RAPID COMMUNICATION

Identification of a Novel Mutation in the CodingRegion of the Grey-Lethal Gene OSTM1 in HumanMalignant Infantile Osteopetrosis

Alfredo Ramırez,1n Julia Faupel,1 Ingrid Goebel,1 Anne Stiller,1 Susanne Beyer,1 Christina Stockle,1

Carola Hasan,2 Udo Bode,2 Uwe Kornak,3 and Christian Kubisch1

1Institut fur Humangenetik, Universitat Bonn, Germany; 2Zentrum fur Kinderheilkunde, Universitat Bonn, Germany; 3Institut fur MedizinischeGenetik, Charite, Humboldt-Universitat Berlin, Germany

Communicated by Arnold Munnich

Autosomal recessive malignant infantile osteopetrosis (ARO) is characterized by severe osteosclerosis,pathologic fractures, hepatosplenomegaly, and pancytopenia. The pathophysiological basis is inadequate boneresorption due to osteoclast dysfunction. In the majority of cases, mutations in either of two human genes causethis fatal disorder: TCIRG1, encoding a subunit of the osteoclast H+-ATPase, and the voltage-gated chloridechannel gene CLCN7. We excluded both genes in a small inbred family with malignant infantile osteopetrosisand undertook linkage analysis of several candidate loci that are involved in murine osteopetrosis. A regionspanning more than 20 cM between the markers D6S1717 and D6S1608 on chromosome 6q21 was found to behomozygous in the affected child. This locus is syntenic to the genomic region harboring the gene for theosteopetrotic mutant mouse grey-lethal (gl). Recently, mutations in a novel gene of unknown function weredescribed in the grey-lethal mouse and in one human patient. Mutation screening of the grey-lethal gene(OSTM1), revealed a homozygous 2-bp deletion in exon 2 (c.415_416delAG) in the affected child. Nomutations could be found in six independent ARO patients who had tested negative for mutations in TCIRG1and CLCN7. In summary, we describe the identification of a novel mutation in the coding sequence of thehuman grey-lethal gene, which is the second OSTM1 mutation found in human ARO, confirmingthe involvement of this gene in the pathogenesis of this severe bone disease. Hum Mutat 23:471–476,2004. r 2004 Wiley-Liss, Inc.

KEY WORDS: osteopetrosis; ARO; grey-lethal; GL; osteoclast; OSTM1

DATABASES:

OSTM1 – OMIM: 607649, 259700 (ARO); GenBank: NM_014028.2

INTRODUCTION

Bone homeostasis requires a tight balance between theactivity of osteoblasts, which produce bone, andosteoclasts, which degrade it. A perturbation of thisequilibrium can lead to a reduction of bone mass, as seenin osteoporosis, or to an abnormal accumulation of bone,as observed in osteopetrosis [Karsenty, 1999; Teitelbaum,2000]. In humans, osteopetrosis is classified according toseverity and age of onset [Lazner et al., 1999].Autosomal recessive malignant infantile osteopetrosis(ARO; MIM# 259700) is characterized by a severeosteosclerosis becoming apparent during the first monthsof life. Common complications are pathologic fractures,hepatosplenomegaly, and pancytopenia. Visual impair-ment and hearing loss can result from cranial nervecompression [Gerritsen et al., 1994b; Thompson et al.,1998], but there is also evidence for a primary retinaldegeneration [Keith, 1968]. Without bone marrowtransplantation, the disease is usually lethal [Gerritsenet al., 1994a].

The genetic heterogeneity of ARO is illustrated byseveral spontaneous mouse mutants (e.g., gl/gl, oc/oc)[Grynpas, 1993; Marks et al., 1985] and by mousemodels developed by gene targeting (e.g., c-Fos, c-Src,Opgl) [Kong et al., 1999; Soriano et al., 1991; Wanget al., 1992] that are osteopetrotic. In most humanorthologs of murine osteopetrosis genes no mutationshave been identified so far [Bernard et al., 1998; Lazneret al., 1999].Recently, mutations in the a3 subunit of the H+-

ATPase gene (Atp6i; human ortholog TCIRG1, MIM#604592) were found to cause an osteopetrotic phenotype

Received 25 October 2003; accepted revised manuscript 13January 2004.

nCorrespondence to: Alfredo Ram|¤ rez, Institut fˇr Humangenetik,Wilhelmstrasse 31, Bonn 53111, Germany.E-mail: a.ramirez@uni-bonn.de

Grant sponsor: BONFOR.

DOI10.1002/humu.20028Published online inWiley InterScience (www.interscience.wiley.com).

rr2004 WILEY-LISS, INC.

HUMANMUTATION 23:471^476 (2004)

in the osteosclerotic (oc) mouse mutant [Scimeca et al.,2000] and in a knock-out mouse model [Li et al., 1999].In addition, mice deficient for the voltage-gated chloridechannel gene Clcn7 (human ortholog CLCN7, MIM#602727) suffer from a severe osteopetrosis [Kornak et al.,2001]. In both cases, mutations in the human orthologshave been found to also cause ARO [Frattini et al., 2000;Kornak et al., 2000, 2001].

The phenotype of the spontaneous mouse mutantgrey-lethal (gl) corresponds very well to human auto-somal recessive osteopetrosis [Grynpas, 1993]. Thehomozygous gl/gl is fully penetrant, and can be rescuedby bone marrow transplantation [Walker, 1993]. Inaddition to the skeletal phenotype, gl/gl mice display agray coat color instead of agouti, which makes a role ofthe Gl gene in melanocyte function likely [Vacher andBernard, 1999]. The gl locus was mapped to mousechromosome 10, first between the Fyn and the Ros geneand later on in a fine mapping approach between the Fyngene and D10Mit148, a region which is syntenic tohuman chromosome 6q16-21 [Chalhoub et al., 2001;Vacher and Bernard, 1999]. Recently, the murine Glgene (osteopetrosis associated transmembrane protein 1,Ostm1) was cloned, which encodes a 338-amino-acidprotein without homology to any known proteinsequence. Analysis of the protein sequence suggests apotential cleavage site for a signal peptide and oneputative transmembrane domain. Mutation screening ofthe human OSTM1 (HUGO aliases: HSPC019, GL;MIM# 607649) gene showed only one splice-sitemutation in 1 out of 19 patients with recessiveosteopetrosis without mutations in TCIRG1 and CLCN7[Chalhoub et al., 2003].

We could not identify mutations in TCIRG1 andCLCN7 in a small inbred family with two membersaffected with infantile malignant osteopetrosis. There-fore, we started to search for the causative gene by acandidate locus approach using linkage analysis. Wereport here the identification of homozygosity in a regionspanning approximately 20 cM on chromosome 6q21 inthis family, which is syntenic to the murine gl locus.Sequencing of the OSTM1 gene revealed a homozygous2-bp deletion in exon 2 in the affected child. In summary,we describe the identification of a novel ARO mutationin the coding sequence of the grey-lethal gene OSTM1.

MATERIALSANDMETHODSPatients

A consanguineous Kuwaiti family with two affected children wasinvestigated; both parents are healthy. The son developed typicalARO symptoms and clinical signs of a neurodegeneration and diedat the age of 6 months, after having received an unsuccessful bonemarrow transplantation at the age of 5 months. No blood samplefrom him was available. His sister developed a severe sepsis at theage of 3 days. She showed the typical signs of a severe ARO,namely an increase in bone density, hepatosplenomegaly, andthrombocytopenia at the age of 2 months. At the age of 3 months,she also developed neurological symptoms, including visualimpairment and increased tone and reflexes of both upper andlower limbs. Due to the poor prognosis, the parents did not agreeto a bone marrow transplantation. After having given informed

consent, blood samples were taken from the parents and theaffected daughter for DNA analysis. The additional six patientsthat were screened for mutations were all sporadic cases. Allshowed the clinical hallmarks of ARO and varying neurologicalsymptoms. Visual loss was reported in two cases, seizures wereobserved in one patient. Two patients died after unsuccessful bonemarrow transplantation.

Genotyping

Genomic DNA was extracted from EDTA blood samplesaccording to standard protocols. Highly polymorphic microsatellitemarkers were selected from candidate regions according to theGenethon and Marshfield maps and public data from the humangenome project (www.ensembl.org/ and http://genome.ucsc.edu/).PCR amplification was performed using standard protocols. Formicrosatellite analyses equal volumes of PCR products and loadingbuffer underwent electrophoresis for 2.5 hr at 70 W on a 6%polyacrylamide gel. Gels were stained with AgNO3 and analyzedby at least two independent persons.

Characterization of the Human GL Locus

A detailed analysis of the human chromosomal region syntenic tothe mouse gl locus was done by comparing the physical andtranscriptional map of the gl locus on mouse chromosome 10, asdescribed by Chalhoub et al. [2001] with the public data of the humangenome project. The gene content and gene order of the human locuson chromosome 6q21 were established by database analysis.

Mutation Screening of Candidate Genes

Using public databases, candidate genes were selected from thecritical region. Their genomic structure was delineated and a PCR-based testing strategy on genomic DNA was established. All exonswere amplified by standard PCR on genomic DNA of the femalepatient and her parents (Family R). Mutation analysis wasperformed by direct sequencing of each amplicon. The genomicstructure of the OSTM1 gene was deduced using the human grey-lethal cDNA GenBank number NM_014028 [Chalhoub et al.,2003]. Mutation numbering is based on the cDNA sequence ofthe GenBank entry NM_014028.2, position +1 corresponds to theA of the ATG translation initiation codon. (Primer sequences areavailable from the authors.) A control sample consisting of 50DNA samples from German individuals was tested for thec.415_416delAG mutation in the OSTM1 gene by restrictiondigestion of the exon 2 PCR product using XcmI. The reaction wasperformed in a 25-ml final volume containing 15 ml of PCR productplus the restriction cocktail.

RESULTSFurther Locus Heterogeneity in Autosomal RecessiveOsteopetrosis

We identified a consanguineous Kuwaiti family withtwo children affected by ARO (Family R). The son wasdeceased and no samples were available, yet bloodsamples were taken from the affected daughter and bothhealthy parents, which are second degree cousins, aftergiving informed consent (for clinical description seeMaterials and Methods section). We performed linkageanalysis for the TCIRG1 locus on human chromosome11q, which did not show homozygosity in the affectedchild in Family R, thereby excluding the H+-ATPase asthe causative gene (see Kornak et al. [2000]). Directsequencing of all coding exons and adjacent splice sites ofthe second ARO gene on 16p13, CLCN7, also revealedno pathogenic mutation.

472 RAMIŁ REZ ETAL.

Testing of Candidate Loci

As the limited size of Family R does not allow for asystematic linkage analysis, we chose a candidate locusapproach to search for homozygosity in syntenic regionsfor osteopetrotic mouse models. We first tested thefollowing loci harboring the orthologs of mouse osteope-trosis genes: SRC on human chromosome 20q11, FOS on14q24, CSF1 on 1p13, and OPGL (coding for RANKL)on 13q14. None of these loci showed homozygosity bydescent, as depicted by the microsatellite haplotypes inFigure 1A–D.

The spontaneous osteopetrotic mutant mouse grey-lethal (gl) gene had been mapped to mouse chromosome10 between the gene Fyn and D10Mit148, the latter ofwhich is located between the genes Apg5l and Bves[Chalhoub et al., 2001]. We tested the syntenic humanregion on chromosome 6q21. As seen in Figure 1E, wefound a homozygous region spanning at least 20 cMbetween markers D6S1717 and D6S1608 on chromo-some 6q21, which encompasses the entire murine gllocus. It therefore seemed likely that the osteopetrosis inFamily R is caused by a mutation in the ortholog of themurine Gl gene.

Candidate GeneApproach and Mutation Analysis

We defined the critical region for positional cloningwith respect to the published mouse mapping data, i.e.,to lie between Fyn and D10Mit148 (located betweenApg5l and Bves). A detailed comparison between thehuman and mouse genomic regions revealed a partial

inversion of the extended human locus with respect tothe mouse locus (Fig. 2). The centromeric ‘‘breakpoint’’of this inversion on human chromosome 6 lies before thegene SIM1 (single-minded (Drosophila) homolog 1) at100.7 Mb and the telomeric breakpoint is at a position117 Mb after a gene called LOC221302. Due to thisinversion, the human critical region is supposed to belocated centromeric to FYN. Structural analyses ofosteoclasts in the gl mouse showed an apparent defectin the structure of the cytoskeleton [Rajapurohitam et al.,2001], which prompted us to first focus our positionalcandidate approach on genes that are known orpredicted to play a role in the organization of thecytoskeleton. We amplified and directly sequenced morethan 10 positional candidate genes of the region,including the genes APG5L, SNX3, WASF1, AIM1,LACE1, and MICAL (data not shown), yet we did notfind any disease-causing mutation. While we conductedthe positional candidate gene approach, the grey-lethalgene (OSTM1 alias GL) was independently cloned. Itconsists of six exons and five introns and spans anapproximately 23-kb genomic sequence [Chalhoub et al.,2003]. Based on the published and public domain data,we performed a mutation analysis in all six exons and thecorresponding intron/exon boundaries of the grey-lethalgene. The analysis of genomic DNA of the patientrevealed a homozygous 2-bp deletion at the beginning ofexon 2 (c.415_416delAG) (Fig. 3). The deletion leads toa premature stop codon at position 150 after 11extraneous amino acids that are translated from thewrong reading frame (p.S139fsX12). As expected, both

Marker/Gene

D14S277D14S77

FOSD14S76D14S61

2 43 6

1 54 3

1 25 2

2 32 4

4 16 5

5 23 2

FOS - 14q24

Mb66.867.369.569.570.1

(A)

Marker/Gene

D1S248D1S2778CSF1

D1S2809D1S252

2 42 4

2 33 5

2 23 4

4 31 3

4 24 3

3 45 1

CSF1 - 1p13

Mb106.1108.0109.4110.2116.3

(C)

MarkerD6S1717D6S1580D6S268D6S1594D6S302D6S261D6S303D6S287D6S1608

1 53 33 22 15 51 21 11 11 2

5 53 52 31 15 42 11 21 42 1

5 53 32 21 15 52 21 11 12 2

gl - 6q21

Mb 99.6103.4107.7108.5112.0114.1116.0119.5120.8

(E)

gl

Marker/Gene

D20S884SRC

D20S85

5 1

2 1

5 6

2 4

1 5

1 2

SRC - 20q11

Mb35.635.737.7

(B)

Marker/Gene

D13S1253D13S1297OPGL

D13S1312

5 42 2

4 4

5 34 3

1 4

4 52 4

4 1

OPGL - 13q14

Mb34.137.137.139.9

(D)

FIGURE 1. Linkageanalysis for selectedcandidate loci inFamilyR.A^D:Haplotypeanalysis for theFOS, SRC,CSF1, andOPGL locus,respectively. E:Haplotype analysis of a regionof chromosome 6q21,which is syntenic to themouse gl locus.The nameof themarkeris given in the left column; the map location in megabases (Mb) with respect to the UCSC Human Genome Browser (http://geno-me.ucsc.edu) is given in the second column.The approximate location of the mouse gl locus is shown by the bracket on the right.The boxed haplotype in (E) indicates the homozygous region by descent,which spans at least 20 Mb.

OSTM1MUTATION IN ARO 473

parents were heterozygous for this deletion. We did notfind the 2-bp deletion in 100 control chromosomes ofGerman origin. The screening of six additional sporadic

patients with autosomal recessive infantile malignantosteopetrosis without mutations in CLCN7 and TCIRG1revealed no further mutations in the OSTM1 gene.

CCNC

100.0

Sim1

51.1

Ccnc

21.3

PRDM13

100.0

Prdm13

21.2

GRIK2

102.1

Grik2

49.6

BVES

105.5

Bves

45.4

APG5L

106.7

Apg5l

44.4

OSTM1

108.4

Ostm1

42.7

WASF1

110.4

Wasf1

41.0

FYN

112.0

Fyn

39.5

COL10A1

116.4

Col10a1

34.3

LOC221302

116.9

Loc221302

33.9

ROS1

117.7

Ros1

52.3

PIST

117.9

Pist

52.5

GJA1

121.7

Gja1

56.5

HDAC2

114.3

Hdac2

37.0

SIM1

100.9

Chr. 6q16-q22_human(UCSC Genome Browser Nov. 2002)

cen. tel.

Chr. 10_mouse(UCSC Genome Browser Feb. 2002)

Chr. 4_mouse

FIGURE 2. Schematic representationof the grey lethal locus and the human syntenic regionon chromosome 6q.The gl critical regiononmousechromosome10 is limited by thegeneFyn and themousemarkerD10Mit148,which is located between thegenesApg5l andBves.The nameof selected genes of the region is given above the bars, the relative location as taken from theUCSCGenomeBrowseris given below the bars.The relative inversion of a subregion of mouse chromosome10 with respect to the conserved syntenic humanregion on chromosome 6q is depicted by the crossed lines in the lower part.

FIGURE 3. A: Sequenceanalysis of the 2-bpdeletion in theOSTM1gene inmembers ofFamilyR. As acomparison, acontrol sequenceis shown on the left. Both parents are heterozygous for the same mutation in this exon (only the father is shown).The daughter ishomozygous for anAG base pair deletion in a stretch of three consecutiveAGs (c.415_416delAG), predicting a frameshift that entailsa premature stop codon11 amino acids after the deletion at position 150 (p.S139fsX12), mutation numbering is based on the Gen-Bankentrywith the accessionnumberNM_014028.2, position +1of thecDNAcorresponds to theAof theATG start codon.Thevirtualtranslation is given in the three letter code. B: Protein alignment between the wild-type (WT) and the predicted protein sequence ofthe deletion. In blue is the identical region of both sequences, in red the extraneous amino acids of the wrong reading frame, thecorrect reading frame of the protein is in black, the putative transmembrane domain (TM) is underlaid.

474 RAMIŁ REZ ETAL.

DISCUSSION

This work describes a small inbred family withinfantile malignant osteopetrosis (ARO) that isnot caused by mutations in the known human osteope-trosis genes. Following a candidate locus approach,we identified a putative locus on human chromosome6q21. We subsequently found a novel mutationwithin the coding sequence of the OSTM1 gene,which was recently reported to underlie osteopetrosisin the grey-lethal mouse mutant and one humanfamily.

The highest levels of the Ostm1 transcript can bedetected in brain, kidney, spleen, osteoclasts, andmelanocytes. It encodes a protein of 338 amino-acidsthat shows no homology to any known protein. Analysisof the protein sequence suggests a potential cleavage sitefor a signal peptide and one putative transmembranedomain [Chalhoub et al., 2003]. The exact function ofthe Ostm1 protein is elusive. Yet, very recently andindependently, the rat ortholog (named Gipn) has beencloned [Fischer et al., 2003], which may act as a E3ubiquitin ligase. It is therefore reasonable to assume thatOstm1 may be involved in the protein degradationmachinery. Compared to other osteopetrotic mousemodels, it is remarkable that gl osteoclasts are not onlycharacterized by an underdeveloped ruffled border, whichis typical for Atp6i and Clcn7 deficiency as well, butadditionally show a defect in cytoskeletal organization.As can be expected, the cells have a strongly reducedresorption activity in vitro. Osteoclast numbers in gl/gltissue sections, however, are higher than in wild-type,suggesting that osteoclast differentiation is basicallyintact, but the terminal maturation into resorbingosteoclasts is disturbed [Rajapurohitam et al., 2001],possibly by an alteration of the proteasome-mediatedprotein degradation.

The mutation found in our patient is a homozygous2-bp deletion in the second coding exon of the OSTM1gene, whereas both asymptomatic parents are hetero-zygous. It is very likely that the consequence is a totalloss of OSTM1 gene function, because the shiftedreading frame results in a premature stop codon thattruncates the protein after 149 amino acids. Hence, theprotein would lack the evolutionarily more conservedC-terminal part, which harbors the putative transmem-brane segment. The human mutation in intron 5identified by Chalhoub et al. [2003], leads to skippingof exon 5 so that a truncated protein is produced thatlacks the putative transmembrane segment as well. Yetfor this splice site mutation, incomplete exon skipping inosteoclasts cannot be excluded, thereby leading to aresidual activity of the Ostm1 protein. In contrast, theframeshift mutation in our patient most probably leads toa complete loss of function. The mutation in the grey-lethal mouse consists of a 7.5-kb deletion encompassingthe entire first exon and thus also entails a complete lossof function. Due to the limited number of human AROcases with OSTM1 mutations, no genotype–phenotypecorrelation can be concluded at the moment.

Additionally, we screened six sporadic patients withARO, who had no mutations in CLCN7 or TCIRG1.However, we did not identify any further OSTM1mutations. This is in line with the findings of Chalhoubet al. [2003], who found only one mutation in 19individuals without mutations in CLCN7 and TCIRG1.This suggests that OSTM1 mutations are a rather rarecause of the disease and that further genes are involved.According to our data and to recently published data[Frattini et al., 2003], approximately 60% of ARO casesare due to mutations in TCIRG1, 10 to 15% showCLCN7 mutations, whereas OSTM1 mutations appearto be very rare (1–3%). This means that in approximately25% of cases no mutation can be identified at themoment, which strongly supports the idea of furtherlocus heterogeneity.Thus, our results confirm that OSTM1 is the third

human ARO gene. Mutations in OSTM1 appear to beless frequent than mutations in TCIRG1 and CLCN7.Further studies are needed to disentangle the exactfunction of the Ostm1 protein in osteoclasts andmelanocytes.

ACKNOWLEDGMENTS

We thank the members of Family R.

REFERENCES

Bernard F, Casanova JL, Cournot G, Jabado N, Peake J, Jauliac S,Fischer A, Hivroz C. 1998. The protein tyrosine kinase p60c-Srcis not implicated in the pathogenesis of the human autosomalrecessive form of osteopetrosis: a study of 13 children. J Pediatr133:537–543.

Chalhoub N, Benachenhou N, Vacher J. 2001. Physical andtranscriptional map of the mouse Chromosome 10 proximal regionsyntenic to human 6q16-q21. Mamm Genome 12:887–892.

Chalhoub N, Benachenhou N, Rajapurohitam V, Pata M, FerronM, Frattini A, Villa A, Vacher J. 2003. Grey-lethal mutationinduces severe malignant autosomal recessive osteopetrosis inmouse and human. Nat Med 9:399–406.

Fischer T, De Vries L, Meerloo T, Farquhar MG. 2003. Promotionof G alpha i3 subunit down-regulation by GIPN, a putative E3ubiquitin ligase that interacts with RGS-GAIP. Proc Natl AcadSci USA 100:8270–8275.

Frattini A, Orchard PJ, Sobacchi C, Giliani S, Abinun M, MattssonJP, Keeling DJ, Andersson AK, Wallbrandt P, Zecca L, NotarangeloLD, Vezzoni P, Villa A. 2000. Defects in TCIRG1 subunit of thevacuolar proton pump are responsible for a subset of humanautosomal recessive osteopetrosis. Nat Genet 25:343–346.

Frattini A, Pangrazio A, Susani L, Sobacchi C, Mirolo M, AbinunM, Andolina M, Flanagan A, Horwitz EM, Mihci E, NotarangeloLD, Ramenghi U, Teti A, Van Hove J, Vujic D, Young T, AlbertiniA, Orchard PJ, Vezzoni P, Villa A. 2003. Chloride channel ClCN7mutations are responsible for severe recessive, dominant, andintermediate osteopetrosis. J Bone Miner Res 18:1740–1747.

Gerritsen EJ, Vossen JM, Fasth A, Friedrich W, Morgan G, PadmosA, Vellodi A, Porras O, O’Meara A, Porta F, Bordigoni P, CantA, Hermans J, Griscelli C, Fischer A. 1994a. Bone marrowtransplantation for autosomal recessive osteopetrosis. A reportfrom the Working Party on Inborn Errors of the European BoneMarrow Transplantation Group. J Pediatr 125:896–902.

OSTM1MUTATION IN ARO 475

Gerritsen EJ, Vossen JM, van Loo IH, Hermans J, Helfrich MH,Griscelli C, Fischer A. 1994b. Autosomal recessive osteopetrosis:variability of findings at diagnosis and during the natural course.Pediatrics 93:247–253.

Grynpas M. 1993. Age and disease-related changes in the mineralof bone. Calcif Tissue Int 53(Suppl 1):S57–S64.

Karsenty G. 1999. The genetic transformation of bone biology.Genes Dev 13:3037–3051.

Keith CG. 1968. Retinal atrophy in osteopetrosis. Arch Ophthal-mol 79:234–241.

Kong YY, Yoshida H, Sarosi I, Tan HL, Timms E, Capparelli C,Morony S, Oliveira-dos-Santos AJ, Van G, Itie A, Khoo W,Wakeham A, Dunstan CR, Lacey DL, Mak TW, Boyle WJ,Penninger JM. 1999. OPGL is a key regulator of osteoclastogenesis,lymphocyte development and lymph-node organogenesis. Nature397:315–323.

Kornak U, Schulz A, Friedrich W, Uhlhaas S, Kremens B, Voit T,Hasan C, Bode U, Jentsch TJ, Kubisch C. 2000. Mutations inthe a3 subunit of the vacuolar H(+)-ATPase cause infantilemalignant osteopetrosis. Hum Mol Genet 9:2059–2063.

Kornak U, Kasper D, Bosl MR, Kaiser E, Schweizer M, Schulz A,Friedrich W, Delling G, Jentsch TJ. 2001. Loss of the ClC-7chloride channel leads to osteopetrosis in mice and man. Cell104:205–215.

Lazner F, Gowen M, Pavasovic D, Kola I. 1999. Osteopetrosis andosteoporosis: two sides of the same coin. Hum Mol Genet8:1839–1846.

Li YP, Chen W, Liang Y, Li E, Stashenko P. 1999. Atp6i-deficientmice exhibit severe osteopetrosis due to loss of osteoclast-mediated extracellular acidification. Nat Genet 23:447–451.

Marks SC Jr, Seifert MF, Lane PW. 1985. Osteosclerosis, arecessive skeletal mutation on chromosome 19 in the mouse.J Hered 76:171–176.

Rajapurohitam V, Chalhoub N, Benachenhou N, Neff L, Baron R,Vacher J. 2001. The mouse osteopetrotic grey-lethal mutationinduces a defect in osteoclast maturation/function. Bone28:513–523.

Scimeca JC, Franchi A, Trojani C, Parrinello H, Grosgeorge J,Robert C, Jaillon O, Poirier C, Gaudray P, Carle GF. 2000. Thegene encoding the mouse homologue of the human osteoclast-specific 116-kDa V-ATPase subunit bears a deletion inosteosclerotic (oc/oc) mutants. Bone 26:207–213.

Soriano P, Montgomery C, Geske R, Bradley A. 1991. Targeteddisruption of the c-src proto-oncogene leads to osteopetrosis inmice. Cell 64:693–702.

Teitelbaum SL. 2000. Bone resorption by osteoclasts. Science289:1504–1508.

Thompson DA, Kriss A, Taylor D, Russell-Eggitt I, Hodgkins P,Morgan G, Vellodi A, Gerritsen EJ. 1998. Early VEP and ERGevidence of visual dysfunction in autosomal recessive osteope-trosis. Neuropediatrics 29:137–144.

Vacher J, Bernard H. 1999. Genetic localization and transmissionof the mouse osteopetrotic grey-lethal mutation. MammGenome 10:239–243.

Walker DG. 1993. Bone resorption restored in osteopetrotic miceby transplants of normal bone marrow and spleen cells. 1975.Clin Orthop Sep(294):4–6.

Wang ZQ, Ovitt C, Grigoriadis AE, Mohle-Steinlein U, Ruther U,Wagner EF. 1992. Bone and haematopoietic defects in micelacking c-fos. Nature 360:741–745.

476 RAMIŁ REZ ETAL.