Main

Mucopolysaccharidosis type I (MPS I) is an autosomal recessive genetic disorder that is caused by the deficiency or absence of α-l-iduronidase, a lysosomal enzyme involved in the degradation of the glycosaminoglycans heparan sulfate and dermatan sulfate. Impaired degradation of these glycosaminoglycans leads to a wide range of clinical manifestations, including hepatosplenomegaly, dysostosis multiplex, coarse facial features, severe arthropathy, hearing loss, visual impairment, restrictive lung disease, upper airway obstruction, valvular heart disease, communicating hydrocephalus, and spinal cord compression. Progressive mental retardation develops in patients at the most severe end of the disease spectrum. Historically, MPS I has been classified into 3 distinct phenotypes, severe (Hurler), mild (Scheie), and intermediate (Hurler/Scheie), but in reality, these 3 phenotypes merely represent different points on a continuous spectrum of disease severity. The 3 phenotypes are usually defined as (1) severe, when onset of symptoms are before 12 months of age, survival is < 10 years, and mental retardation manifests before the age of 3 years; (2) mild, when onset of symptoms are after 5 years of age, survival is normal, and mental retardation is absent; and (3) intermediate, when onset of symptoms is between 1 and 6 years, survival is variable, and mental retardation is absent or mild but not present before 3 years of age.1

Biochemical and immunological techniques have been only partially successful in predicting clinical severity. Although it is reasonable to assume that there must be some residual enzyme activity in milder phenotypes, it is very difficult to demonstrate such residual activity in cultured cells.2,3 Isolation and characterization of the human α-l-iduronidase gene has made it possible to identify primary disease-causing mutations and to attempt genotype-phenotype correlations.

The α-l-iduronidase gene (IDUA) is situated on chromosome 4p16.3 and contains 14 exons.4 To date, approximately 100 mutations, including missense, nonsense, splice site, deletions, and insertions, have been identified throughout the gene. The 3 phenotypes appear to be caused primarily by different combinations of mutant alleles at the IDUA gene locus. Severe patients are predicted to have mutations on both alleles that prevent the production of any functional enzyme, whereas mild patients are predicted to have at least one allele that allows for some residual enzyme activity. Patients with the intermediate phenotype have at least one allele that allows for a trace amount of residual enzyme activity to prevent severe disease.5 As an additional level of complexity, 30 nonpathogenic alleles have been reported that may influence the levels of α-l-iduronidase activity in normal individuals and may be responsible for the variable disease severity reported in some patients with the same pair of disease-causing mutations.4,5

METHODS

A search of the medical literature was performed in the biomedical database MEDLINE for the years 1970 to 2002. Search terms used to identify citations that provided information on patients’ genotypes and resultant phenotypes included the subject headings “mucopolysaccharidosis I/genetics,” “DNA mutational analysis,” and “mutation.” The search identified 30 relevant citations.1–30

The online Human Gene Mutation Database (http://archive.uwcm.ac.uk/uwcm/mg/hgmd0.html) was also searched for MPS I mutations, and one additional mutation was found that was not cited in MEDLINE.31

Every effort was made to prevent counting patients more than one time. This was accomplished by noting the authors and their affiliations in addition to the patient’s ethnicity, date of birth, and sex, where indicated. This information was cross-referenced among all the citations and consideration was taken to avoid duplications. Several articles specifically mentioned which patients’ genotype/phenotype correlations had been previously published.

Search terms used to identify reports of allele frequency included the subject headings “mucopolysaccharidosis I” and “gene frequency.” This literature search also included any mention of the words “common,” “commonly,” “frequent,” “frequently,” or “frequency” if present within 3 words of “mutation” or “mutations.” The search identified 11 relevant citations.4,5,7,9,10,12–14,16,18,20

RESULTS

This review identified 18 nonsense, 45 missense, 9 splice site (intron), and 25 insertion/deletion mutations in MPS I patients (Tables 1–4). Most of these mutations are associated with the severe phenotype.

Table 1 Identified homozygous and heterozygous nonsense mutationsa
Full size table
Table 4 Identified insertion and deletion mutations
Full size table

Nonsense mutations

All nonsense mutations reported to date are believed to result in the total lack of enzyme activity (null mutation). Accordingly, patients who are homozygous for a particular nonsense allele or heterozygous for two different nonsense alleles have presented with the severe phenotype (Table 1).

Patients who are heterozygous for a nonsense allele and another type of mutation (missense, splice site, or insertion/deletion) have presented with a wide range of clinical phenotypes depending on the severity of the second allele. If a patient presented with the severe phenotype and had one allele that was a nonsense mutation, then the other allele was deemed also to be severe. Likewise, if a patient presented with a less severe phenotype and had one nonsense allele, then the second allele was assumed to allow for some residual enzyme function.

Missense mutations

Although the substitution of a single amino acid can severely impair enzyme function, missense mutations are most likely to be compatible with some residual IDUA activity.4 The majority of patients with mild or intermediate phenotypes have had at least one missense mutation. R89Q is the most common mild mutation, and in general, patients who have a least one allele with R89Q will present with the mild form of the disease. However, there are many reports of missense mutations in patients with severe disease. Table 2 lists known missense mutations along with the identified second allele and resulting phenotype. The missense mutation, P533R, is unusual in that it has been identified in the homozygous state in patients with mild, intermediate, intermediate-severe, and severe phenotypes. Although there is no explanation for this wide variability, MPS I disease progression appears to be less severe in P533R/W402X heterozygotes and P533R homozygotes compared to W402X or Q70X homozygotes or W402X/Q70X heterozygotes.27

Table 2 Identified missense mutations
Full size table

Splice-site mutations

Most splice-site mutations profoundly affect normal splicing, leading to a very unstable mRNA and thus, the severe phenotype when associated with a second null mutation.5,8,20,22,25 The exception is the 678-7g>a (IVS 5-7g>a) mutation, which produces a small amount of normal IDUA mRNA and protein that prevents significant storage of substrate. This mutation has been identified in association with a null mutation in several mild patients.5,25 Table 3 lists known splice site mutations with the identified second allele and resulting phenotype.

Table 3 Identified splice site mutations
Full size table

Deletions/insertions

Twenty-three small deletions or insertions have been detected, with most causing severe MPS I. Only two insertions (396insAC, 974ins12) and one deletion (deltaD444/445) have been identified in mild patients in association with a severe allele.17 Table 4 lists known insertion and deletion mutations with the identified second allele and resulting phenotype.

Mutation frequency

Overall, the most commonly reported mutations have been the nonsense mutations, W402X and Q70X, although their frequencies vary in MPS I patients from different ethnic backgrounds. For instance, the W402X allele is frequent (45%–60%) in North America, Australia, Spain, and the United Kingdom, but is less common in Scandinavia (17%) and Italy (11%).1,4,5,20 On the other hand, Q70X has a much higher frequency in Scandinavia (62%) versus other European countries (16%), North America (17%), and Australia (17%).4,5,17,18 In Japan, no patient has been shown to carry either the W402X or Q70X mutation, whereas the R89Q and 704ins5 mutations are common.16 There is a high frequency of the homozygous P533R mutation in Moroccan patients, most likely as a result of the high occurrence of consanguineous unions.10 Table 5 includes reports from the literature of the frequency of mutations in different parts of the world.

Table 5 Frequency of mutations (%) in different parts of the world
Full size table

Polymorphisms and other nonpathogenic alleles

Polymorphisms, defined as benign genetic variations present in > 1% of the normal (unaffected) population as well as other rare nonpathogenic alleles, have been described in the IDUA gene on the same allele (cis) as a known disease causing mutation. To date, 30 nonpathogenic sequence variants have been detected in the IDUA gene.5 It is not known what effect any of these sequence variants may have on the stability and processing of the IDUA mRNA, or the activity, stability, or transport of α-l-iduronidase. However, these polymorphisms likely contribute to the variability in α-l-iduronidase activity seen in healthy individuals, and it is also likely that they modify the severity of MPS I disease when present in combination with known MPS I mutations.21 For example, the A361T polymorphism, which was present in a patient with the mild allele R89Q, was thought to potentiate the deleterious effect of the R89Q mutation by decreasing the activity of the mutant α-l-iduronidase, thus altering the clinical phenotype from mild to intermediate.22 A second example is of a severe patient with 3 IDUA sequence changes: 974ins12, P496R, and the polymorphism A591T.5 As stated, the combination of the 974ins12 mutation with null mutations has been reported in patients with the mild to intermediate phenotype, implying that 974ins12 mutation allows for some residual enzyme activity. The severe phenotype in this patient was suggested to be caused by the A591T polymorphism acting in cis with the 974ins12 mutation.5

Mutation location

Attempts to predict the clinical phenotype based on the type of amino acid change or the location of the molecular lesion have been largely unsuccessful. Two mutations, R89Q and R89W, are associated with a mild phenotype, suggesting that the arginine (R) at position 89 is not vital for enzyme function.3 These mutations are in close proximity to H82P and A75T, which have resulted in severe phenotypes. On the other hand, the missense mutation A75P (with Q70X) has been reported in patients with a mild phenotype, suggesting that the replacement of alanine (A) by proline (P), rather than threonine (T), allows some functional enzyme to be produced.5

Another example is the close proximity of the 3 mild to intermediate mutations L490P, R492P, and P496L, which suggests that they lie in a location that can tolerate a nonconservative substitution to or from proline (P). However, the mutations, R489P and P496R, have been associated with the severe phenotype.2,5 Therefore, it appears that the substitution of a proline (P) for an arginine (R) at codon 496 causes greater functional disruption than replacement by a leucine (L) residue.

A final example involves the missense mutations, X654G and X654C, both of which alter the normal stop codon at the end of the coding sequence and result in a frameshift. X654G, when associated with a null mutation, results in an intermediate phenotype, whereas X654C, when present in the homozygous state, results in the severe phenotype.2,4

DISCUSSION

One of the main purposes of mutation identification is to be able to establish genotype-phenotype correlations that will allow for a patient’s clinical phenotype to be predicted from the genotype. Currently, we can only predict that nonsense mutations will invariably cause severe MPS I disease, if present on both IDUA alleles. The clinical consequences of amino acid substitutions or of other mutations that are not clearly null can be predicted only by looking at the phenotype of the patients in which these mutations have previously been identified. Even then, predictions must be made with caution because the same mutation may lead to varying severity due to the combination of IDUA alleles, attenuating polymorphisms, and other rare sequence variants, genetic background, and environmental effects.20 In addition, many patients have at least one private mutation making phenotype prediction difficult. Collection of genotype and phenotype data in a central registry will be helpful, and toward this end, Genzyme Corporation has established an MPS I Disease Registry (http://www.MPSIRegistry.com).

At this time, determination of genotype of MPS I patients is not a routine procedure, and if performed, is usually limited to a panel of nine recurrent mutations (W402X, Q70X, A327P, L218P, 474-2a>g, R89Q, P533R, A75T, and 678-7g>a). Future techniques need to be pursued to enable prediction of disease progression in patients with unique, rare, partially defined, and unknown genotypes. By further examining the stability of the IDUA mRNA and protein, better insight into the effect of a mutation on enzyme activity may be gained. A direct correlation between residual α-l-iduronidase activity in cultured MPS I fibroblasts and phenotypic severity appears promising, but the methodology is sophisticated and has been performed on only a small number of patients on a research basis.3

The accurate prediction of genotype-phenotype correlations in MPS I has significant implications given the wide spectrum of disease severity possible and the choice of treatment options. The only two specific treatments, bone marrow transplantation32–35 and the recently approved enzyme replacement therapy, recombinant human α-l-iduronidase (Aldurazyme [laronidase]),36–38 have very different risk-benefit profiles. Bone marrow transplantation has been shown to improve many of the somatic symptoms and stabilize cognitive functioning in patients with severe disease.32–35 However, because of its associated high morbidity and mortality, bone marrow transplantation is generally reserved for MPS I patients with severe disease who are under 2 years of age and have preserved cognitive function. Aldurazyme has demonstrated efficacy in improving the noncentral nervous system features of MPS I disease, specifically pulmonary function and walking capacity, but it is not expected to cross the blood-brain barrier and have an impact on cognitive function. Combination bone marrow transplantation and enzyme replacement therapy, particularly in young patients, is a third treatment option that warrants further clinical investigation. With the prospects of newborn screening for MPS I in the future, it will become all the more important to have the best predictive information available for families and caregivers to make the most rational choice for treatment.