Main

Brachytelephalangic chondrodysplasia punctata (BCP) describes a group of heterogeneous conditions with overlapping phenotypes. X-linked recessive BCP (CDPX1) (OMIM no. 302950), originally recognized by Sheffield et al.,1 is a panethnic congenital rare disorder that affects males. The most characteristic clinical features are as follows: 2,3 (i) Chondrodysplasia punctata, or stippled epiphyses, observed on X-ray. These minimally involve the ankle and distal phalanges but can also include long bones, vertebrae, hips, costochondral junctions, hyoid bone, and tracheal and bronchial cartilage. Chondrodysplasia punctata can be observed initially in the second trimester of pregnancy by ultrasound but tends to improve or disappear by age 2–3 years. (ii) Brachytelephalangy, or shortening of the distal phalanges. A characteristic finding on X-ray is an inverted triangle appearance of the distal phalanges, with punctata on either side. (iii) Nasomaxillary hypoplasia, referring to absence or hypoplasia of the anterior nasal spine, causing a flat nasal base, reduced nasal tip protrusion, shortened columella, and vertical grooves within the alae nasi. (iv) Proportionate short stature. Another commonly observed characteristic is mixed conductive and sensorineural hearing loss. Most children have normal intellect and life span, but comorbidities can be present, including compression of the cervical spinal cord associated with cervical vertebral abnormalities or stenosis of the upper and lower airways that result from extensive calcifications within the tracheal and bronchial cartilage.3,4 Intrafamilial differences in disease severity have also been documented.3

The only known genetic cause of CDPX1 is defects in arylsulfatase E (ARSE). The ARSE gene is located at Xp22.3, within a cluster of contiguous arylsulfatase genes that share high sequence homology, escape X inactivation, and have pseudogenes on the Y chromosome.5 There are 17 sulfatases in the human genome; all share extensive sequence homology and contain a highly conserved cysteine that undergoes a unique posttranslational modification essential for their catalytic activity.6,7,8 ARSE is localized to the Golgi membranes,9 and its transcript has been identified in multiple tissues.5 Its protein product is a 589-amino-acid and 60 kD precursor, which is subject to N-glycosylation at four potential sites (asparagine residues 58, 125, 258, and 344) to give a mature 68 kD form. Aside from the conserved active site domain, there are two predicted transmembrane domains, through which it is likely to be anchored to the Golgi cisternae. Its physiological substrate is unknown. Enzymatic assays to determine endogenous ARSE activity in fibroblasts have not been successful thus far because of interference with other highly homologous arylsulfatases.9 The ability of ARSE to hydrolyze the fluorogenic 4-methylumbelliferyl (4MU) sulfate permits the use of this artificial substrate to measure its activity in an in vitro system.5 Studies showed that the enzyme activity is optimum at neutral pH.5,9

By 2008, more than 140 male patients with a CDPX1 phenotype had been reported, and of these, around one-fourth were associated with Xp deletions or chromosomal rearrangements. Only 67 male probands had undergone ARSE mutation analysis, with mutations identified in 31.3 Functional analysis was performed in 8 of 15 reported missense alleles (Supplementary Table S1 online) by expressing the mutant ARSE complementary DNA in mammalian COS1 cells and measuring ARSE activity using 4MU sulfate.5 In these experiments, all missense alleles had negligible activity.9,10

Considering the phenotypic resemblance of warfarin embryopathy in early gestation to BCP, Franco et al.5 (1995) demonstrated that ARSE activity is inhibited in the presence of warfarin, an anticoagulant drug that decreases amounts of active vitamin K. These findings suggested that cases of BCP without ARSE mutations could be caused by inhibition of a normal ARSE enzyme in fetal development by reducing levels of vitamin K. Because placental transport of vitamin K is normally reduced,11 vitamin K deficiency in the mother would be expected to result in greater vitamin K deficiency in the fetus. This hypothesis is attractive given that there have been several reports of BCP in offspring of both sexes in which gestational vitamin K deficiency was suspected, including mothers with severe hyperemesis gravidarum,12 small-intestinal obstruction,13 small-bowel syndrome,14 pancreatitis,15 or biliary lithiasis.16 In those cases in which ARSE mutations were evaluated in their affected offspring, none were found.3,13 Another group of children with BCP are born to mothers with autoimmune diseases.17,18,19 It was proposed that maternal autoantibodies might perturb fetal vitamin K metabolism,20 thereby reducing the etiology of all phenocopies to disturbances in vitamin K metabolism at a critical fetal time period. Although this hypothesis remains to be proven, it was not clear how many patients with BCP might be phenocopies due to maternal vitamin K deficiency states.

In 2008, through the Collaboration Education and Test Translation program (CETT),21 sponsored by the National Institutes of Health Office of Rare Diseases Research, we established a pilot project to identify patients with CDPX1 and its phenocopies. The collaborative group included the clinical laboratory, which performed ARSE gene sequencing in patients referred with CDPX1 phenotypes and collected clinical information from the referring physician using a clinical data sheet designed for the CETT program for CDPX1 (Supplementary Table S2 online). The research laboratory reviewed the clinical data sheets to evaluate genotype–phenotype correlations and performed functional analysis of the novel missense alleles identified. Because the majority of ARSE mutations identified thus far have been unique, functional analysis was important to demonstrate pathogenicity. This was also a unique opportunity to prospectively determine the incidence of CDPX1 phenocopies.

Materials and Methods

CDPX1 phenotype

CDPX1 phenotype was defined as male sex, nasomaxillary hypoplasia, brachytelephalangy, and punctate calcifications on X-ray in children younger than 3 years.

Clinical data collection

The data sheet was designed to sample all the major features described in patients with BCP reported in the literature since 1978. In addition, it contained broad questions to identify other etiologies, including maternal vitamin K deficiency and autoimmune disease, that have been associated with BCP. The final data sheet was approved by the program staff and review board of the CETT program. Informed consent was obtained for the collection of clinical data, along with the consent for ARSE gene sequencing, according to the CETT program requirements. Referring physicians were asked to complete this form (shown in Supplementary Table S2 online) at the time of sample submission. The clinical data sheets were deidentified and then sent to the research laboratory for review. See Table 1 for the breakdown of patients involved in this study.

Table 1 Breakdown of patients and data sheets collected in this project
Full size table

Literature review

We searched PubMed from 1995 to 2010 to identify ARSE mutations in which functional analysis had not been done. Search terms used were CDPX1, ARSE, X-linked recessive CDP, and BCP.

Molecular analysis

All 11 exons and intronic flanking sequences of the ARSE gene were Sanger sequenced bidirectionally. In females, targeted array comparative genomic hybridization with exon-level resolution was performed concurrently with sequencing. Targeted mutation analysis was performed in affected male relatives of probands with ARSE mutations and in mothers and other female relatives to evaluate carrier status.

Mutant complementary DNA constructs

Wild-type ARSE complementary DNA (wtARSE) was engineered in the mammalian expression vector, pALTER-MAX (Promega, Madison, WI) by inserting ARSE exon 2, obtained by RT-PCR from a liver RNA library, into the ARSE complementary DNA clone (GenBank: AA887688.1). It contained the open reading frame, bps 1–1856, including 86 nucleotides of the 3′ untranslated region. wtARSE diverges from the reference sequence (NM_000047.2) by having the synonymous substitutions c.4T→C and c.1771A→G, and the polymorphisms reported in dbSNP: c.495 T→C, c.1270 G→A, and c.1692 C→T. Site-directed mutagenesis of wtARSE was performed using the QuikChange II Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA) and resequenced after mutagenesis for confirmation. Primers used to generate ARSE mutations are listed in Supplementary Table S3 online.

COS cell transfections

COS1 cells were maintained in Dulbecco’s Modified Eagle Medium supplemented with 10% fetal bovine serum at 37 °C and 5% CO2. Overnight cultures were seeded at 10 × 105 cells per 21 cm2 culture plate and cotransfected with 5 μg of ARSE and 2.5 μg of Renilla reniformis luciferase (Promega) plasmids using 15 μl of FuGENE HD (Roche Diagnostics, Mannheim, Germany) according to manufacturer’s instructions. Transfections were performed in triplicate. Cells were harvested 48 h after transfection, washed with phosphate-buffered saline, trypsinized, and resuspended in Dulbecco’s Modified Eagle Medium with 10% fetal bovine serum, centrifuged at 1,000 rpm for 4 min and resuspended in 200 μl of lysis buffer: 0.1 mmol/l Tris HCl, 0.15 mol/l NaCl, and 1% Triton X-100 at pH 7.5. The lysate was incubated on ice for 20 min, then centrifuged at 20,800 rcf × 20 min at 4 °C. Supernatant was collected and protein concentration was measured by Bradford assay. For time course experiments, 0.5 μg of Renilla plasmid was used and lysis buffer was 0.05 mol/l Tris HCI at pH 7.5.

ARSE assay

The reported assay5,12 was modified as follows: incubation mixture (100 μl) contained 20 μg of lysate and 0.2 mmol/l 4MU sulfate (Sigma-Aldrich, Saint Louis, MO) in 0.05 mol/l Tris HCI buffer (pH 7.5). Incubation was performed for 3 h at 37 °C in a 96-well plate; the reaction was stopped with 1.8 ml of glycine–carbonate buffer (pH 10.7). Fluorescence was determined (Wallac1420 Multilabel Counter, software version 3.00; PerkinElmer, Waltham, MA) at 365 nm (excitation) and 460 nm (emission). Each sample was run in triplicate and the experiment was repeated three times. Time course experiments on selected mutations were stopped at 0 h, 30 min, 1 h, 2 h, and 2 h 30 min. Renilla luminescence was measured by the Dual-Luciferase Assay and a GloMax 96 Microplate Luminometer (Promega).

Results

Clinical and molecular analysis

Twenty nine male probands were referred prospectively through the CETT program, and clinical data sheets were collected on 19. Three additional male probands with ARSE missense alleles were contributed by the Molecular Genetics Laboratory of the Royal Devon and Exeter National Health Service Foundation Trust, and clinical data sheets were collected on two of them. From these 21 data sheets, the frequency of relevant clinical features in ARSE mutation-positive and mutation-negative probands was collated in Supplementary Table S4a,b online, respectively.

ARSE mutations were identified in 17/29 male probands from the CETT project and included four complete gene deletions, 1 single-codon deletion, and 12 missense mutations. The mutations identified are shown in Table 2 . All missense alleles were predicted to be damaging by either PolyPhen22 or SIFT23 analysis, except for T306A, which was predicted to be benign. None of these mutations were polymorphisms identified in dbSNP (V132).24 All residues affected by the mutations were evolutionarily conserved among all ARSE proteins and human arylsulfatate C (ARSC).

Table 2 ARSE mutations detected in this study
Full size table

In addition, four females with clinical features of BCP were referred for ARSE mutation analysis. No mutations were found, and clinical data were obtained in only one patient. She had nasomaxillary hypoplasia, flattened nose, anteverted nostrils, brachytelephalangy, ichthyosis, delayed motor, and cognitive development. At the age of 11 months, there was no evidence of chondrodysplasia punctata, and her karyotype was normal. The distribution of all patients included in this project is shown in Table 1 .

ARSE functional analysis

We performed functional analysis on the 12 novel missense alleles, the single codon deletion identified in our cohort, and 7 novel missense alleles reported previously in which functional analysis had not been done3,5,25,26 ( Table 2 ). These 20 ARSE alleles were engineered by site-directed mutagenesis and coexpressed in COS1 cells with Renilla luciferase to control for transfection efficiency. Whole-cell lysates were incubated with 4MU sulfate for 2 h, and ARSE activity was measured by detection of 4MU product. We found that all the mutant alleles had negligible ARSE activity, suggesting that they were pathological ( Figure 1 ). However, we observed minimally increased fluorescence in some alleles, especially those located in the C-terminal region, suggesting potential residual activity. To determine if residual activity was present, time course experiments were performed for nine selected alleles. We found that none of these alleles showed an increase of ARSE activity over time as compared with wtARSE ( Figure 2a–d ).

Figure 1
figure 1

Activity of ARSE missense proteins. Top, mutations are listed on the X axis from N to C terminus of ARSE. C86A alters the putative active site of ARSE and was used as a negative control. S156N is a reported ARSE polymorphism. Missense proteins in orange were subsequently tested for residual activity. The Y axis represents the ratio of averaged fluorescence of the mutant proteins to wtARSE. Bottom, expression of Renilla plasmid cotransfected with each ARSE mutant protein. Y axis shows measured Renilla luminescence units. Experiments represent three independent replications. ARSE, arylsulfatase E; wtARSE, wild-type ARSE.

Full size image
Figure 2
figure 2

Time course experiments for selected alleles. (a) G57S; (b) G73S, G137A, T306A; (c) G391R, G355S, T409M; (d) G434S and A463T show negligible activity over time compared with wtARSE. The Y axis represents 4-methylumbelliferyl-fluorescence (1.0s) (counts). Error bars represent the SD of the measurements. Experiments represent three independent replications. ARSE, arylsulfatase E; wtARSE, wild-type ARSE.

Full size image

Discussion

Spectrum of ARSE mutations

Twenty nine male probands with CDPX1 phenotype were referred through the CETT program. Molecular analysis identified ARSE mutations in 17, providing a mutation identification rate ~58%, similar to previous reports.3,5,10 Although the majority of ARSE mutations identified overall have been unique, several reported ARSE mutations occurred more than once in unrelated families, including G73S, G137A, G317A, T481M, P578S, and W581X. Of these, only G73S and T481M represent transitions at CpG dinucleotides. We also noted that the maternal carrier frequency may be high. Although only nine mothers were tested in this study, all were found to be carriers of the ARSE mutations identified in their sons. Similarly, nine mothers of previously reported males with CDPX1 who had carrier testing were found to be carriers.3,25,27 Maternal carriers were identified among both ARSE deletion and point mutation alleles (see Table 2 ). We suggest that the high frequency of maternal carriers, if proven, could be secondary to de novo mutations occurring more often in the germline of the maternal grandfather. Alternatively, given that patients with CDPX1 often survive and reproduce, the frequency of new mutations may be low.

Absence of genotype–phenotype association

The majority of missense substitutions were located in residues that were conserved in evolution for both ARSE and other arylsulfatases, ASA, ASB, ASC, ASD, ASH, and ASF. All missense alleles were distributed along most of the full protein and had negligible activity as compared with wtARSE. Of note, there were no missense alleles identified in exon 2, which contains the start methionine codon, and none in the putative transmembrane domains. These results, taken together with indistinguishable phenotypes of patients with ARSE gene deletions and nonsense and missense alleles, indicate the absence of an association between mutation and disease severity. This is further supported by variation in phenotype severity between affected males within the same family.3

It is possible that the missense alleles interfere with the proper localization of the protein or result in a catalytically inactive or misfolded and degraded protein. Daniele et al.9 showed that the ARSE missense alleles, R11P, G137V, G245R, and G492Y, when overexpressed in COS cells had negligible activity but retained stability and Golgi location, suggesting that these residues may be important for the catalytic activity of ARSE.

ARSE protein modeling

ARSE is most closely related evolutionarily to ARSC and likely derives from this gene by duplication.6 The amino acid identity between ARSE and ARSC is 51%. Recently, the crystal structure of ARSC, a transmembrane endoplasmic reticulum protein, was determined.28 The crystal structure implies that the active site rests near the membrane surface, suggesting an intriguing role for the lipid bilayer in catalysis. Sequence analysis predicts that ARSE also contains two putative membrane-spanning domains, which likely anchor ARSE to Golgi membranes.9 To further evaluate the effect of ARSE missense alleles on protein function, we modeled ARSE onto ARSC29 using the automated alignment mode in SWISS-MODEL.30 Structural 3D assessment for this model was created using the PyMOL Molecular Graphics System (Version 1.2r3pre; Schrödinger, LLC). Using this model, we visualized the location of the ARSE missense alleles identified in this article, as well as three known polymorphisms ( Figure 3 ). Of note, we found that most mutations clustered around the active site (shown in red in Figure 3 ), similar to previously evaluated ARSC missense mutations suspected of impairing catalytic activity.28 A smaller group of three mutations clustered around an ARSC mutation, also known to be inactive, that was positioned in a loop that might associate with the membrane ( Figure 3 ). Of note, the known polymorphisms were all positioned toward the external surface. As previously predicted, there were no mutations located in the transmembrane domains. This structural comparison suggests that ARSE missense mutations inactivate ARSE by impairing its catalytic site.

Figure 3
figure 3

Model of ARSE. ARSE was superimposed on arylsulfatase C crystal structure and visualized by PyMOL. β Sheets, connecting loops, and α helices are green. The 18-amino-acid active site is red. ARSE residues involved in mutations are depicted as blue spheres, and polymorphisms are yellow spheres (see Table 1 ). ARSC mutation residues analyzed by Ghosh (2004) (ref. 28) are light blue spheres. The two transmembrane domains are shown as parallel α helices at the bottom of the molecule. ARSC, arylsulfatate C ARSE, arylsulfatase E.

Full size image

Clinical comparison of BCP patients with and without ARSE mutations

Clinical data sheets were returned on 21/32 male probands (including those from the United Kingdom) but were not always fully completed. Nevertheless, the frequency of clinical findings (Supplementary Table S4a,b online) was similar to those reported before.3 In exception, rhizomelia and ichthyosis, atypical findings for CDPX1, were reported once in patients with ARSE mutations R90G and D47N, respectively. Ganglioneuroblastoma was reported once in a patient without an ARSE mutation. Cervical spine abnormalities, frequently reported in patients with BCP, were not reported in the group without ARSE mutations in this study. Most surprising was the absence of information related to potential phenocopies in the ARSE mutation–negative group. The limitations of this data collection included its nonmandatory requirement to be completed. In addition, more specific questions on maternal vitamin K deficiency states should have been included. These features limit the interpretation of the data.

Etiologies for BCP in patients without ARSE mutations

From the 12 male probands identified without ARSE mutations, there were no mothers reported with conditions previously associated with gestational vitamin K deficiency. With the caveat that the clinical data sheets might be incomplete for maternal history, it seems likely that genetic heterogeneity or other environmental factors not considered also contribute to mutation-negative patients. Future studies could evaluate the X-linked ARSE paralogs ARSD, ARSF, and ARSH. ARSD has been sequenced in some patients with BCP previously without mutations reported.5,25 Improved understanding of ARSE function would help to identify additional pathway proteins that, when defective, could be candidates for causing BCP.

One of these pathways may involve vitamin K, as fetal vitamin K deficiency in early gestation may be responsible for an unknown proportion of BCP phenotypes.3,12,13,16 Supporting these observations, offspring of pregnant women receiving warfarin showed shortening and flattening of the nose, epiphyseal stippling, and cerebral hemorrhage.31 Warfarin inhibits the VKORC1 complex, which recycles vitamin K. Offspring of pregnant rats treated with warfarin showed calcification and disruption of hypertrophic chondrocytes in long-bone epiphyses. These studies indicate that early gestational warfarin exposure results in bone dysplasia, whereas later exposure affects coagulation factors. Reduced bone γ-carboxyglutamate levels in the pups suggested that this phenotype was related to inhibited synthesis of vitamin K–dependent skeletal proteins.32 Vitamin K–dependent proteins are activated in the endoplasmic reticulum through posttranslational modification by γ-glutamyl carboxylase. γ-Glutamyl carboxylase adds CO2 to a bound glutamate residue to form γ-carboxyglutamate. Vitamin K is a cofactor in this reaction and is simultaneously oxidized.33,34 γ-Carboxyglutamate or Gla residues bind calcium with high affinity. Vitamin K–dependent proteins involved in bone metabolism are bone Gla protein (osteocalcin), expressed in osteoblasts and odontoblasts and matrix Gla protein, i.e., expressed in vascular smooth muscle cells and chondrocytes.35 Matrix Gla protein acts as an inhibitor of extracellular matrix mineralization and prevents vascular calcification.36 Mutations in MGP cause Keutel syndrome, which features chondrodysplasia punctata, midfacial hypoplasia, peripheral pulmonary stenosis and progressive soft tissue, and vascular calcifications.37,38 We speculate that ARSE might undergo γ-carboxylation and have functions similar to matrix Gla protein and osteocalcin, by catalyzing the desulfation of a critical growth plate component that inhibits calcification. Alternatively, ARSE may require vitamin K for normal growth plate mineralization in a pathway independent of γ-carboxylation.

In this project, the CETT program fostered collaboration between clinicians, researchers, and clinical laboratories. The direct depositing of mutations from the clinical laboratory into a public database39 allows access by the entire biomedical community for further investigation.

Disclosure

The authors declare no conflict of interest.