Introduction

Cartilage-hair hypoplasia (CHH; MIM 250250), or metaphyseal dysplasia, McKusick type, is an autosomal recessive disease that has diverse clinical manifestations. The most prominent features are short stature and hypoplastic hair. Other common features include metaphyseal dysplasia of the long and short tubular bones, defective immunity, predisposition to malignant tumors including lymphoma, ligamentous laxity, hypoplastic anemia, and Hirschsprung disease. Variations in clinical severity are remarkable both between and within families (Makitie and Kaitila 1993). In addition, CHH variants with only skeletal manifestations (metaphyseal dysplasia without hypotrichosis: MIM 250460) have been reported (Verloes et al. 1990; Castriota-Scanderbeg et al. 2001; Bonafe et al. 2002; Nakashima et al. 2003).

CHH is caused by mutations in the RMRP gene (Ridanpaa et al. 2001), which encodes a 267-nucleotide RNA component of RNase MRP (RMRP: ribonuclease mitochondrial RNA processing) (Ridanpaa et al. 2001; Bonafe et al. 2002). To date, 73 RMRP mutations have been reported (Ridanpaa et al. 2001, 2002; Bonafe et al. 2002, 2005; Nakashima et al. 2003; Kuijpers et al. 2003; Harada et al. 2005; Hermanns et al. 2005; Thiel et al. 2005). The 70A>G mutation is the most prevalent; it comprises 92% of mutations seen in the Finnish population and is commonly seen in other populations (Ridanpaa et al. 2002). RMRP mutations also cause a variant form of CHH (Bonafe et al. 2002), anauxetic dysplasia (MIM 607095) (Thiel et al. 2005), and Omenn syndrome (Roifman al. 2006). Mutations in RMRP can be classified into three categories according to the position of the mutation, described as follows: (1) promoter mutations involve duplication or insertion of several nucleotides between the TATA box and the transcription initiation site; (2) transcript mutations are one- or two-nucleotide changes within the transcribed region; (3) 5′ end mutations are insertions or duplications in the 5′ end of the transcript (Ridanpaa et al. 2002). The molecular pathogenesis of these mutations and genotype–phenotype associations are yet to be clarified.

To explore further the range of RMRP mutations in CHH, we examined the RMRP gene in Japanese CHH patients. Our study reveals several novel and common mutations, and characterizes a unique mutation spectrum and founder mutations in the Japanese population.

Subjects and methods

Patients and mutation analysis

A total of 11 CHH patients were included in this study. The inclusion criteria were generalized metaphyseal dysplasias of long and short tubular bones with or without extra-skeletal disorders known to be associated with CHH, including hair hypoplasia and immunologic disorders. RMRP mutations had previously been identified in two patients (CHH1, 2) (Nakashima et al. 2003).

Peripheral blood, hair, or fingernails were obtained with informed consent from the patients and their parents. Genomic DNA was extracted using standard procedures. Polymerase chain reaction (PCR) direct sequencing analysis for RMRP was performed as described (Nakashima et al. 2003).

Haplotype analysis

Seven patients (CHH1, 2, 4–8) who had RMRP mutations common in Japanese CHH were subjected to haplotype analysis. We constructed linkage disequilibrium (LD) blocks containing the RMRP gene using genotype data from 44 Japanese individuals in the HapMap phase I database (International HapMap Consortium 2005). The haplotype structure with its tag SNPs was determined using Haploview (Barrett et al. 2005). We genotyped five tag SNPs using the Taqman assay with a PRISM 7900 sequence detector (Applied Biosystems). Haplotypes for chromosomes harboring the two common Japanese mutations were determined by genotyping tag SNPs for seven patients and their parents.

Results

Identification of RMRP mutations

We screened nine patients for RMRP mutations and identified mutations in six (Table 1). Of these patients, four had hair hypoplasia or immunological disorders, and two showed skeletal changes only (CHH variant). Three mutations were novel, and three were recurrent (Nakashima et al. 2003; Bonafe et al. 2005). Other than 182G>T, these mutations have not been seen in populations other than the Japanese. We suppose that the three novel mutations are not polymorphisms because they were not detected in 65 Japanese controls. Four mutations were point mutations in the transcribed region, and two were duplications in the promoter region. All patients were compound heterozygotes.

Table 1 RMRP mutations and clinical phenotypes
Full size table

Genotype–phenotype association

Eight Japanese patients with RMRP mutations were evaluated. None of these patients possessed the full spectrum of skeletal, hair, and immunological disorders characteristic of classical CHH. Three had skeletal changes only (CHH variant), and hair hypoplasia was seen in just two patients, who were siblings. In most cases, birth length and height were not severely affected, although two patients who were short from birth developed very severe short stature (<−8 SD). Patient CHH4, who had transcript mutations on both alleles, showed a mild phenotype.

Haplotype analysis of mutation alleles

Including the two previously screened cases (CHH1 and 2), the two recurrent mutations (17-bp dup at +3 and 218G>A) were found in five patients, respectively. Because these mutations are so common among Japanese CHH cases, we suspected the presence of founders. To investigate this possibility, we determined the haplotype structure of RMRP in seven patients (CHH1, 2, 4–8) who had the common mutations.

Haplotypes were assessed by PCR direct sequencing, using the four SNPs within the RMRP region that showed minor allelic frequencies greater than 10% in the Japanese population (Nakashima et al. 2003; International HapMap Consortium 2005). Our analysis revealed that the two recurrent mutations were contained within distinct haplotypes (Fig. 1).

Fig. 1
figure 1

The haplotype structure of Japanese CHHs in the RMRP region. The two common mutations are contained within distinct haplotypes (colored), respectively. SNP positions from the transcription start site of RMRP (denoted as +1) are indicated at the top. The telomere is to the left. Under the Chromosome heading, F and M indicate the paternal and maternal chromosomes, respectively

Full size image

Next, we determined the haplotype block structure for the region surrounding RMRP using genotype data from 44 Japanese individuals in the HapMap database. RMRP was contained in a block spanning ∼20 kb (Fig. 2 top). The haplotype block was represented by seven haplotypes with >1% frequency (Fig. 2 bottom). The haplotype of the chromosome containing the 17-bp dup at +3 mutation was determined by parental–children transmission in three (CHH1, 5, 8) of five disease chromosomes. All three had haplotype III (AAAGG), which is seen in 13.6% of the Japanese population. The other two patients might have this haplotype, but the phase was not determined. Likewise, we determined the haplotype in two of five chromosomes containing the 218A>G mutation (CHH2 and CHH5). Both had a rare haplotype (ATTGA) that was not seen among the 44 Japanese individuals used in the haplotype block analysis. The three other patients might also possess this haplotype, but the phase was not determined (Fig. 3).

Fig. 2
figure 2

The linkage disequilibrium (LD) block and haplotype structure around RMRP in the Japanese population. Top Genomic structure and the LD block containing RMRP. LD was evaluated using D′ statistical analysis. Bottom The haplotype structure and frequency of the LD block containing RMRP. The block was represented by seven haplotypes. *Tag SNPs: SNP1, rs10972549; SNP2, rs1538537; SNP3, 3750430; SNP8, rs4878625; SNP10, rs2071675

Full size image
Fig. 3
figure 3

The haplotype structure of the region around RMRP. The two common mutations are contained within distinct haplotypes (colored). The telomere is to the right. SNP positions are indicated from the transcription start site of RMRP (denoted as +1)

Full size image

Discussion

The mutation spectrum for RMRP in the Japanese population is unique. The 70A>G founder mutation that is prevalent in Western populations (Ridanpaa et al. 2002) has not been reported in Japanese individuals. On the other hand, the 218A>G and 17-bp dup at +3 mutations are common in Japanese patients, but have not been reported in other populations. We have shown that the two common mutations are contained within rare distinct haplotypes, indicating the presence of unique founders among Japanese CHH patients. Because these haplotypes have been defined by tag SNPs, our results will be useful in detecting CHH mutations and their carriers in Japanese CHH patients.

We saw no patient who possessed all of the skeletal, hair, and immunological features characteristic of classical CHH. Extra-skeletal features are not prevalent in the Japanese population, although immunological disorders may yet develop in this young set of patients. As described previously (Nakashima et al. 2003), all patients were compound heterozygotes for insertions or duplications in the promoter or 5′ transcribed regions and point mutations in the transcribed region. The former mutations cause marked decreases in gene transcription and, hence, the substantial effects on phenotype; in contrast, the latter mutations have only mild effects on transcription (Nakashima et al. 2003). This is consistent with the mild phenotype observed in patient CHH4, who had two point mutations. Although it is clear that RMRP mutations produce a spectrum of phenotypes, from CHH variants with only skeletal manifestations (Bonafe et al. 2002) at the mild end to anauxetic dysplasia (Thiel et al. 2005) at the severe end, further studies are necessary to better delineate these phenotype–genotype correlations.