Background Mucolipidoses II and III alpha/beta (ML II and ML III) are lysosomal disorders in which the essential mannose 6-phosphate recognition marker is not synthesised on to lysosomal hydrolases and other glycoproteins. The disorders are caused by mutations in GNPTAB, which encodes two of three subunits of the heterohexameric enzyme, N-acetylglucosamine-1-phosphotransferase.
Objectives Clinical, biochemical and molecular findings in 61 probands (63 patients) are presented to provide a broad perspective of these mucolipidoses.
Methods GNPTAB was sequenced in all probands and/or parents. The activity of several lysosomal enzymes was measured in plasma, and GlcNAc-1-phosphotransferase was assayed in leucocytes. Thirty-six patients were studied in detail, allowing extensive clinical data to be abstracted.
Results ML II correlates with near-total absence of phosphotransferase activity resulting from homozygosity or compound heterozygosity for frameshift or nonsense mutations. Craniofacial and orthopaedic manifestations are evident at birth, skeletal findings become more obvious within the first year, and growth is severely impaired. Speech, ambulation and cognitive function are impaired. ML III retains a low level of phosphotransferase activity because of at least one missense or splice site mutation. The phenotype is milder, with minimal delays in milestones, the appearance of facial coarsening by early school age, and slowing of growth after the age of 4 years.
Conclusions Fifty-one pathogenic changes in GNPTAB are presented, including 42 novel mutations. Ample clinical information improves criteria for delineation of ML II and ML III. Phenotype–genotype correlations suggested in more general terms in earlier reports on smaller groups of patients are specified and extended.
Statistics from Altmetric.com
Mucolipidosis II (ML II: MIM 252500), originally called inclusion cell or I-cell disease, is an autosomal recessive lysosomal disorder, clinically evident from birth with progressive course and fatal outcome in childhood.1–3 Mucolipidosis III alpha/beta (ML III: MIM 252600), originally described as pseudo-Hurler polydystrophy, is allelic to ML II and has a closely related pathogenesis.4 5 It is also an autosomal recessive disorder, with onset in early childhood, slowly progressive course and fatal outcome from early adulthood. ML III gamma (MIM 252605) is a non-allelic mucolipidosis which has so far been observed mainly in Middle-East countries.6 7
The term “mucolipidosis” was introduced in 1970 by Spranger and Wiedemann8 to describe several conditions with features of both the mucopolysaccharidoses and the sphingolipidoses. Only ML II and ML III are pathologically similar. ML I, now more descriptively called sialidosis, is due to deficiency of neuraminidase. ML IV is a neurodevelopmental disorder with retinal degeneration and is caused by mutations in the MCOLN1 gene. The activity of the lysosomal hydrolases is normal in ML IV.
ML II, ML III and ML III gamma result from the deficiency of the heterohexameric enzyme, UDP-N-acetylglucosamine–lysosomal hydrolase N-acetylglucosamine-1-phosphotransferase (EC 184.108.40.206), trivially termed UDP-GlcNAc-1-phosphotransferase (GlcNAc-1-PT or GNPT). The enzyme is composed of two α, two β and two γ polypeptides. It catalyses the first step in the biosynthesis of the mannose 6-phosphate (M6P) recognition marker in lysosomal hydrolases and other glycoproteins in the cis-Golgi cisterns.9–11 This modification does not occur in ML II and is deficient in ML III fibroblasts. Without the M6P marker, the lysosomal glycoproteins end up in the culture medium of fibroblasts in vitro, where the catalytic activity is significantly increased. Concomitantly, the intracellular activity of most lysosomal enzymes is considerably reduced in cultured cells.12 13 The former finding correlates directly with the increased specific activity of lysosomal enzymes in plasma and other body fluids of patients with ML II and ML III.13–16
The GNPTAB gene contains 21 exons, spans about 80 kb of cDNA on chromosome 12q23.3 and encodes a 1256-amino acid precursor protein that contains the catalytic domain and contributes also to substrate recognition.6 17–20 Proteolytic cleavage at the Lys928–Asp929 bond yields the mature α and β polypeptides. Homozygous or compound heterozygous mutations in GNPTAB result in either ML II or ML III as amply proven by the reports of 37 different pathogenic mutations.21–28 The GNPTG gene with 11 exons located at 16p13.3 encodes the γ polypeptide with a possible substrate-recognition role in the GNPT hexameric enzyme complex. Mutations in the GNPTG gene cause ML III gamma.6 7 19 29
Here we present 51 mutations in GNPTAB, 42 of them novel, in a cohort of 63 patients (61 probands). The report includes clinical and biochemical findings in addition to the patients' genotypes. The conclusions pertain to improving the criteria for delineation of ML II and ML III and for distinction of relatively rare patients with clinical findings intermediate between the two reference phenotypes. The results confirm, specify and extend the phenotype–genotype correlations suggested in more general terms in earlier reports on smaller groups of patients.
Subjects and methods
Patients and families
This study describes the clinical, biochemical and molecular findings in 61 ML probands (63 patients), the largest cohort reported so far. The majority of patients were from the USA, but probands were of various ethnic backgrounds and from wide ranging geographic areas including the UK (one), Canada (one), Japan (one), Australia (three) and New Zealand (two). Further diversity is represented because of foreign-born parentage (England, Venezuela, Pakistan, India, Mexico and Guyana). GNPTAB was sequenced in 56 probands. Five probands who died before the study are also included, as their GNPTAB genotypes were determined by mutation screening in the parents.
Thirty-four probands (36 patients) were termed the “core” group of subjects, because either they were clinically evaluated by at least one of the authors or detailed review of the clinical records was possible. Of historical interest, one of the two unrelated patients in whom ML II (termed I-cell disease at the time) was originally delineated 40 years ago is included.1 2 The core group of patients consisted of 13 probands (14 patients) with ML II, 14 probands (15 patients) with ML III and seven probands with a phenotype called ML “intermediate”.
Twenty-seven patients were assigned to the “non-core” group of subjects, as more limited clinical information accompanied the samples sent to the Greenwood Genetic Center for laboratory investigations and none was clinically evaluated by the authors.
Clinical data on the core patient group only were used to provide the basis for correlating phenotype and mutant GNPTAB genotype, but findings in the non-core group of patients were most often supportive of the conclusions regarding phenotype–genotype correlations.
The plasma activity of several lysosomal hydrolases including total β-d-N-acetylglucosaminidase (EC 220.127.116.11), β-d-glucuronidase (EC 18.104.22.168), β-d-galactosidase (EC 22.214.171.124) and α-l-fucosidase (EC 126.96.36.199) was measured by the methods of Thomas et al.30 Arylsulfatase A (EC 188.8.131.52) activity in plasma was measured by the method of Baum et al31 with minor modifications.
GlcNAc-1-phosphotransferase in leucocytes
GlcNAc-1-phosphotransferase was assayed in samples of peripheral blood leucocytes by a previously published procedure.26 However, it was observed that GlcNAc-1-phosphotransferase activity was lost significantly during the storage of leucocytes from healthy controls and obligate heterozygotes. Because of this sample deterioration during storage, we used activity data only when control and patient leucocytes were collected around the same time.
Screening of GNPTAB mutation
PCR and sequencing
Individual exons were amplified with primers previously published by Tiede et al.21 After confirmation of successful amplification by gel electrophoresis, the amplicons were purified on QIAquick columns or plates (QIAGEN, Valencia, California, USA). Sequencing was performed with a standard BigDye (Applied Biosystems (ABI), Foster City, California, USA) protocol followed by purification using DyeEx columns or plates (QIAGEN). Products were run on an ABI3730 DNA Sequencer. Finished sequence data were compared with normal controls and the published GNPTAB text sequence using Sequencher software (Gene Codes Corporation, Ann Arbor, Michigan, USA).
Patients with ML II
The ML II core group contains 14 patients (including one twin pair; four males and 10 females). Seven patients have died, the youngest at 3 days of age and the oldest at 8 years 3 months. Four patients have been examined by one of the authors (SC, JL).
Prenatal ultrasound studies documented abnormalities in five of 10 pregnancies, including oligohydramnios (two), intrauterine growth retardation (two), echogenic cardiac foci (two) and short femurs (one). Two patients had preterm deliveries. Anthropometric measurements at birth were below average in most instances, and seven patients were small for gestational age (tables 1 and 2).
All ML II patients had craniofacial and/or skeletal abnormalities noted on the first day of life, but time to diagnosis ranged from 3 weeks to 2.5 years. Cranial size remained proportional to body size. Early dysmorphic features included facial coarseness with depressed nasal bridge and shallow orbits, metopic prominence and thickened alveolar ridges (figure 1). The hypertrophied gums contributed to the notation of narrow, deeply furrowed palates in five newborns. Although the craniofacial features may be suggestive, true craniosynostosis is not a component of ML II. Unfortunately, this descriptive label was assigned to four patients in this group, and two had craniectomies before the recognition of the diagnosis ML II.
Congenital abnormalities requiring orthopaedic attention in infancy included hip dysplasia or dislocation (three patients), scoliosis (one patient), hand contracture (four patients), bowed limbs (five patients, including two with club feet). Three patients were thought to have neonatal rickets or osteogenesis imperfecta. Statural growth failed completely before or by 24 months of age.
All patients with ML II had delayed and deficient neuromotor development. Some probands could never sit unsupported. Only 1/14 patients achieved unaided walking. Nearly all made vocal sounds, but verbal expression remained limited to a few words, poorly and hoarsely pronounced. Cognitive development and receptive language skills fit best into the range of moderate intellectual disability, but formal psychometric testing could not be performed in most patients.
Most patients were poor feeders leading to enteral support (nasogastric or gastric tube) for 7/14 patients. All patients experienced increased morbidity of recurrent respiratory infections. Four had tracheotomies. Mitral and aortic valve thickening was the most common problem detected by echocardiography. Hypertrophy of ventricular walls was a consistent feature in the longer surviving probands. Umbilical and inguinal hernias were noted in five and two patients, respectively. Two males had hypospadias. Five patients had documented corneal clouding, the earliest noted at 5 months of age. Detailed observations on each patient with ML II are available from the authors.
Patients with ML III
The ML III core group contains 15 patients (including one sib pair; five males and 10 females). As of 1 July 2008, patient ages ranged from 10 years to 45 years. Ten of the patients have been examined by one of the authors (SC).
Eight of the 15 pregnancies had prenatal ultrasound studies. None showed fetal anatomical abnormalities. One patient was delivered preterm. All were average size or above for gestational age (tables 1 and 2).
Figure 2 shows manifestations in ML III. The time period from initial skeletal signs of disease to appropriate diagnosis varied from 6 months to 10 years with an average of 5 years. Suspected diagnoses included juvenile rheumatoid arthritis, various mucopolysaccharidoses, Niemann–Pick type B, cerebral palsy and arthrogryposis congenita. Orthopaedic abnormalities were documented at birth in only two patients, both of whom had hip dysplasia. The patient with the latest onset of skeletal symptoms began complaining of joint stiffness and loss of flexibility at 7.5 years. For the others, contractures of the hands were evident between 2.5 and 4 years of age. Decreased range of motion of the shoulders was noted by 4 years of age in four patients. One of the two patients appropriately diagnosed before the onset of symptoms was evaluated for lysosomal disease in infancy because routine placental pathology had revealed abnormal foamy cells. This child was entirely asymptomatic at 7 months of age when the diagnosis was established. The second presymptomatic diagnosis was made in a 2-year-old, 4 years before this child's complaints of hip pain and difficulty with walking. This patient had been initially evaluated concurrently with an affected 6-year-old sibling.
Early developmental milestones are known for 12 of the 15 patients in the ML III core. Mild gross motor delays were noted in seven patients, with walking the most common delayed milestone, achieved at 16 months for one and 17–24 months for the other patients in this group. Although all 15 patients achieved independent ambulation in early childhood, all experienced progressive joint stiffness and pain. By their mid-teens, 11 patients relied on assistive devices (wheelchair, motorised scooter) for longer distances. One proband with atypical spinal cord complications lost independent ambulation at 15 years. Another proband was limited to a few painful steps by 20 years.
In contrast, the oldest patient in this cohort had bilateral hip and knee replacements and maintained independent ambulation far longer. This patient's first hip replacement was performed at age 25 years and revised 19 years later. One of the patients with recognised neonatal hip dislocations underwent bilateral hip replacement at age 15 years. A third patient had knee replacement surgery at age 25 years. Four patients underwent spinal fusion procedures. Carpal tunnel release was performed in 10 of 15 patients, often multiple times. Two patients had inguinal hernia repair and four patients had small, uncomplicated umbilical hernias.
Myringotomy tube placement was performed repeatedly in seven patients for chronic or recurrent otitis media. Recurrent respiratory infections, including pneumonia, occurred in five of 15 patients. The patient with the most severe tracheomalacia required tracheotomy and tracheal stents. Even among patients without frequent respiratory illness, intubations for surgical procedures were often described as complicated, necessitating the use of fibreoptic equipment.
Ophthalmological findings included mild corneal clouding documented on slit lamp examination in five patients, juvenile glaucoma in one, and myopia in six. All but one patient had at least mild cardiac valve thickening, most commonly involving the mitral and secondly the aortic valves. A single patient had aortic valve replacement at age 9 years, but this was probably not entirely related to ML III. This patient's father had required replacement of his bicuspid aortic valve. One patient was treated for congestive heart failure from mid-teens until the time of death at age 25 years.
Ten of the core ML III patients were administered the Kaufman Brief Intelligence Test, Second Edition (K-BIT). The results confirmed normal IQ in this group of patients, with a mean verbal IQ of 85 (range 71–106), non-verbal IQ of 85 (range 71–106) and composite IQ of 82 (range 67–112). Most patients required at least minimal special educational support or classroom modifications. Among the adult patients, high school diplomas were achieved by all but one. Two patients attended college, and one achieved a bachelor's degree. Detailed observations on each ML III patient are available from the authors.
Patients with ML intermediate
In table 1 the clinical findings in seven core group probands are summarised separately because, from the outset, features of both ML II and ML III were noticed. In general these patients with intermediate phenotypes tended to have physical findings similar to, but milder than, ML II and clinical courses more reminiscent of ML III.
ML II and ML III belong to the larger group of metabolic disorders associated with osteochondrodystrophic changes that have been termed dysostosis multiplex.32 The complete absence of GlcNAc-1-phosphotransferase in ML II results in earlier and more severe changes than those seen in ML III, but also in qualitative differences, as is best demonstrated in hand films. The tubular hand bones are significantly shortened and the diaphyses become extremely widened in the period between infancy and age 3 years in the ML II patient (figure 3). The submetaphyseal regions of the metacarpals remain narrow, resulting in a pointed configuration at both ends. In ML III, the tubular hand bones are of normal or near-normal length and more importantly maintain near-normal diaphyseal tubulation (figure 3). Contractures, especially of the more distal small finger joints, and the ensuing claw-like deformity, are mainly due to hardening of tendons and capsular soft connective tissue. The contractures and small-joint stiffening occur much earlier and are more severe in ML II, but also occur consistently in ML III.
Lysosomal enzymes in plasma
Plasma activity of several lysosomal hydrolases in all affected patients was significantly increased over controls. Comparison of the mean value for the patient group versus controls showed a 6.7–60-fold increase in plasma activity (α-l-fucosidase (×6.7), arylsulfatase A (×8.4), β-d-galactosidase (×13.7), total β-d-N-acetylglucosaminidase (×13.9) and β-d-glucuronidase (×60)). Parents of affected patients (obligate carriers) also had higher circulating enzyme activity (1.3–3.5×) than controls.
The mean value for each lysosomal enzyme was compared between the ML II and ML III groups using an unpaired t test, assuming unequal variance, using Stata/SE V10. β-d-Glucuronidase and α-l-fucosidase were the only enzymes that showed significantly different (p = 0.0129 and 0.0140, respectively) activities between these two patient groups.
GlcNAc-1-phosphotransferase in leucocytes (n = 9)
GlcNAc-1-phosphotransferase activity was found to be completely abolished in the single ML II sample, varied between 0.84% and 10% in the control sample in four ML III samples, and was 3% of the control level in the single ML intermediate sample. The specific activity of this enzyme in two obligate heterozygotes was calculated to be 36.2% (ML III carrier) and 71.4% (ML II carrier) of control activity.
Both pathogenic GNPTAB mutations were identified and characterised in 60 of 61 probands by routine sequencing (tables 3 and 4). The single exception is a patient with ML II in whom only one of the two pathogenic mutations has been fully characterised. The second change is a large insertion that may be similar in some fashion to the retrotransposition of an Alu element in GNPTAB, as has been reported by Tappino et al.28
Fifty-one different GNPTAB pathogenic alterations were identified, including 10 missense (MS), 11 nonsense (NS), 22 frameshift (FSh), and 7 splice site (Spl) alterations and the above-mentioned undefined structural rearrangement. Nine mutations that had been previously reported in the literature accounted for almost half (46.7%) of the mutant alleles in this cohort of 61 probands.21–26 28 33 The most common was the FSh mutation, c.3503delTC, occurring in 18 ML II and four ML III patients (table 3). The next most common mutation allele was the c.3335+6T>G splice alteration found in the heterozygous state in 11 ML III probands. The 42 novel mutations, including the incompletely elucidated structural rearrangement in one ML II patient, are shown in table 3 by proband and table 4 by exon. Most (38/42) of the novel mutations were detected in only one or in two alleles. Only three of the novel alterations (c.342delCA, c.1123C>T and c.1399delG) were encountered more often and identified in five, four and eight alleles, respectively. With the exception of two Hispanic patients in whom the frameshift c.616del ACAG was found in the heterozygous state, each recurrent mutation was detected in people of various ethnicities.
Fifteen patients were found to be homozygous for a single pathogenic gene mutation and 46 were compound heterozygotes (table 3). Seven of the homozygotes carried unique mutant alleles, suggesting that parental consanguinity was likely even in cases where it was not recognised.
The majority of patients (29/36) in the core group clearly exhibited the ML II or ML III phenotype (table 2). The ML II phenotype correlates directly with those GNPTAB genotypes that are homozygous or compound heterozygous for two alleles expected to produce no or nearly no gene product (table 3). That these “null” or “amorph” alleles are associated with absence or near-absence of GlcNAc-1-phosphotransferase activity was proven strictly only in one patient with ML II in our study, but shown in a number of similar patients by others.21 25 26
In the core group, 12/13 probands with ML II were either homozygous for FSh mutations inducing a premature stop codon (four patients) or compound heterozygous for either two different FSh mutations (three patients) or a FSh and NS mutation (five patients). The same clinical consequence was obvious in the remaining patient whose genotype was a compound heterozygote with a FSh mutation and a thoroughly rearranged mutant allele. The direct correlation observed is corroborated by similar genotypes in 15/19 ML II patients in the non-core group (NS//FSh, three patients; FSh homozygotes, six patients; FSh//compound heterozygotes, four patients; NS homozygotes, two patients).
The null alleles in this cohort of patients with ML II have affected nine different exons, ranging from exon 2 to exon 19. Hence, mutation type rather than intragenic location has determined the phenotypic outcome. Premature truncation of either the α or β subunit in GlcNAc-1-phosphotransferase is likely to abolish the enzyme activity. This is supported by the homogeneous clinical effect of the null alleles wherever located in the gene.
The presence of at least one hypomorph (MS or Spl) allele in the GNPTAB mutant genotype appears to protect against ML II and results in the clinically milder ML III in all instances—for instance, 15/15 patients in the core group (FSh//Spl, four patients; FSh//MS, one patient; Spl//NS, four patients; MS//NS, three patients; MS//MS, two patients; Spl//MS, one patient) and in 7/7 subjects in the non-core ML III group (FSh//Spl, three patients; FSh//MS, one patient; Spl//NS, one patient; Spl//Spl, one patient; Spl//MS, one patient) (table 3). Spl and MS mutations are termed hypomorph alleles because, instead of total inactivation, they often result in significantly reduced enzyme activity due to the decreased amount of, and/or ineffective, enzyme protein. That the mutant genotype in patients with ML III may allow some residual GlcNAc-1-phosphotransferase activity was confirmed in the leucocytes of four ML III patients, with enzyme activities ranging between 0.84% and an estimated 10%.
Although the clinical variability among patients with ML III is more extensive than in ML II, all ML III patients in the genotypic category with one null and one hypomorph allele, irrespective of the locations within either the α or β subunit, are phenotypically well within the criteria defining this form of mucolipidosis (table 2). Comparison of the phenotypes in two patients with identical FSh/Spl genotypes, but with the FSh mutation of different parental origins, did not reveal any objective difference. A similar conclusion was reached by comparing the phenotypes in two patients with the identical MS//NS genotype, while taking differences in age well into account.
Our data also indicate that patients with ML III and GNPTAB genotypes composed of two hypomorph alleles are not consistently clinically milder than patients with one null and only one hypomorph allele. For example, one such MS/Spl patient with only minor facial changes and mild dysostosis multiplex has, since adolescence, developed severe thoracolumbar scoliosis. Another MS/MS patient, now in her forties, had a late onset and slowly evolving ML III phenotype that, however, required bilateral hip replacement at age 25 years. Painful and crippling hip disease is, at least from adolescence, a major and consistent feature of ML III. A third patient with a milder than average ML III phenotype had compound heterozygosity for a MS mutation in GNPTAB exon 12 and a deletion of three consecutive nucleotides in exon 3 resulting only in the loss of Val77 or Val78.
A minority (7/36) of patients were found to have an intermediate phenotype, even when taking into account the inherent clinical variability within either reference phenotype. Findings in the seven patients so classified are given separately in table 1. They tended to have earlier appearance of craniofacial and skeletal signs than patients with ML III (first signs before age 3 years), but experienced a clinical course in terms of developmental progress, ambulation, speech and survival quite similar to patients with ML III. With one exception, these patients have at least one MS mutation, either coupled with a FSh (four patients), a Spl (one patient), or another MS mutation (one patient) (table 3). The exception is a patient who is homozygous for a FSh mutation in exon 4. This specific 2 bp deletion in exon 4 has subsequently been identified in additional patients, also homozygous. The milder than expected phenotype is remarkably consistent. All other patients with two FSh alleles have fallen into the ML II phenotype.
The mutation spectrum reported here is consistent with the 37 mutations found in eight previous studies.21–28 34 35 The previously reported alterations are evenly spread across the GNPTAB gene, with frameshift-causing mutations being the most common (16/37 = 43%), as in our study (22/51 = 43%). The mutation encountered most often, both in the literature data and in this study (c.3503delTC), has been reported as the single causal mutation in the GNPTAB gene in a French-Canadian founder population.33 In a recent report, Otomo et al36 analysed GNPTAB in 40 Japanese patients. The most common mutation in that population (c.3565C>T, R1189X) was detected in four of our probands: two US Caucasians, one Hispanic and one New Zealander. Of the 14 mutations reported as novel by Otomo et al, we report four similar mutations at the same location or within 1 bp, with similar effect, including one MS. We suspect mutations will become less ethnically specific as GNPTAB analysis is transitioned into the diagnostic setting, with the possible exception of the exon 2 duplication reported in six Japanese patients.36
The issue of clinical significance must be considered in the 10 cases with MS mutations. To address their significance, we have run each MS alteration through the basic prediction software programs PolyPhen-Polymorphism Phenotyping (http://genetics.bwh.harvard.edu/pph/) and SIFT-Sorting Intolerant from Tolerant (http://blocks.fhcrc.org/sift/SIFT_BLink_submit.html). Four of the alterations (c.1402T>A→p.C468S, c.1514G>A→p.C505Y, c.2867A>G→p.H956R and c.3053A>G→p.D1018G) were classified as probably damaging and not tolerated, which is consistent with a high likelihood of pathogenicity. Five alterations (c.10C>A→p.K4Q, c.44C>A→p.S15Y, c.569A>T→p.D190V, c.1001G>A→p.R334Q and c.1196C>T→p.S399F) were predicted to be possibly damaging and poorly tolerated by both PolyPhen and SIFT. Alterations p.K4Q, p.D190V and p.S399F have been reported previously in patients with ML.25 26 Only the c.1042A>C→p.I348L alteration was predicted to be a benign and tolerated change. Here it was detected in a non-core patient with ML II in combination with another clearly pathogenic FSh alteration. On the basis of the fact that it does not block phenotypic expression of the latter (table 3), p.I348L has been accepted as being detrimental.
There is ample clinical, radiographic and pathology-based evidence that ML II and ML III are systemic disorders mainly of connective tissue.32 37–39 In ML II, the intensive egress of structurally abnormal “lysosomal” glycoproteins into the extracellular matrix (ECM) correlates not only with early cessation of endochondral ossification and linear growth, but also with apparently highly overactive intramembranous ossification leading to extreme diaphyseal widening in the metacarpals (figure 3) probably caused by abnormal signalling effects triggered in the damaged connective tissue. The transient phenomenon of periosteal cloaking around the diaphyses of long bones in infants and the periarticular punctuate calcifications in some neonates and even in third-trimester fetuses with ML II38 40 may be the result of this ECM-generated mechanism. Similar findings in the hand bones of patients with geleophysic dysplasia with ADAMTSL2 mutations have been ascribed to dysregulation of transforming growth factor (TGF)-β signalling.41 Dysregulated signalling of TGF-β is triggered by either abnormal fibrillin-1, the major constituent of extracellular microfibrils, or mutant TGF-β receptors in Marfan syndrome and Loeys–Dietz syndrome, respectively.42 Apparently the progressive features in the ECM are non-specific and triggered by various abnormalities in connective tissue, as they are also observed in patients with some of the glycosaminoglycan or oligosaccharide storage disorders.
Murine models of ML II and III do not have cytoplasmic inclusions in fibroblasts or mesenchymal cells and do not have prominent skeletal and connective tissue abnormalities.43 Feline ML II is spontaneously occurring, displays inclusion bodies in cultured fibroblasts, and shows skeletal and joint abnormalities, making the feline a better but still not ideal model that needs further study.44 Studies of animal models of mucopolysaccharidosis VI suggest that proinflammatory cytokines alter expression of several matrix metalloproteinases, enzymes involved in ECM degradation.45 In ML III, a variable fraction of lysosomal glycoproteins still acquires the M6P recognition marker and is correctly routed to, and functional within, lysosomes. It is apparently sufficient to maintain some statural growth and to postpone clinical expression of the abnormal ECM signalling. Diaphyseal widening in the small hand bones, if present at all, is much milder in ML III. The abnormal signalling is, however, the likely cause of progressive stiffening of weak connective tissue, of the osteopenic osteochondrodysplasia in the hips and spine overt in the longer ML II survivors and in ML III patients.
In conclusion, the clinical work has confirmed the diagnosis of either ML II or ML III in 29/36 probands in the core group and in the large majority of the non-core patients. In only seven probands were both clinical features and course of the disorder found to straddle the boundaries of variability set for delineation of the two reference phenotypes. Their disorder was termed ML intermediate.
ML II, recognisable at birth, often causes intrauterine growth impairment and sometimes the prenatal “Pacman” dysplasia.40 The main postnatal manifestations of ML II include gradual coarsening of neonatally evident craniofacial features, early cessation of statural growth and neuromotor development, dysostosis multiplex, and major morbidity by hardening of soft connective tissue about the joints and in the cardiac valves. Fatal outcome occurs often before or in early childhood. ML III, with clinical onset rarely detectable before 3 years of age, progresses slowly with gradual coarsening of the facial features, growth deficiency, dysostosis multiplex, restriction of movement in all joints before or from adolescence, and painful gait impairment by prominent hip disease. Cognitive handicap remains minor or absent even in the adult, often wheelchair-bound patient, with variable though significantly reduced life expectancy.
The phenotypic spectrum shown to be more dichotomous than continuously variable has been directly correlated with the GNPTAB mutant genotypes detected. Patients with ML II were found to be either homozygotes or compound heterozygous for NS or FSh mutant alleles, expected to produce no or nearly no functional GlcNAc-1-phosphotransferase and hence complete failure to sort lysosomal enzymes devoid of the necessary M6P recognition marker to the lysosomal compartment. ML III represents the clinical outcome of GNPTAB homozygous or compound heterozygous mutant genotypes comprising at least one hypomorph (MS or Spl) allele and resulting in up to 10% of residual GlcNAc-1-phosphotransferase activity. The milder more chronic course underscores the important role of GlcNAc-1-phosphotransferase in physical growth and in quality and maintenance of connective tissue.
The intermediate phenotypes found in only a few subjects in the patient cohort confirm the overall direct genotype–phenotype correlation. More detailed clinical studies in progress indicate that clinical subgroups may be recognisable within the group of patients with intermediate ML and provide information on the effect of MS mutations in specific locations of the GNPTAB gene.
The diagnosis of ML II or ML III is established in the metabolic laboratory for patients with characteristic physical and radiographic features. Urine screens support the diagnosis. Urinary excretion of glycosaminoglycans is normal while oligosaccharide excretion may be excessive. Assay of multiple plasma acid hydrolases shows 5–20-fold increased activity compared with controls. Specific activities of lysosomal hydrolases in peripheral leucocytes are normal. GNPTAB mutation screening provides solid confirmation of the diagnosis of either ML II or ML III and informs valuably about the patient's prognosis. Assay of GlcNAc-1-phosphotransferase activity is not routinely available and the finding of highly hyperactive acid hydrolases in the plasma does not differentiate ML II from ML III, further contributing to the utility of molecular analysis in these diseases.34 35 In addition, early knowledge of the GNPTAB genotype should favour gathering clinical information more prospectively and may influence the development of successful treatments.
We thank the following: the patients and their families who participated in this study; the physicians and counsellors who referred patients and provided records; David Sillence, Paul Orchard, Virginia Proud, Omar Abdul-Rahman, John Belmont, Brendan Lee, Thaddeus Kurczynski, Andrea Atherton, Stephen Romansky; Jonathan Rollins for statistical analysis; Jurgen Spranger for consideration and advice; and the clinical geneticists, counsellors and nurses who evaluated the families at the Greenwood Genetic Center. Support for patient travel and evaluation was provided, in part, by ISMRD, the International Advocate for Glycoprotein Storage Diseases, the National MPS Society, and the Genetics Endowment of South Carolina.
An additional table is published online at http://jmg.bmj.com/content/vol47/issue1
Competing interests None.
Patient consent Obtained.
Provenance and Peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.