Genotype and Phenotype Characterization of Patients with Mucopolysaccharidosis IV-A in Chile

Abstract

Introduction

Morquio syndrome or mucopolysaccharidosis type IV-A (MPS IV-A) is an autosomal recessive disease caused by biallelic variants in the GALNS gene, encoding the lysosomal enzyme GalN6S, responsible for glycosaminoglycan keratan sulfate and chondroitin-6-sulfate degradation. Studies have shown that the degree of evolutionary and chemical divergence of missense variants in GalN6S when compared to ancestral amino acids is associated with the severity of the syndrome, suggesting a genotype-phenotype correlation. There is little information on Latin American patients with MPS IV-A that replicate these findings. This study aimed to characterize the phenotype and genotype from patients with MPS IV-A, who are under Enzyme Replacement Therapy at the Children’s Neuropsychiatry Service of the Hospital Clínico San Borja Arriarán, Santiago, Chile, and to determine if there is any association between genotype and phenotype with those findings.

Methods

Information was collected from medical charts, all patients went through a GalN6S enzymatic activity measurement in leukocytes from peripheral blood, and the GALNS gene was sequenced for all cases.

Results

12 patients with MPS IV-A were recruited, all patients presented multisystem involvement, mostly skeletal, and 75% of cases underwent surgical interventions, and cervical arthrodesis was the most frequent procedure. In regards of the genotype, the two most frequent variants were c.319+2T>C (n = 10, 41.66%) and p.(Arg386Cys) (n = 8, 33.33%), the first one was previously described in 2018 in a patient from Chile [Bochernitsan et al., 2018].

Conclusion

This is the first time that a genotype-phenotype correlation has been studied by analyzing the variants effect on the molecular structure of human GalN6S and the evolutionary conservation degree of affected residues in a cohort of patients in Chile. Albeit our work could not find statistically significant associations, we may infer that the evolutionary conservations of affected amino acids and the effect of variants on enzyme structure may play a main role. Further analyzes should consider a meta-analysis of published cases with genotype data and larger samples and include other variables that could provide more information. Finally, our data strongly suggest that variant c.319+2T>C could have a founder effect in Chilean patients with MPS IV-A.

Introduction

Mucopolysaccharidosis type IV-A (MPS IV-A), also known as Morquio A disease, is an autosomal recessive lysosomal storage disease belonging to the mucopolysaccharidoses group, caused by the deficiency of one of the enzymes responsible for the Glycosaminoglycan (GAG) degradation in the lysosome, which causes an accumulation of GAGs in different tissues at the cellular level [Hendriksz et al., 2013]. MPS IV-A is the result of biallelic variants in the GALNS gene located at the 16q24.3 locus that encodes the N-acetylgalactosamine-6-sulfatase GalN6S lysosomal enzyme [Tomatsu et al., 1991], responsible for GAG keratan sulfate and chondroitin-6-sulfate catabolism [Kresse et al., 1982]. Its most obvious characteristic is skeletal dysplasia as the GAGs are mostly accumulated in bone and joint tissues. Consequently, patients with MPS IV-A manifest craniofacial dysmorphia and multisystemic involvement [Wraith, 1995]. However, the clinical manifestations are variable, from severe (classical phenotype) to mild (attenuated phenotype) conditions, which are classified according to the age of onset, height, and degree of multisystemic involvement [Hendriksz et al., 2013].

Several studies have demonstrated the presence of clinical similarities among siblings with MPS IV-A [Tylki-Szymanska et al., 1998; Rekka et al., 2012], providing some elements to suggest the existence of a genotype-phenotype association. In 2000, the effect of variants on the GalN6S function was identified for the first time, establishing that certain types of variants were associated to the severity of the disease [Sukegawa et al., 2000]. These findings were then corroborated through the use of bioinformatic tools [Sudhakar, 2011]. During 2012, the GalN6S human structure was determined using X-ray crystallography, which characterized the catalytic domain of the enzyme [Rivera-Colon et al., 2012]. This study enabled us to understand genotype-phenotype correlations from a structural perspective. However, other studies have also shown that the structure of the enzyme is not the only determining factor in the clinical expression of the disease. Indeed, the degree of evolutionary conservation of the mutated regions and the chemical conservation of amino acids was found to be crucial as well [Tomatsu et al., 2006]. More recently, other studies have described the correlations between genotype, phenotype, and QS levels in blood and urine [Tomatsu et al., 2004; Dung et al., 2013]. Of note, most genotype-phenotype correlation studies have been carried out in Caucasian populations or people from high-income countries, and there is a scarcity of studies that replicate these findings in other populations such as Latinos. Therefore, our work aimed to describe the genotype and phenotype from a cohort of patients with a confirmed diagnosis of MPS IV-A under control at the Pediatric Neuropsychiatry Service of the San Borja Arriarán Hospital in Santiago, Chile, and to determine a genotype-phenotype association with those findings.

Materials and Methods

Patients and Clinical Information

All patients with a confirmed diagnosis of MPS IV-A by enzymatic activity measurement in leukocytes from peripheral blood and under medical monitoring and treatment at the Children's Neuropsychiatry Service of the Hospital Clínico San Borja Arriarán (Chile) were invited to take part in this study. Eligible participants were recruited between June 2017 and June 2018, and each one of them gave their consent to be included in this study. Demographic data (gender, ethnicity, and city of origin) were collected. Clinical data (height and multisystemic clinical and radiographic findings), medical history, and enzymatic activity were compiled for analysis. The presence of severe (classical phenotype) or mild forms (attenuated phenotype) is determined considering the onset of symptoms, height, and the degree of multisystem involvement [Hendriksz et al., 2013]. For this purpose, patients with age of onset <1 year, a final height ≤120 cm, and a greater degree of organ involvement were considered to suffer from classic phenotype. Patients with symptoms starting during the second decade of life, a final height ≥140 cm, and a lower degree of organ involvement were classified as an attenuated phenotype.

An informed consent was signed by the adult patients, while the underage patients signed an assent prior to the authorization from their legal guardians. This study was conducted in accordance with the Declaration of Helsinki and was previously reviewed and approved by the Ethics Committee of the Central Metropolitan Health Service.

GALNS Genetic Analysis

DNA Extraction and Sequencing

A peripheral blood sample was obtained from patients and then transferred to filter paper. This paper was sent for genomic DNA extraction and molecular analysis of the GALNS gene to the Medical Genetics Service laboratory of the Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil. Bidirectional Sanger sequencing of all exons and exon-intron boundaries of the GALNS gene was performed using previously published primers.

Bioinformatic Analysis

The variants obtained from the molecular analysis were interpreted according to the American College of Medical Genetics and Genomics (ACMG) guidelines [Richards et al., 2015] using the InterVar program (http://wintervar.wglab.org) [Li and Wang, 2017] and Alamut Visual version 2.11 (Interactive Biosoftware, Rouen, France) using the canonical transcript NM_000512/ENST00000268695. The distribution and frequencies of variants in the GalN6S enzyme distribution were plotted using the MutationMapper program (http://www.cbioportal.org/mutation_mapper) [Cerami et al., 2012]. For missense variants, structural analysis was performed using the 4fdj entry code from the Protein Data Bank database [Burley et al., 2018] in the Chimera version 1.12 program (https://www.cgl.ucsf.edu/chimera/download.html) [Pettersen et al., 2004], and an evolutionary conservation analysis was perform with the Clustal Omega program (https://www.ebi.ac.uk/Tools/msa/clustalo/) [Zerbino et al., 2018]. For this purpose, the GalN6S canonical protein sequences described in humans were used (Homo sapiens, 28 ENSP00000268695) along with the corresponding orthologs in chimpanzees (Pan troglodytes, ENSPTRP00000014434), mice (Mus musculus, ENSMUSP00000015171), rats (Rattus norvegicus, ENSRNOP000015171), cats (Felis catus, ENSFCAP00000012076), hens (Gallus gallus, ENSGALP00000053922), frogs (Xenopus tropicalis, ENSXETP00000024579), zebrafish (Danio rerio, ENSDARP00000124843), and fruit flies (Drosophila melanogaster, FBpp0311532). These sequences were obtained from the Ensembl genome database version 95 (Annex 2) [Zerbino et al., 2018]. For consequence prediction of splice-site variants, the NNSPLICE program version 0.983 (http://www.fruitfly.org/seq_tools/splice.html) was used [Reese et al., 1997].

Statistical Analysis

Continuous variables were described using medians and interquartile ranges. Categorical variables were described using absolute and relative frequencies. The enzyme activity was described as a continuous variable in the tables, but for statistical analysis, it was converted in a categorical one (detectable or undetectable, considering 20 nmol/17 h/mg as the cutoff) [Shams et al., 2017]. For analyses of genotype-phenotype correlations, patients were organized according to several dichotomous independent variables: (1) variant type (those with at least one missense mutation vs. those with protein truncating mutations only, i.e., nonsense mutations, splicing, deletions, among others); (2) zygosity (homozygous vs. compound heterozygous); (3) degree of evolutionary conservation reported in patients with at least one missense variant (highly conserved residue[s] variant[s] vs. evolutionarily variable residue[s]); (4) impact on GalN6S structure in those patients with at least one missense variant (those patients with missense variant[s] affecting processing, catalytic site, or hydrophobic core versus those with only superficial missense variant[s]), according to enzyme structure [Rivera-Colon et al., 2012]. Residual enzyme activity and clinical phenotype were the dependent variables.

The Fisher’s exact test was performed for bivariate analyses using the statistical software STATA version 13.1. A two-tailed p < 0.05 value was considered as statistically significant.

Results

General and Clinical Characteristics of the Sample

Twelve patients with MPS IV-A were eligible for this study. The median age at enrollment was 13.79 years (25th percentile 11.20–75th percentile 25.49). Sex distribution presented a similar proportion between groups (7 women and 5 men). The distribution by ethnic group showed 11 patients from Chile (8 from admixed groups and 3 from Amerindian populations) and one Colombian patient from admixed origin. Eleven patients presented the classic phenotype and 1 patient the attenuated phenotype (Table 1).

Table 1.

Symptoms exhibited in patients with MPS IV-A

Patient	Sex	Ethnic origin	City of birth	Age of suspicion, years	Age of confirmation, years	Age of recruitment, years	ERT start age, years	Time on ERT, years	Height, cms	Skeletal compromise	Neurological compromise	Cardiological compromise	Respiratory compromise	Visceral compromise	Ophtalmological compromise	Hearing compromise	Phenotype
1	F	Mestizo	Punta Arenas	1.83	6.75	10.58	7.58	3.00	95	HL bilateral DM	Motor DD-CSS	Mild LVH	Bronchial asthma	HM	No	Mild SNHL	Classic
2	M	Mestizo	Santiago	5.25	33.41	37.08	34.00	3.08	109	DM	Motor DD-CSS	Mild LVH	Mild RRC	IH	CO	Severe SNHL	Classic
3	F	Mestizo	Sevilla Valle	0.90	1.50	19.75	17.08	0.41	98	DM	Motor DD-CSS myelopathy	No	Mild RRC	HM	CO	No	Classic
4	F	Mestizo	Melipilla	2.25	7.08	24.16	20.66	2.41	112	DM	Motor DD-CSS	No	Mild OSAS	No	CO	Mild SNHL	Classic
5	M	Mestizo	Aysén	1.91	4.58	12.75	9.08	3.41	102	DM	Motor DD	Mild LVH	Central apneas	No	No	No	Classic
6	F	Mestizo	Aysén	0.08	0.50	8.50	5.16	3.41	89.5	DM	No	SA	No	No	No	Mild SNHL	Classic
7	F	Mestizo	Punta Arenas	2.08	3.50	31.16	27.66	3.66	95	HL bilateral DM	Motor DD-CSS	Mild LVH	Mild RRC	HM	CO	Mild SNHL	Classic
8	M	Amerindian	Puerto Montt	4.08	10.58	14.83	11.58	3.25	113	DM	Motor DD-CSS	Mild LVH	No	No	No	SNHL TAH	Classic
9	F	Amerindian	Calbuco	1.91	4.16	29.50	25.66	3.91	95.5	DM	Motor DD-CSS myelopathy	Mild AI Mild TI	Mild RRC	No	CO	Severe SNHL	Classic
10	F	Mestizo	Puerto Montt	3.75	6.50	11.41	7.83	3.58	99.5	HL bilateral DM	Motor DD-CSS	No	Mild OSAS	HM	No	SNHL TAH	Classic
11	M	Amerindian	Quemchi	1.75	1.91	4.50	2.91	1.66	88	DM	Motor DD-CSS	No	No	UH	No	TAH	Classic
12	M	Mestizo	Ancud	10.08	11.25	12.08	11.66	0.50	148	Bilateral EDFH	No	Mild TI	No	No	CO	No	Attenuated

Skeletal compromise was the main feature. All patients – including the patient with attenuated phenotype – manifested orthopedic complications to a variable degree. SA, sinus arrythmia; RRC, restrictive respiratory compromise; EDFH, epiphyseal dysplasia of the femoral head; DM, dysostosis multiplex; TAH, tonsillar and adenoid hypertrophy; CHL, conductive hearing loss; IH, inguinal hernia; HM, hepatomegaly; SNHL, sensorineural hearing loss; LVH, left ventricular hypertrophy; UH, umbilical hernia; AI, aortic insufficiency; MI, mitral insufficiency; TI, tricuspid insufficiency; HL, hip luxation; CO, corneal opacities; CSS, cervical spinal stenosis; DD, developmental delay; OSAS, obstructive sleep apnea syndrome.

Height was monitored using the Center for Disease Control (CDC) and the MPS IV-A growth charts [Montano et al., 2008]. All patients with the classic phenotype were below the third percentile in CDC charts, while the patient with the attenuated phenotype is located at the 20th percentile.

The symptoms of MPS IV-A patients are presented in Table 1. All patients presented with multisystem involvement, specially skeletal. All of them had orthopedic complications to a variable degree, including the patient with the attenuated phenotype, and 9 of 12 patients required surgical interventions. Cervical arthrodesis was the most frequent surgical procedure, performed in 9 patients, followed by femoral osteotomy and hip reconstructive surgery in 3 patients. Developmental delay, predominantly motor, was found in 10 patients, and 8 showed severe cervical spinal stenosis, of whom 2 patients evolved with myelopathy. Nine patients presented with a variable degree of hearing loss, predominantly sensorineural type; 8 patients had cardiac and pulmonary dysfunction; 6 patients showed corneal opacities on physical examination; and 4 patients had hepatomegaly.

Regarding treatment, all patients have received ERT. Eleven of them are under ERT to date, whereas one quit voluntarily due to an adverse allergic reaction. While the median age of ERT start was 11.62 years (25th percentile 7.76–75th percentile 21.91), the youngest patient was 2 years old and the oldest one was 34 years old. Regarding duration of treatment, the ERT lasted 3.16 years (25th percentile 2.22–75th percentile 3.45); the shortest duration was 5 months and the longest one was 3 years and 11 months.

Genetic Analysis of the Sample

The GALNS gene was sequenced for all patients in our sample. In total, 6 different pathogenic variants were identified in the 24 alleles of 12 patients; 5 of them were missense and 1 was splice-site variant. One of the missense variants had not been described in the literature before. While the detected variants and their features are presented in Table 2, their locations in GalN6S domain are shown in Figure 1. Regarding zygosity, half of the patients were homozygous, and the other half were compound heterozygous.

Table 2.

GALNS gene molecular genetic analysis

Patient	Zygocity	Genomic position (hg19)	Exon/intron	CDS	Protein	Variant type	Database (code)	Reference	Enzyme activity (nmol/17h/mg)	Phenotype	Probable location or effect according to Rivera-Colon et al.	AAa degree of evolutionary conservation	ACMG Criteria of pathogenicity	ACMG interpretation
1	Compound heterozygous	chr16:88891261 chr16:88908303	exon 11	c.1156C>T	p.(Arg386Cys)	Missense	dbSNP (rs118204437)	11, 13, 15, 20, 21, 31, 32	2.00	Classic	Hydrophobic core	Conserved	PS3, PM1, PM2, PP3, PP5	Pathogenic
			intron 3	c.319+2T>C	p.?	Splice-site	ClinVar (VCV001048188.4)	1			NA	NA	PVS1, PS3, PM2, PP3, PP5	Pathogenic
2	Homozygous	chr16:88891261	exon 11	c.1156C>T	p.(Arg386Cys)	Missense	dbSNP (rs118204437)	11, 13, 15, 20, 21, 31, 32	1.00	Classic	Hydrophobic core	Conserved	PS3, PM1, PM2, PP3, PP5	Pathogenic
3	Homozygous	chr16:88898507	exon 9	c.901G>T	p.(Gly301Cys)	Missense	dbSNP (rs118204443)	11, 13, 32, 33	0.80	Classic	Hydrophobic core	Vertebrate-specific	PS3, PM1, PM2, PP3, PP5	Pathogenic
4	Homozygous	chr16:88891260	exon 11	c.1157G>A	p.(Arg386His)	Missense	dbSNP (rs1221167717)	15	0.01	Classic	Hydrophobic core	Conserved	PS3, PM1, PM2, PP3, PP5	Pathogenic
5	Compound heterozygous	chr16:88891261 chr16:88908303	exon 11	c.1156C>T	p.(Arg386Cys)	Missense	dbSNP (rs118204437)	11, 13, 15, 20, 21, 31, 32	0.00	Classic	Hydrophobic core	Conserved	PS3, PM1, PM2, PP3, PP5	Pathogenic
			intron 3	c.319+2T>C	p.?	Splice-site	ClinVar (VCV001048188.4)	1			NA	NA	PVS1, PS3, PM2, PP3, PP5	Pathogenic
6	Compound heterozygous	chr16:88891261 chr16:88908303	exon 11	c.1156C>T	p.(Arg386Cys)	Missense	dbSNP (rs118204437)	11, 13, 15, 20, 21, 31, 32	0.00	Classic	Hydrophobic core	Conserved	PS3, PM1, PM2, PP3, PP5	Pathogenic
			intron 3	c.319+2T>C	p.?	Splice-site	ClinVar (VCV001048188.4)	1			NA	NA	PVS1, PS3, PM2, PP3, PP5	Pathogenic
7	Homozygous	chr16:88891261	exon 11	c.1156C>T	p.(Arg386Cys)	Missense	dbSNP (rs118204437)	11, 13, 15, 20, 21, 31, 32	1.05	Classic	Hydrophobic core	Conserved	PS3, PM1, PM2, PP3, PP5	Pathogenic
8	Compound heterozygous	chr16:88889099 chr16:88908303	exon 12	c.1262G>A	p.(Gly421Glu)	Missense	ClinVar (RCV001578603.3)	16	0.00	Classic	Surface	Vertebrate-specific	PS3, PM1, PM2, PP3, PP5	Pathogenic
			intron 3	c.319+2T>C	p.?	Splice-site	ClinVar (VCV001048188.4)	1			NA	NA	PVS1, PS3, PM2, PP3, PP5	Pathogenic
9	Homozygous	chr16:88908303	intron 3	c.319+2T>C	p.?	Splice-site	ClinVar (VCV001048188.4)	1	0.00	Classic	NA	NA	PVS1, PS3, PM2, PP3, PP5	Pathogenic
10	Compound heterozygous	chr16:88891261 chr16:88908303	exon 11	c.1156C>T	p.(Arg386Cys)	Missense	dbSNP (rs118204437)	11, 13, 15, 20, 21, 31, 32	0.00	Classic	Hydrophobic core	Conserved	PS3, PM1, PM2, PP3, PP5	Pathogenic
			intron 3	c.319+2T>C	p.?	Splice-site	ClinVar (VCV001048188.4)	1			NA	NA	PVS1, PS3, PM2, PP3, PP5	Pathogenic
11	Homozygous	chr16:88908303	intron 3	c.319+2T>C	p.?	Splice-site	ClinVar (VCV001048188.4)	1	0.00	Classic	NA	NA	PVS1, PS3, PM2, PP3, PP5	Pathogenic
12	Compound heterozygous	chr16:88908303 chr16:88902673	intron 3	c.319+2T>C	p.?	Splice-site	ClinVar (VCV001048188.4)	1	48.10	Attenuated	NA	NA	PVS1, PS3, PM2, PP3, PP5	Pathogenic
			exon 6	c.569A>G	p.(Tyr190Cys)	Missense	dbSNP (rs886042058)	NR			Surface	Partially conserved	PS3, PM1, PM2	Likely Pathogenic

AA, amino acid; CDS, coding sequence; dbSNP, single nucleotide polymorphism database; NR, not reported.

^aDegrees of evolutionary conservation already published or estimated are described in this report.

Fig. 1.

Distribution (x axis) and frequency (y axis) of the pathogenic variants in the GalN6S enzyme using the MutationMapper program. The N-terminal portion of the catalytic domain is shown in green, and the C-terminal portion of the catalytic domain of GalN6S is shown in red.

In terms of allelic frequency, the c.319+2T>C variant was the most frequent (n = 10, 41.66%), followed by p.(Arg386Cys) (n = 8, 33.33%), p.(Gly301Cys) (n = 2, 8.33%), p.(Arg386His) (n = 2, 8.33%), p.(Gly421Glu) (n = 1, 4.16%), and p.(Tyr190Cys) (n = 1, 4, 16%). All variants were interpreted as pathogenic according to ACMG guidelines [Richards et al., 2015].

The c.319+2T>C mutation is located in the splice site of exon 3 and intron 3, affecting the donor site. The in silico prediction resulted in the use of an alternative donor site in intron 3 and the expected addition of 271 nucleotides to exon 3. However, this addition would lead to a premature stop codon (p.A107Gfs* 53), which may cause nonsense-mediated decay of the allele (Fig. 2). This effect can be deduced from the null enzymatic activity present in patients homozygous for this variant (Table 2). This variant was previously described in 2018 [Bochernitsan et al., 2018]. In our study, we found 2 homozygous patients and 6 compound heterozygous patients for this variant. All affected patients come from the extreme south of the country, including the 3 Amerindian patients in this sample.

Fig. 2.

In silico prediction of the c.319+2T>C mutation using the NNSPLICE program suggests the use of an alternative donor site in intron 3 and the expected incorporation of 271 nucleotides into exon 3. The frameshift, however, could lead to a premature stop codon (p.A107Gfs*53) causing nonsense-mediated decay.

The p.(Arg386Cys) missense mutation results in a substitution of arginine for cysteine in exon 11. This mutation is located in the protein’s hydrophobic core [Rivera-Colon et al., 2012] (Fig. 3). Albeit the affected residue has been previously described as conserved [Tomatsu et al., 2006], our alignment revealed a not very conserved residue, diverging with chimpanzee and fruit fly (glutamine and aspartic acid, respectively) (Fig. 4). This substitution was first described in patients in 1995 [Ogawa et al., 1995] and subsequently by various authors [Tomatsu et al., 1997; Bunge et al., 1997; Sukegawa et al., 2000; Tomatsu et al., 2004; Tomatsu et al., 2006]. We found 2 homozygous patients and 4 compound heterozygous patients for this variant in this study, this being the first time it has been described in the Chilean population.

Fig. 3.

Modeling of the amino acids affected by variants (in red) in the bimeric GalN6S enzyme (monomers in gray and light blue). The calcium ion in the enzyme active site in green.

Fig. 4.

Analysis of evolutionary conservation of amino acids affected by variants in 9 species. The affected residues are highlighted with blue rectangles, and the corresponding variant is written on it.

The p.(Gly301Cys) missense mutation results in a substitution of glycine for cysteine in exon 9. This mutation is located in the protein’s hydrophobic core [Rivera-Colon et al., 2012] (Fig. 3). Albeit the affected residue has previously been described as vertebrate specific [Tomatsu et al., 2006], our analysis showed a higher degree of conservation as several species, including fruit fly, presented the same residue (Fig. 4). This variant was first described in patients in 1997 [Kato et al., 1997] in the Colombian population and then detected in subsequent publications [Bunge et al., 1997]. In our study, we found 1 homozygous patient for this variant belonging to the only foreign patient in this sample, from the city of Sevilla Valle, Colombia.

The p.(Arg386His) missense mutation results in a substitution of arginine for histidine in exon 11 and is located in the protein’s hydrophobic core [Rivera-Colon et al., 2012] (Fig. 3) and, according to Tomatsu et al. [2006], would affect a conserved residue, which differs from the previously described alignment (Fig. 4). It was first described in a Chilean patient in 2004 [Tomatsu et al., 2004]. We found 1 homozygous patient for this variant in our study.

The p.(Gly421Glu) missense mutation results in a substitution of glycine for glutamic acid in exon 12. This mutation is located on the protein surface [Rivera-Colon et al., 2012 ] and is also in contact with the other GalN6S monomer (Fig. 3). Regarding the degree of evolutionary conservation of the affected residue, this would be specific to vertebrates, diverging with the fruit fly, where the residue is arginine (Fig. 4). It was reported previously in a heterozygous subject [Morrone et al., 2014]. We found 1 compound heterozygous patient for this variant in our study.

The p.(Tyr190Cys) missense mutation results in a tyrosine substitution for cysteine in exon 6. This mutation is located on the protein surface [Rivera-Colon et al., 2012] (Fig. 3). Regarding the degree of evolutionary conservation of the affected residue, this would be scarcely conserved since the original residue is phenylalanine in rats, mice, and fruit flies (Fig. 4). This mutation is only reported in the dbSNP (rs886042058) and ClinVar (RCV000286677.1) databases and is described as an uncertain significance variant. We found 1 compound heterozygous patient for this variant in our study.

Genotype-Phenotype Correlation

Five variants were associated to the classic phenotype, in order of the following frequency: c.319+2T>C; p.(Arg386Cys); p.(Gly301Cys); p.(Arg386His); and p.(Gly421Glu), whereas the variant p.(Tyr190Cys) was associated to an attenuated phenotype in this sample. When analyzing the effect of the different independent variables on both the phenotype and for the enzymatic activity, the results were not statistically significant in terms of the type of variant and zygosity (p = 1), degree of evolutionary conservation of the amino acids (p = 0.25), and the effect of the variant on the structure of GalN6S (p = 0.20).

Discussion

Twelve MPS IV-A patients were demographically, clinically, and genetically characterized in order to predict whether the type of GALNS variant is associated with the phenotype. The main elements for clinical suspicion were developmental delay predominantly motor associated to skeletal manifestations, specially hyperlordosis and joint hypermobility. Although patient #12 presented an attenuated form of the disease, the remaining patients displayed the classical phenotype. Regarding the age of diagnostic confirmation, most of the patients were confirmed in the school period, with a suspicion and confirmation gap of approximately 2 years and 7 months. The gap for diagnostic confirmation by measurement of enzymatic activity is explained by the access difficulty within the public sector, especially before the enactment in Chile of Law No. 20,850 in 2014, which created a financial protection system financing for high-cost diagnoses and treatments [Medina and Kottow, 2015]. Patient #2 was excluded from this estimation since his confirmation took 28 years and 2 months for the reasons previously stated.

Patients with classic phenotype presented with multisystem involvement but with a variable degree, which explains the great phenotypic heterogeneity. All of these classic patients presented with multiple dysostosis that ranged from moderate to severe. Although the patient with the attenuated phenotype did not develop all the typical skeletal manifestations, he presented with joint hypermobility and bilateral femoral epiphyseal dysplasia. This supports previous studies that stated skeletal dysplasia is present in patients with MPS IV-A, to some extent, seen in all phenotypes [Montano et al., 2003].

Regarding the GALNS gene molecular analysis, more than 400 pathogenic variants have been described until now [Tomatsu et al., 2005; Laradi et al., 2006; Wang et al., 2010; Shams et al., 2017], most of them are single nucleotide. In our series, most variants were missense as previously described in the literature. Within this group, the most frequent variant was p.(Arg386Cys), which is also the most reported variant, accounting for approximately 6% of all variants [Zanetti et al., 2021].

A recently published review [Zanetti et al., 2021] collected, pooled, and analyzed all GALNS gene variants reported to date. This study collected the genotype from 1,190 patients distributed worldwide. Colombia and Brazil were the countries that presented the least allelic heterogeneity. The most common variant in Brazil was p.(Ser341Arg) and could be considered a founder mutation [Bochernitsan et al., 2018]. As presented here, p.(Arg386Cys) was the most widely distributed variant worldwide and specially in Colombia and Brazil. The remaining missense variants p.(Gly301Cys), p.(Arg386His), and p.(Gly421Glu) occur at a lower frequency [Tomatsu et al., 2005; Dung et al., 2013; Olarte-Avellaneda et al., 2014], while p.(Tyr190Cys) has not been previously reported in the literature and is registered in the databases as a variant of uncertain significance.

Finally, the splice-site variant c.319+2T>C was previously reported in one patient from Chile as well [Bochernitsan et al., 2018]. It affects the donor site of intron 3 and would have an impact on the intron splicing. It can be deduced that c.319+2T>C would lead to mRNA degradation and, therefore, null synthesis of the enzyme causing a severe clinical picture, although confirmatory studies such as GALNS mRNA quantification and sequencing (under the influence of some nonsense-mediated decay inhibitor, such as puromycin) and enzyme quantification should be performed to determine the specific mechanism of disease of this variant. However, because patients #9 and #11 are homozygous for this variant and they had undetectable enzymatic activity, associated to a severe phenotype, it is plausible to deduce that this variant is under nonsense-mediated decay. The classic phenotype association with splice-site variants has been previously reported for other variants [Khedhiri et al., 2012; Olarte-Avellaneda et al., 2014]. Interestingly, 8 patients that carry this splice-site variant come from the southernmost regions of our country, and 3 of them are Mapuche. The geographic and ethnic data strongly suggest that this variant would be specific to the extreme south of Chile and could have a founder effect. However, to confirm this hypothesis, it is necessary to carry out haplotype studies on these patients and their relatives, as well as to determine its minor allele frequency at the population level. Depending on the results of these analyses, this could have public health implications, including newborn screening of MPS IV-A targeted to this geographical region.

The missense variants p.(Arg386Cys) and p.(Arg386His) are located in the protein’s hydrophobic core [Rivera-Colon et al., 2012], affecting enzyme folding and therefore its activity, which is reflected in a severe clinical phenotype. The relationship between the classic phenotype and these variants has already been described in the literature [Bunge et al., 1997; Tomatsu et al., 1997; Sukegawa et al., 2000; Tomatsu et al., 2004; Tomatsu et al., 2006]. Although both variants have been reported to affect a conserved residue [Tomatsu et al., 2006], in our study, the alignment revealed a discrepancy, which shows that R386 is only conserved in vertebrates. This finding may be explained by the fact that orthologs GalN6S protein sequences have been updated over time. Futhermore, Tomatsu et al. grouped the conserved and vertebrate-specific residues as “highly conserved,” which may also explain the divergence with our results. The p.(Arg386Cys) mutation was found in 6 patients and the p.(Arg386His) mutation in 1 patient in this study; all of them displayed a classic phenotype. These findings reinforce the available evidence about the relationship between variants that affect highly conserved residues and a more severe clinical presentation [Tomatsu et al., 2006].

The p.(Gly301Cys) missense variant is located in the protein’s hydrophobic core, affecting its folding and therefore its functioning, which causes a severe clinical phenotype. The relationship between the classic phenotype and this variant has already been described in the literature [Bunge et al., 1997; Kato et al., 1997]. Regarding the conservation of the affected residues, it has been reported as specific for vertebrates [Tomatsu et al., 2006]. However, our alignment demonstrates that this variant affects a highly conserved residue. The p.(Gly301Cys) mutation was found in a patient with a classic phenotype consistent with that already described in the literature [Tomatsu et al., 2006].

The p.(Gly421Glu) missense variant is located on the GalN6S surface and may affect the interaction between both monomers, altering the intermolecular contact required for the correct function, which would lead to a severe clinical phenotype [Rivera-Colon et al., 2012]. The relationship between the classic phenotype and this variant has already been described in the literature [Rivera-Colon et al., 2012]. The degree of conservation of the affected residue has not been previously reported. In our study, the alignment of the affected residue showed vertebrate-specific conservation. The p.(Gly421Glu) mutation was found in a patient with a classic phenotype as previously reported [Morrone et al., 2014].

The p.(Tyr190Cys) missense variant has not been previously reported in the literature. According to our analyses, this variant is located superficially in the enzyme, but its side chain is immersed in the hydrophobic core (Fig. 3). Therefore, this residue would be partially superficial or partially located in the hydrophobic core [Rivera-Colon et al., 2012]. Although these authors do not establish a fifth category, the physicochemical properties of these residues are between those located in the hydrophobic core and those in the surface. Regarding the conservation analysis of the affected residue, this revealed discordance in mouse, rat, and fruit fly regarding humans, in whom the corresponding residue is phenylalanine instead of tyrosine. However, since both amino acids have a side chain with a nonpolar aromatic ring, they are chemically very similar, and the residue would therefore be partially conserved. In this study, the p.(Tyr190Cys) variant was found in patient #12, the only one in this sample with an attenuated phenotype. If we consider that the other allele carries the c.319+2T>C splice-site variant that would generate nonsense-mediated decay and no enzyme synthesis, the residual enzymatic activity seen in this patient would be explained by a partial defect caused by this variant. Therefore, this mutation would not have an uncertain significance as previously reported in the dbSNP and ClinVar databases but would be likely pathogenic. However, given the location of this variant and considering the degree of conservation of the affected residue, it would only have a slight effect on GalN6S enzymatic activity, which is consistent with the patient’s phenotype.

Regarding the statistical analysis of the genotype-phenotype correlations, we were not able to find significant associations. However, the degree of evolutionary conservation and the location of variants had lower p values than the other studied variables, for both the phenotype and enzymatic activity. A larger sample size may find significant associations among these variables and the clinical and biochemical phenotype of patients with MPS IV-A. Finally, it is important to highlight that both the current available evidence and the data presented here do not allow a full understanding of how the different variables associated with the genotype influence the GalN6S enzymatic activity and the clinical presentation of patients with MPS IV-A. Although a relationship between the location of the variants in the hydrophobic nucleus and a severe alteration in protein folding is recognized, it is not yet possible to explain how some variants located in the surface have a similar clinical and biochemical impact than those located in the active site. On the other hand, regarding the degree of evolutionary conservation, Tomatsu’s proposal is helpful in some cases but misleading in others, when they grouped both conserved and vertebrate-specific residues as “highly conserved,” giving as reference the analysis of missense variants in patients with hemophilia [Wacey et al., 1994]. A more detailed study of the variants that are located partially on the surface or partially in the hydrophobic core could unveil unknown molecular aspects that would lead to a better genotype-phenotype correlation. Further studies may validate the current correlations among the location, degree of evolutionary conservation and other independent factors of the variants and their effects on the GalN6S enzymatic activity, and the clinical manifestations of MPS IV-A in Chileans patients.

Conclusions

This is the first time that a genotype-phenotype association has been studied by analyzing the variants’ effect on the molecular structure of human GalN6S and the evolutionary conservation degree of affected residues in a cohort of patients in Chile. Regarding the reported variants, all cases with the classic phenotype had variants that affected highly conserved or vertebrate-specific ones, which may lead to folding alteration or protein degradation. Concerning the variant p.(Tyr190Cys), its effect was predicted by a combination of the patient’s enzymatic activity, the genotype, and the use of several bioinformatic tools. These analyses allowed us to reclass the p.(Tyr190Cys) variant as pathogenic according to the ACMG criteria.

Albeit our work could not find statistically significant associations, genotype-phenotype correlations may be inferred, where the evolutionary conservations of affected amino acids and the effect of variants on enzyme structure may play a main role. However, several factors could limit this interpretation such as the subjective classification of the patients (clinically and biochemically) and the sample size. Further analyzes should consider a meta-analysis of published cases with genotype data and larger samples and include other variables that could provide more information, such as the GAG levels in urine and blood before and after the start of ERT. We think that this approach may be replicated in the study of the hereditary neurological diseases, which will contribute to a better understanding of their pathophysiology and the design of new therapeutic strategies.

As compared to the data previously reported in the literature, in this study, the most frequent pathogenic variant detected was c.319+2T>C variant. Interestingly, the patients who carried this variant shared geographic and ethnic features which strongly suggest that this variant would be specific to our country and could have a founder effect.

Statement of Ethics

An informed consent was signed by the adult patients, while the underage patients signed an assent prior to the authorization from their legal guardians. This study was conducted in accordance with the Declaration of Helsinki and was previously reviewed and approved by the Ethics Committee of the Servicio de Salud Metropolitano Central.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

The authors received no financial support for the research, authorship, or publication of this article.

Author Contributions

José Cárdenas-Galli contributed to the draft and submission of the manuscript. Diane Vergara and Victor Faundes contributed to the genetic analysis of the data and provided a critical review of the manuscript. Scarlet Witting contributed as the attending physician for one of the patients. Fernanda Balut contributed as the attending physician for one of the patients. Patricio Guerra contributed as the attending physician for one of the patients. Sebastián Silva contributed as the attending physician for one of the patients. José Tomás Mesa contributed as the attending physician for one of the patients. Javiera Tello contributed as the attending physician for one of the patients. Álvaro Retamales contributed as the attending physician for one of the patients. Andrés Barrios contributed as the attending physician for one of the patients. Fernando Pinto contributed as the attending physician for one of the patients. Lucy Mónica Troncoso contributed as the attending physician for some of the patients and provided a critical review of the manuscript.

Funding Statement

The authors received no financial support for the research, authorship, or publication of this article.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.