Clin Genet 2002: 61: 146–151 Copyright C Munksgaard 2002Printed in Denmark. All rights reserved

CLINICAL GENETICS0009-9163

Short Report

RMRP gene sequence analysis confirms acartilage-hair hypoplasia variant with onlyskeletal manifestations and reveals a highdensity of single-nucleotide polymorphisms

Bonafe L, Schmitt K, Eich G, Giedion A and Superti-Furga A. RMRP L Bonafe a, K Schmitt b, G Eich c,gene sequence analysis confirms a cartilage-hair hypoplasia variant with A Giedion a and A Superti-Furga a

only skeletal manifestations and reveals a high density of single-nucleotide a Division of Metabolism and Molecularpolymorphisms. Paediatrics and Division of Radiology,Clin Genet 2002: 61: 146–151. C Blackwell Munksgaard, 2002 University Children’s Hospital, Zurich,

Switzerland, b Landeskinderklinik Linz, Linz,Mutations in the RMRP gene that codes for an RNA subunit of the MRP Austria, c Division of Paediatric Radiology,RNAse complex are the cause of cartilage-hair hypoplasia (CHH; MIM Kantonsspital, Aarau, Switzerland250250). We tested the hypothesis that recessive metaphyseal dysplasiawithout hypotrichosis (M1M 250460), a disorder presenting with shortstature and metaphyseal dysplasia similar to CHH, but lacking hairanomalies, immunodeficiency and other extra skeletal features, might beallelic to CHH. We identified four mutation-carrying alleles segregatingwith the skeletal phenotype in two unrelated boys and their parents. Oneallele carried the common Finnish mutation π70A » G; the remainingthree carried π195C » T, π238C » T, and dupAAGCTGAGGACG at Key words: cartilage-hair hypoplasia –ª2. Sequencing 120 alleles from a control group revealed an unusually genetic mutations – metaphysealhigh density of single-nucleotide polymorphisms in and around the dysplasia – RMRP gene – single-nucleotideRMRP gene: the biological significance of this finding is unclear. polymorphismsWe conclude that recessive metaphyseal dysplasia without hypotrichosis Corresponding author: Professor Andreais a variant of CHH, manifesting only as short stature and metaphyseal Superti-Furga, Division of Metabolism anddysplasia. Precise diagnosis of this form of metaphyseal dysplasia is not Molecular Paediatrics, University Children’swithout importance because of recessive inheritance with corresponding Hospital, Steinwiesstrasse 75, CH-8032recurrence risk, as well as because of potential complications such as Zurich, Switzerland. Tel.: 41-1-266-7722;

fax: 41-1-266-7167;anaemia, susceptibility to infections and the increased likelihood ofe-mail: asuperti/access.unizh.chdeveloping cancer. The short stature and metaphyseal changes associated

with cone-shaped epiphyses of the hands should raise the diagnostic Received 29 October 2001, revised andpossibility of a CHH-related disorder that can then be confirmed by accepted for publication 3 Decembermutation analysis. 2001

The gene for cartilage-hair hypoplasia (CHH; met-aphyseal dysplasia, McKusick type; MIM 250250)has recently been identified (1). Alone out of allthe group of skeletal dysplasias, it codes not for aprotein, but for a RNA subunit of an RNAse com-plex. The pathogenesis of CHH is not yet under-stood, but RMRP may be a housekeeping gene,which would explain the pleiotropism of the dis-order.

A spectrum of mutations was identified in aseries of patients with classical CHH (1). Thesubjects were mostly of Finnish origin, but in-

146

cluded a Swiss child with severe CHH and im-munodeficiency who had undergone bone mar-row transplantation (2). There is a group of pa-tients who have short stature and radiographicchanges similar to those of CHH, but who alsohave normal hair, no immunodeficiency andnone of the other extra skeletal changes associ-ated with CHH [‘metaphyseal dysplasia withouthypotrichosis’, MIM 250460 (3)]. The presentstudy was undertaken to test whether RMRPmutations could be identified in patients belong-ing to this group.

RMRP mutations in skeletal variant of CHH

Subjects and methodsSubjects

Case 1

This boy was born at term to parents of Swiss andDanish extraction, and had a normal birth lengthof 52cm (P50–90). The subject’s growth declinedto P10 during the first year of his life and toslightly below P3 by 3years of age. At 7.5years ofage, he was 109cm in height, 8cm below P3 forSwiss children, but 2cm above P90 for Finnish pa-tients with CHH (4).

The subject was referred to us at 3years of agebecause of his short stature and moderate varusdeformity of the knees. Radiographic examinationat that time showed modest metaphyseal changes.Cartilage-hair hypoplasia was considered in thedifferential diagnosis and a haematological-im-munological work-up was performed. The sub-ject’s haemoglobin level was borderline low for hisage at 109gLª1, and total IgG (4.23gLª1) andtotal lgA (0.22gLª1) in plasma were also border-line low. A more extensive work-up at 7years ofage showed a slightly decreased level of haemo-globin (110gLª1), but normal values for severalparameters of cellular immunity (i.e. mitogen andantigen stimulations, lymphocyte subpopulations,expression of surface antigens, and adhesion mol-ecules). Therefore, an immune deficiency was ex-cluded.

The boy’s mother noted a susceptibility to highfever during childhood, but unusual or serious in-fections were never found. The subject had var-icella at 6years of age with pronounced efflor-escences, but no complication. Resection of hyper-trophic adenoids was done at 7years of age, alsowithout complications. He grows abundant hair ofdark blond to light brown colour that has neverbeen considered abnormal; hair thickness meas-urements have not been performed. Physical ex-amination showed moderately short stature withrelatively short legs (Fig.1) and mild joint laxity.The radiographic features are shown in Fig.1; al-though the diagnosis of a CHH variant was con-sidered at 3years of age, it was only at 7years thatmore characteristic features of CHH were seen,namely the metaphyseal irregularities and particu-larly the phalangeal cone-shaped epiphyses of thehand. The pelvis and the spine did not show pecu-liarities.

Case 2This boy was born at term to parents of Austrianorigin. His birth length was normal (48cm, P25),but his growth declined to slightly below P3 at 1(68.5cm) and 2years of age (77cm). At 3.2years of

147

age, the subject’s height was 84cm (below P3 forthe reference paediatric population, but above P90for Finnish patients with CHH): his growth chan-nel has not changed since then, and at 10years ofage he was 116cm tall [13cm below P3 for con-trols, but 1cm above P90 for Finnish patients withCHH (4)].

The subject was referred to us for re-evaluationof his short stature (Fig.1). He had relatively shortlegs and laxity of the small joints of the hands.Radiographic examination revealed metaphysealstriations and phalangeal cone-shaped epiphysesof the hand (Fig.1). Metaphyseal dysplasias with-out hypotrichosis or a CHH variant were con-sidered in the differential diagnosis. A haematolog-ical-immunological work-up at 10years of ageshowed normal blood haemoglobin (126gLª1),plasma IgG (9.25g lª1), IgG subclasses and plasmalgA (1.03g lª1) for his age. A lymphocyte subpopu-lation study gave normal results. There was no ab-normal susceptibility to infection during child-hood, with varicella infection producing severe ef-florescences but no complication, and the subject’sgeneral health has been excellent. As in case 1, hishair has always been considered normal; its thick-ness has not been measured.

Methods

Genomic DNA was obtained with informed con-sent from both patients and their parents, and ana-lysed by polymerase chain reaction (PCR) ampli-fication and direct sequencing of the RMRP gene.To obtain a reference sequence and ascertain poly-morphisms in the RMRP gene, genomic DNAfrom 60 anonymized unrelated controls of differ-ent ethnicities (27 Swiss, nine Italians, two French,five Germans, five Dutch, three British, one Span-ish, one Belgian, one Czech, one Austrian, fourAmericans and one Australian) was also studiedby PCR amplification and sequencing.

The sequence of the most 3ƒ primers describedby M. Ridanpää et al. (1) is not contained in thesequence deposited at GenBank (M29916). Thus,we amplified the RMRP gene using a new primerCHH-11-F, located 5ƒ of the TATA box and the 3ƒprimer RM4R previously published by M. Rid-anpää et al. (Fig.2). Polymerase chain reactionamplification of the RMRP gene as a single frag-ment of <700bp was performed in a PE9700Thermal Cycler (Perkin Elmer) by using thePCRmix Promega kit (39 cycles with annealingtemperature 56 æC). We subsequently sequenced theRMRP amplified fragment in both directions usingthe end-standing primers CHH-11-F and RM4R.From the sequence obtained with primer RM4R,a second reverse internal primer CHH-14-R, closer

Bonafe et al.

Fig.1. Clinical and radiographicfeatures of the two boys with a skeletalpresentation of cartilage-hairhypoplasia (CHH): Top row: Case 1 at 3years of age. There is moderate,proportionately short stature with mildvarus deformity of the legs. Note theabundant hair. Knee radiographs takenat 7years of age showing metaphysealirregularities with radiolucent areas andcoarser trabeculation at the knees. Handradiographs taken at the same ageshowing moderate metacarpalshortening, as well as multiplephalangeal epiphyses with half-moon ortrapezoid shapes (arrows) which are mostevident at the middle phalanges. Bottomrow: This photograph of case 2 (chosenby the patient himself) shows him at 9years of age. There is short stature, butno other features of CFF. Radiographsof the subject’s knees at 10years of ageshowing metaphyseal changes which areslightly more pronounced than those seenin case 1, with striations extending fromthe metaphyses towards the diaphyses.The hand radiographs are similar tothose of case 1, but the changes areslightly more pronounced, possiblybecause of the subject is older. Themetacarpals are moderately shortenedand there are cone-shaped phalangealepiphyses with rounded half-moonshapes (arrows) throughout the proximaland medial phalangeal rows,occasionally with some irregularitiesproducing a trapezoid or ‘tent-shaped’appearance. The half-moon shapedcones are virtually specific for CHH(19).

to the end of the transcribed region, was designedand used to sequence the first amplicon (Fig.2).The ABI Prism BigDyeTerminator Ready Reactionkit and protocol (Applied Biosystems) were usedfor sequencing in a ABI Prism 310 Genetic Ana-lyzer.

ResultsExtending the sequence of the RMRP gene

We determined the sequence beyond the 3ƒ of thepublished data and used that to position a closerprimer (CHH-14-R) more suitable for bi-direc-tional overlapping sequencing. In Fig.2, 224nt ofgenomic sequence additional to that deposited inGenBank (M29916) is presented.

Mutation analysis

We identified several single base differences whichwere present both in the controls and in one of the

148

CHH families: three of these map just 5ƒ of theTATA box, one maps between the TATA box andthe transcribed region, and two map immediately3ƒ of the coding region. Two single nucleotidechanges are located within the transcribed region(π156G/C and π177C/T), but involve nucleotideswhich are not conserved throughout species (5, 6).All these putative polymorphisms (with one excep-tion,π272t/c) were present at the homozygousstate in some of the controls (Table1). Three ad-ditional sequence changes were found only at theheterozygous state in single control subjects (Table1): ª24c/g (located just after the TATA box),π250C/T and π58insA. These nucleotides are nothighly conserved through species (6), and accord-ing to the alignment of E. Sbisa et al., there is anadditional adenosine at π58 positions in both Xen-opus and Saccaromyices cerevisiae (5). Neverthe-less, it cannot be excluded that these changesmight represent pathogenic alleles at the hetero-zygous state.

RMRP mutations in skeletal variant of CHH

Fig.2. Extended genomic sequence ofthe RMRP gene (see GenBankM29916). The transcribed region isindicated in capital letters. Theasterisk denotes the 5ƒ end of thesequence contained in GenBank. TheTATA box is boxed. The mutationspreviously described by M. Ridanpääet al. in 2001 (1) are indicated as thesequence variation below the wild-type(point mutations) or as arrows markingthe position of insertions/duplications.The newly recognized point mutationsand duplication are indicated in boldtype and as a bold arrow, respectively.The putative polymorphisms areindicated in italic type. The primersused for amplification (CHH-11-F andRM4R) and sequencing (CHH-11-Fand CHH-14-R) are also indicated.

Four mutations were identified in the two CHHfamilies. In family 1, the affected child was a com-pound heterozygote for the duplication of 12

Table 1. Newly identified putative pathogenic mutations and polymorphisms in the RMRP gene

Sequence changes* Frequency in 120 Homozygous Conservedcontrol alleles controls out of 60 through species

New putative pathogenic mutationsdup(aagctgaggacg) at ª2 Not found – –π195C » T Not found – Yesπ238C » T Not found – Yes

New polymorphisms†ª58t/c 76/44 8 –ª56a/g 100/20 3 –ª48c/a 56/64 17 –ª6g/a 100/20 3 –π156G/C 100/20 3 Noπ177C/T 100/20 3 Noπ272t/c 118/2 0 –π274t/c 82/38 5 –

Other sequence changesª24c/g 119/1 0 –π250C/T 119/1 0 NoinsA at π58 119/1 0 No

*Numbering refers to the first transcribed nucleotide (π1). Nucleotides located within the transcribed region are indicated with capitalletters.†In addition to the haplotype deposited in GenBank, the co-segregation of some single-nucleotide polymorphisms allowed us topropose three additional probable haplotypes: (1) haplotype a, ª58c and ª48a; (2) haplotype b, ª58c, ª48a and π274c; and (3)haplotype c, ª56g, ª48a, ª6a, π156C and π177T.

149

nucleotides (AAGCTGAGGACG) between theTATA box and the first nucleotide of the codingregion, and for a π195C»T transition within the

Bonafe et al.

transcribed region (Table1, Fig.2). These muta-tions were present in the heterozygous state in thefather and in the mother, respectively. The mat-ernal allele in case 1 also carried the changesª56a/g, ª48c/a, ª6g/a, π156G/C and π177C/T,all putative polymorphisms found at the homo-zygous state in the mother and in several controls(Table1). Thus, the index case was heterozygous ata total of six distinct nucleotides within the ana-lysed region. In family 2, the affected child wascompound heterozygous for transition π70A»G[the Finnish mutation previously described (1)]and transition π238C»T. Also in this case, themutations were heterozygous in the father andmother, respectively. The 12nt duplication on thepaternal allele of patient 1 is located two nucleo-tides before the transcription initiation site. Otherinsertion/duplications have been previously de-scribed in the same 5ƒ-flanking region between theTATA box and the transcribed region (Fig.2).Both the π195C»T and the π238C»T tran-sitions affect nucleotides which are conservedthrough all compared species. The π195C»T isonly two nucleotides away from the π193G»Atransition identified by M. Ridanpää et al. (l). Nei-ther the duplication nor the two point mutationswere found in 60 unrelated control subjects (120alleles).

DiscussionMolecular analysis of the RMRP gene

To establish a mutation analysis procedure forthe RMRP gene, we first had to determine thesequence of the untranslated region at the 3ƒ-endof the gene between the end of the published se-quence and the published reverse primer.

Sequencing the gene in a population of controlindividuals, we identified several single nucleotidepolymorphisms (SNPs), none of which has beenreported previously (1). A possible explanation isthe more heterogeneous ethnic composition of ourcontrol group. This number of SNPs within a gen-omic stretch of <400bp is surprising since it sig-nificantly exceeds the average SNP density in thehuman genome, which has recently been estimatedat less than one in a thousand nucleotides (7).Whatever the biological significance of the highfrequency of polymorphisms in and around theRMRP gene, this fact must be taken into accountwhen performing molecular diagnostics for CHH,as exemplified by case 1, who was heterozygous atsix single nucleotides in the genomic segment thatwe studied. A high number of SNPs within a shortDNA segment that can be amplified by a singlePCR may be useful for population studies.

150

The 12nt duplication on the paternal allele ofcase 1 results in a lengthening of the distance be-tween the TATA box and the transcription initia-tion site. It is similar in nature to three insertion/duplication mutations found by M. Ridanpääet al. in three other CHH families (including ourpatient with severe CHH) (1). Those insertionshave been shown to result in reduced transcriptlevels. We assume that the duplication we foundhas similar effects. The two transitions π195C»T and π238C»T found in the CHH families onlyreside in residues conserved in all species as it ap-pears from comparisons between yeast, Arabi-dopsis thaliana, Xenopus laevis, mouse, rat, cowand human RNA sequences (5, 6), whereas the twoidentified polymorphisms located inside the tran-scribed region are both not highly conservedthrough species. We noted that neither the pre-viously described mutations π70A »G andπ193G »A, nor the two point mutations iden-tified here, map to those regions which had beenidentified as base-pair conserved and considered tobe important for secondary structure (6). This mayimply either that predictions concerning secondarystructure must be refined, or alternatively, thatthese nucleotides are important not for secondarystructure, but rather for interaction with the pro-tein component(s) of RNAse MRP or with sub-strate RNAs. If the latter is true, it is conceivablethat different mutations may affect binding of dif-ferent RNA substrates (1, 8), and thus explain thepleiotropism of the disorder as well as the vari-ability of the expressed phenotype (8, 11).

Clinical and nosological aspects

In their milestone paper on the identification ofRMRP as the CHH gene, M. Ridanpää et al. em-phasized the pleiotropism of the disorder (1). In-deed, the Finnish group has highlighted several as-pects of CHH (9), such as growth retardation (4),anaemia (10, 11), Hirschsprung disease (12), sus-ceptibility to cancer (13), increased mortality (14),and particularly, the various facets of the immuno-deficiency (15, 16). Our observations do not em-phasize the pleiotropism, but rather, the variableexpression of this disorder.

There are strong indications for the variable se-verity of CHH. First, most of the complicationsdescribed in the Finnish population are not obli-gate. Secondly, intrafamilial variability has beenobserved (17). Lastly, the literature contains de-scriptions of individuals with metaphyseal dys-plasia who, because of the absence of susceptibilityto infections and of hair changes, have been classi-fied as CHH variants or CHH-like, or just as meta-physeal dysplasia (3, 18). A. Verloes et al. (3) re-

RMRP mutations in skeletal variant of CHH

viewed a series of such patients who are listed inMIM under the entry ‘recessive metaphyseal dys-plasia without hypotrichosis’ (MIM 250460) andconsidered the possibility that the condition mightbe allelic to true CHH. The results obtained in thetwo boys studied by the present authors confirmthat there is a purely skeletal presentation of CHH.Thus, MIM entry no. 250460 may be incorporatedto no. 250250, cartilage-hair hypoplasia.

The reclassification of the two cases described inthe present study raises the question of whether ornot they are at risk for the other known compli-cations of the disorder, particularly of immuno-deficiency and an increased risk of cancer. Furtherstudies aimed at determining whether there aremore precise predictive factors for infections and/or cancer would be useful.

Retrospectively, an important diagnostic telltalesign in both our patients was the pattern of phal-angeal cone-shaped epiphyses of the hands (seeFig.1 and its legend). These were present in bothboys and were indistinguishable from those ob-served in Finnish individuals with CHH (19).Cone-shaped epiphyses are not seen in either oneof the other metaphyseal dysplasias. Their speci-ficity lead us to pursue the diagnostic hypothesisof CHH in our patients and to undertake the mol-ecular study. However, it must be noted that thecone-shaped epiphyses were only seen when theboys were 7 and 9years old, respectively: they werenot seen in patient 1 at 3years of age. The diffi-culty of radiographic diagnosis of CHH in the firstyears of life is well known (20). Although the find-ings in these two cases should not be taken as ofgeneral value, the pattern of radiographic changesin the hands of patients with CHH is quite specific,and when cone-shaped epiphyses are seen in com-bination with isolated metaphyseal dysplasia, wesuggest that CHH is likely and mutation analysisof RMRP will confirm the diagnosis.

Acknowledgements

We thank the two boys and their parents for consenting to pub-lish their photographs. We are grateful to Beat Steinmann,Gabor Matyas and Sheila Unger for commenting the manu-script. This study was supported by Swiss National Foundationgrant no. 31-57272.99.

References

1. Ridanpää M, van Eenennaarn H, Pelin K, et al. Mutationsin the RNA component of RNase MRP cause a pleiotropichuman disease, cartilage-hair hypoplasia. Cell 2001: 104:195–203.

151

2. Berthet F, Siegrist CA, Ozsahin H, et al. Bone marrowtransplantation in cartilage-hair hypoplasia: correction ofthe immunodeficiency but not of the chondrodysplasia.European J Pediatr 1996: 155: 286–290.

3. Verloes A, Pierard GE, Le Merrer M, Maroteaux P. Reces-sive metaphyseal dysplasia without hypotrichosis. A syn-drome clinically distinct from McKusick cartilage-hair hy-poplasia. J Med Genet 1990: 27: 693–696.

4. Makitie O, Perheentupa J, Kaitila I. Growth in cartilage-hair hypoplasia. Pediatr Res 1992: 31: 176–180.

5. Sbisa E, Pesole G, Tullo A, Saccone C. The evolution of theRNase P- and RNase MRP-associated RNAs: phylogeneticanalysis and nucleotide substitution rate. J Mol Evol 1996:43: 46–57.

6. Schmitt ME, Bennett X, Dairaghi DJ, Clayton DA. Sec-ondary structure of RNase MRP RNA as predicted byphylogenetic comparison. FASEB J 1993: 7: 208–213.

7. Sachidanandam R, Weissman D, Schmidt SC, et al. A mapof human genome sequence variation containing 1.42 mil-lion single nucleotide polymorphisms. Nature 2001: 409:928–933.

8. Tollervey D, Kiss T. Function and synthesis of small nu-cleolar RNAs. Current Opinion Cell Biol 1997: 9: 337–342.

9. Makitie O, Kaitila I. Cartilage-hair hypoplasia-clinicalmanifestations in 108 Finnish patients. Eur J Pediatr 1993:152: 211–217.

10. Makitie O, Rajantie J, Kaitila I. Anaemia and macro-cytosis-unrecognized features in cartilage-hair hypoplasia.Acta Paediatr 1992: 81: 1026–1029.

11. Makitie O, Juvonen E, Dunkel L, Kaitila I, Shines MA.Anemia in children with cartilage hair hypoplasia is relatedto body growth and to the insulin-like growth factor sys-tem. J Clin Endocrinol Metab 2000: 85: 563–568.

12. Makitie O, Kaitila I, Rintala R. Hirschsprung disease as-sociated with severe cartilage-hair hypoplasia. J Pediatr2001: 138: 929–931.

13. Makitie O, Pukkala E, Teppo L, Kaitila I. Increased inci-dence of cancer in patients with cartilage-hair hypoplasia.J Pediatr 1999: 134: 315–318.

14. Makitie O, Pukkala E, Kaitila I. Increased mortality in car-tilage-hair hypoplasia. Arch Dis Child 2001: 84: 65–67.

15. Makitie O, Kaitila I, Savilaliti E. Susceptibility to infectionsand in vitro immune functions in cartilage-hair hypoplasia.Eur J Pediatr 1998: 157: 816–820.

16. Makitie O, Kaitila I, Savilaliti E. Deficiency of humoralimmunity in cartilage-hair hypoplasia. J Pediatr 2000: 137:487–492.

17. van der Burgt I, Haraldsson A, Oosterwijk JC, van EssenAJ, Weemaes C, Hamel B. Cartilage hair hypoplasia, meta-physeal chondrodysplasia type McKusick: description ofseven patients and review of the literature. Am J MedicalGenet 1991: 41: 371–380.

18. Castriota-Scanderbeg A, Dallapiccola B, Mingarelli R,Kozlowski K. Distinctive metaphyseal chondrodysplasiasimulating cartilage hair hypoplasia. Am J Med Genet2001: 99: 289–293.

19. Giedion A. Phalangeal cone-shaped epiphyses of the hand:their natural history, diagnostic sensitivity, and specificityin cartilage hair hypoplasia and the trichorhinophalangealsyndromes I and II. Pediatr Radiol 1998: 28: 751–758.

20. Le Merrer M, Maroteaux P. Cartilage hair hypoplasia ininfancy: a misleading chondrodysplasia. Eur J Pediat 1991:150: 847–851.