig. 3a) are these homodimer residues: monomer 1 (chain A) at W43, V48, Y49, P52, S53, R54, L68, P70, Q250, R251, Y254, L293, A296, P297, E298, S302, P305, F306, R319, V335, E337, Q397, P420, Q422, N423, V424, S425, E443, L444, R446, D447, P448, R451, F452, P453, L454, and E460 interact with monomer 2 (chain B) at W43, V48, Y49, P52, S53, R54, E55, P70, Q250, R251, R253, Y254, L293, A296, P297, E298, Q299, S302, N303, P305, F306, L307, R319, V335, E337, Q397, F412, P420, Q422, N423, V424, E443, R446, D447, P448, R451, L454, S455, S458, and E460 (Fig. 3).

Fig. 3.

Rearrangement of the GALNS region around exon 1 deletion. a Full-length model for the entire wild-type (WT) GALNS structure that shows dimer interaction. b Zoom into the interaction region for the WT GALNS showing the native conformation for homodimer interface. c Full-length model for the entire exon 1 deletion GALNS structure that shows dimer interaction with compromised structure. d Zoom into the interaction region for the exon 1 deletion GALNS showing the altered conformation for homodimer interface.

Exon 1 Deletion GALNS Homodimer Interactions

The N-terminus half maintains a different set of residues around the revised homodimer interface during attempted complex formation (Fig. 3c), which include these homodimer residues: monomer 1 (chain A) at W43, V48, E51, P52, S53, R54, L68, P70, N71, Q250, R251, L293, P297, E298, S302, P305, F306, L307, R319, V335, E337, Q397, P420, Q422, N423, V424, E443, R446, D447, P448, R451, F452, and E460 interact with monomer 2 (chain B) at W43, V48, E51, P52, S53, R54, L68, P70, N71, Q250, R251, L293, P297, E298, S302, P305, F306, L307, R319, V335, E337, Q397, P420, Q422, N423, V424, E443, R446, D447, P448, R451, F452, and E460 (Fig. 3).

Root-Mean-Square Deviation Assessments

The deletion of exon 1 residues (M A A V V A A T R W W Q L L L V L S A A G M G A S G A P Q P P N I L L L L M D D) significantly disrupts the structure around this region, which is visually evident and an essential interaction (Fig. 3c, d) [Reumers et al., 2005; Schymkowitz et al., 2005; Caulfield, 2011; Caulfield and Medina-Franco, 2011; Lopez-Vallejo et al., 2011; Caulfield and Devkota, 2012; Zhang et al., 2013]. Additionally, we measured the root-mean-square deviation (RMSD) for the residues mentioned above between the wild-type and variant structure, which revealed a large structural deviation due to the mutation. The RMSD at a cutoff of 6 Å from the site of missing residues as given above resulted in >8 Å deviation for all residues, >7 Å for the backbone residues, and >6.5 Å for the C-alpha residues only. By any metric of RMSD, those are significant changes indicative of a different conformational state that has the potential to alter function.

Energetics Assessment

Object Stability as Metric of Structural Integrity

For wild type versus the variant, we found that the stability of the object from energetic calculations for ΔG per amino acid (averaged over entire object) is greater for the exon 1 deletion variant than the wild-type protein with 608.5 and 317.51 kcal/aa × mol × Å² total, respectively, such that exon 1 deletion and wild-type protein are 1.24 and 0.608 kcal/aa × mol × Å², per amino acid, respectively [Reumers et al., 2005; Schymkowitz et al., 2005; Caulfield, 2011; Caulfield and Medina-Franco, 2011; Lopez-Vallejo et al., 2011; Caulfield and Devkota, 2012; Zhang et al., 2013]. A ΔG of 317.51 kcal/(mol × Å²) is normal for a protein of 522 aa size (Fig. 3a). However, the exon 1 deletion variant in the N-terminus region generates a much larger positive energy (both stability and for homodimer interaction), thus is quite disruptive, and the Monte Carlo analysis showed some significant rearrangement for the N-terminus around residues W43 through P70, which are demonstrably more prone to conformational variability and play an important role in proper complex stabilization (Fig. 3c, d).

Homodimer Interaction Energies for Complex Stability

More important are the interaction energies (ΔΔG) for the monomer 1:monomer 2 interface that maintain the dimer structure, which are compromised in the exon 1 deletion structure (Fig. 3). The ΔΔG values for the wild type and exon 1 deletion variant are −3.88 and 12.68 kcal/(mol × Å²⁾, respectively. The negative energies are attractive, and the positive energies are considered repulsive, so the overall impression is that the exon 1 deletion is prone to not forming dimers as readily. The net difference of >16 kcal/mol between the 2 homodimers is many magnitudes difference based on log scaling of these energies, so the damaging nature of this deletion is very evident.

However, the protein with exon 1 deletion results in a change of protein function that could come as a consequence of the alteration of the region of residues near to the homodimer site and also the general interaction. The molecular model for the full structure and its truncated form are given (Fig. 3a, b, 4a, c) using our state-of-the-art methods, which have been established [Caulfield, 2011; Caulfield and Medina-Franco, 2011; Caulfield et al., 2011, 2014, 2015; Lopez-Vallejo et al., 2011; Caulfield and Devkota, 2012; Abdul-Hay et al., 2013; Zhang et al., 2013; Fiesel et al., 2015a, b; Ando et al., 2017; Puschmann et al., 2017]. Overall, it is demonstrated that the exon 1 deletion is quite deleterious to the protein's stability and function.

Fig. 4.

GALNS molecular model for the full-length human sequence consisting of 522 amino acids and exon 1 deletion variant (482 aa). All protein residues shown in licorice rendering and using standard element coloring (O-red, N-blue, H-white, S-yellow), where carbons are shown in orange, blue to indicate the monomer derived for dimer. Homodimerization and interaction residues are shown, colored by a GALNS monomer (blue, orange, gray). a Full-length model for the entire wild-type (WT) GALNS structure that shows dimer interaction. Each monomer of the dimer is indicated by a different color of ribbon (blue, orange). b Poison-Boltzmann electrostatic surface for the WT GALNS structure. c Full-length model for the entire exon 1 deletion structure of GALNS with defective dimer interaction is shown. Monomers are similarly indicated as in panel a (light blue and light orange ribbons). d Poison-Boltzmann electrostatic surface for the exon 1 deletion GALNS structure.

Discussion

Biallelic pathogenic variants in GALNS encoding N-acetylgalactosamine-6-sulfate sulfatase are associated with autosomal recessive MPS IVA. In around 16% of patients with Morquio syndrome, molecular confirmation cannot be reached with traditional gene sequencing unless complemented by other diagnostic modalities. Even then, it is reported that in around 5% of patients, a second pathogenic allele evades detection, as in our reported case. One explanation is the second pathogenic variant may lie in a deep intronic region that is not covered by commercially available standard sequencing panels as performed above. One other possibility is chromosome 16 isodisomy. These phenomena have been described in other genetic and metabolic defects [Atwal, 2014; Atwal et al., 2014, 2016; Donti et al., 2016; Sunde et al., 2016; Blackburn et al., 2018].

In a recent report by Caciotti et al. [2018], 3 patients who previously did not harbor a second pathogenic mutation were found to have a deep intronic mutation leading to aberrant splicing events. To rule out aberrant splicing events in these patients, mRNA analysis was needed; however, per our case, it may not be feasible to perform mRNA analysis routinely in the clinic presently, as funding and insurance coverage are limiting. We suspect that in our patient, like the previous reports, the second disease-causing pathogenic variant may be a deep intronic pathogenic variant affecting splicing. Given no dominant forms of Morquio have been described, and this does not generally fit with the pathomechanism of an enzyme deficiency, we are confident there is a second pathogenic allele in our patient. This is evidenced by the severe enzyme deficiency observed. This second variant is likely a mild one, allowing the production of some, although minimal, residual enzyme activity. The second mild variant may confer the “mild” phenotype observed in our patients, which also allowed the phenotype to be mild enough to support 2 successful pregnancies.

Our case presented with an attenuated form of Morquio syndrome, nevertheless this phenotype was immediately apparent to the biochemical geneticist (author P.S.A.) she consulted with. Indeed, the patient was already aware of a suspected diagnosis in childhood, thus one could argue she has been on a 30+ year diagnostic odyssey. Misdiagnosis of spondyloepiphyseal dysplasia tarda in cases of Morquio syndrome has been described previously and is more of a concern in attenuated forms such as our patient. The utilization of BioMarin's Simply Test for MPS aided in getting her the precise diagnosis quickly without having to ensure insurance coverage for this testing. She was subsequently initiated on ERT at an outside institution, and it will be interesting to follow her symptomology longitudinally for improvements.

The protein modeling described above demonstrates molecular alterations between the wild-type and mutant GALNS proteins. The main differences to note are at the homodimer interaction point, and the net difference of >16 kcal/mol between the wild-type and mutant homodimers at the interaction site; this difference is very significant based on log scaling of these energies, and results in higher state of repulsion of the homodimers in the mutant protein. This repulsion is evidenced by the conformational changes and increased distance between amino acid resides at the homodimer interaction sites in the mutant GALNS. These sophisticated protein modeling techniques allow for detailed insights into dynamic protein interactions and a much richer mechanistic insight into enzyme biology; we can explore the pathomechanisms at orders of magnitude higher scrutiny compared to rudimentary protein prediction techniques such as Polyphen/SIFT2. The protein modeling is demonstrated an additional research avenue of exploration into how these abnormal proteins do not function like wild-type. As such, some protein mutations such as early exon deletions that may result in null alleles may not benefit from protein analysis in this manner.

Finally, our patient has had 3 first trimester miscarriages followed by 2 successful pregnancies. She was considered high-risk during her pregnancy and the delivery because of her skeletal dysplasia, and hence a C-section was recommended prior to delivery. Interestingly this C-section resulted in a hernia, which is observed in higher frequency in MPS disorders when compared to the general population.

Both her children were born healthy and did not have any phenotype suggestive of skeletal dysplasia. As the standard of care for MPS patients has improved, leading to an increase in life expectancy, a growing number of MPS adults may consider reproduction [Stewart et al., 2016b]. Fertility studies on humans with MPS are limited, and no significant fertility issues in females with MPS have been reported [Stewart et al., 2016a]. Stewart et al. [2016b] noted that women with MPS are of small stature, with obstructive and restrictive respiratory disease, cardiac issues, potential for spinal cord compression, hepatosplenomegaly, and musculoskeletal abnormalities, all of which may interfere with normal pregnancy and delivery or exacerbate due to pregnancy. The literature on this subject is quite limited, and most published case reports describe pregnancy and delivery in patients with MPS type I [Stewart et al., 2016b]. In one case, a woman with severe MPS IVA developed polyhydramnios and dyspnea in the 21st week of gestation, and she subsequently delivered a premature baby at 28 weeks (1.18 kg, 38 cm) [Stewart et al., 2016a]. The baby had anemia, hyperbilirubinemia, apnea, and bronchopulmonary dysplasia; however, it was discharged in a medically stable condition [Stewart et al., 2016a]. Whether to use ERT during pregnancy and lactation is a complex issue for the treating physician. The risk of exacerbation of the disease associated with the cessation of ERT must be carefully weighed before deciding to stop use during pregnancy and/or lactation. It is essential that more reports describing the outcome of ERT use in patients during pregnancy/lactation with MPS be published to aid in the future management in high-risk cases as the current literature is extremely limited.

Regarding ERT use and pregnancy outcomes in MPS, one case reports a female with MPS I who discontinued ERT use after she became pregnant and, through normal delivery, gave birth prematurely to a female weighing 1.25 kg at 29 weeks gestation. Of note, after discontinuation of ERT her condition rapidly declined [Stewart et al., 2016a]. In another case, a patient is described that did not discontinue ERT use for pregnancy nor while lactating for 3 months. The infant from this pregnancy did not go on to develop any antidrug (laronidase) antibodies and no laronidase was detected in the breast milk [Stewart et al., 2016a].

Overall, the number of pregnancies described in MPS patients remains relatively low. Awareness must be provided to women with MPS about the risk and complications associated with pregnancy and delivery and subsequently being able to take care of their child [Stewart et al., 2016b]. Furthermore, patients with MPS must be able to overcome the financial demands related to such a pregnancy [Stewart et al., 2016a]. Since MPS patients have a significant anesthetic risk secondary to their underlying cardiac condition, most physicians, as in our patient's example, prefer a scheduled cesarean section under epidural or spinal anesthesia to a natural delivery (as the latter could require an emergency cesarean section potentially under general anesthesia) [Stewart et al., 2016b]. In one case, as reported by Drummond et al. [2015], complete paraplegia was evident immediately postoperatively after an apparently uneventful lumbar epidural using general anesthesia in a patient with Morquio Type A syndrome with moderate thoracic spinal stenosis. Finally, while many issues can occur during pregnancy and delivery, in most reported cases, the children were born healthy [Stewart et al., 2016b]. A further area of research will be to determine the effects of skeletal complications and cardiovascular complications on fertility issues of patients with MPS IVA.

In summary, we describe the clinical and biochemical phenotype of a patient with a novel exon 1 deletion, we demonstrate a neoteric protein modeling technique that can be utilized as confirmation of pathogenicity of novel variants in Morquio and other genetic disorders, and we detail the pregnancy outcomes in this patient and review current literature of pregnancy outcomes in Morquio syndrome.

Statement of Ethics

Ethical approval for this study was provided by the Ethics Committee of the Mayo Clinic, USA. Informed consent and the permission for the publication of the accompanying images were obtained.

Conflict of Interest Statement

All authors declare that there are no conflicts of interest and no disclosures.

Funding Sources

We acknowledge receipt of BioMarin Medical Education Grant “2019-RS-003407”.

Author Contributions

Pavalan Selvam, Angita Jain, Jessica Abbott, and Abhimanyu S. Ahuja contributed to the conception of the work, review of the clinical interpretation of the molecular data, and drafting of the manuscript. Anvir Cheema, Katelyn A. Bruno, and Herjot Atwal contributed to the drafting of the manuscript. Thomas Caulfield contributed to the development and analysis of the molecular modeling work and review of the manuscript. Irman Forghani revised the manuscript from a clinical standpoint. Paldeep S. Atwal contributed to the design, conception, and supervision of the project and did a thorough revision of the scientific work and content.

Data Availability Statement

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author. Given the confidential nature of the patient records, it is not publicly available.

Supplementary Material

Supplementary data