Familial tumoral calcinosis and hyperostosis–hyperphosphataemia syndrome are different...

6
SCIENTIFIC ARTICLE Familial tumoral calcinosis and hyperostosishyperphosphataemia syndrome are different manifestations of the same disease: novel missense mutations in GALNT3 Leo Joseph & Sandra N. Hing & Nadege Presneau & Paul ODonnell & Tim Diss & Bernadine D. Idowu & Selvanayagam Joseph & Adrienne Margaret Flanagan & David Delaney Received: 27 May 2009 / Revised: 27 August 2009 / Accepted: 20 September 2009 / Published online: 15 October 2009 # ISS 2009 Abstract Objective To report on the biochemistry and clinical and genetic findings of two siblings, the younger sister presenting with recurrent bone pain of the radius and ulna, and medullary sclerosis, and the older brother with soft tissue calcific deposits (tumoral calcinosis) but who later developed bone pain. Both were found to be hyperphosphaturic. Materials and methods The index family comprised four individuals (father, mother, brother, sister). The affected siblings were the offspring of a non-consanguineous Indian family of Tamil origin. Bidirectional sequencing was performed on the DNA from the index family and on 160 alleles from a population of 80 unrelated unaffected control individuals of Tamil extraction and 72 alleles from individuals of non-Tamil origin. Results Two symptomatic siblings were found to harbour previously unreported compound heterozygous missense UDP-N-acetyl-D-galactosamine: polypeptide N- acetylgalactosaminyltransferase 3 (GalNAc-transferase; GALNT3) mutations in exon 4 c.842A>G and exon 5 c.1097T>G. This sequence variation was not detected in the control DNA. This is the first report of siblings exhibiting stigmata of familial tumoral calcinosis and hyperostosishyperphosphataemia syndrome with docu- mented evidence of autosomal recessive missense GALNT3 mutations. Conclusion The findings from this family add further evidence to the literature that familial tumoral calcinosis and hyperostosishyperphosphataemia syndrome are man- ifestations of the same disease and highlight the importance of appropriate metabolic and genetic investigations. Keywords Bone neoplasms . Genetics . Mutation . Musculoskeletal diseases . GALNT3 Leo Joseph and Sandra N. Hing contributed equally to this paper L. Joseph : S. Joseph Department of Orthopaedic Surgery, Vinodhagan Memorial Hospital and Dr. Josephs Ortho Clinic, Thanjavur, India S. N. Hing : B. D. Idowu : A. M. Flanagan : D. Delaney Department of Histopathology, Royal National Orthopaedic Hospital NHS Trust, Stanmore, Middlesex, UK P. ODonnell Department of Radiology, Royal National Orthopaedic Hospital NHS Trust, Stanmore, Middlesex, UK N. Presneau : A. M. Flanagan Cancer Institute, University College London (UCL), London, UK T. Diss : A. M. Flanagan Department of Histopathology, University College London Hospital (UCLH) NHS Trust, Rockefeller Building, London, UK P. ODonnell : A. M. Flanagan Institute of Orthopaedics and Musculoskeletal Science, University College London (UCL), Stanmore, UK A. M. Flanagan (*) Institute of Orthopaedics and Musculoskeletal Science, Stanmore, Middlesex HA7 4LP, UK e-mail: [email protected] P. ODonnell The Institute of Orthopaedics and Musculoskeletal Science, University College London (UCL), London, UK Skeletal Radiol (2010) 39:6368 DOI 10.1007/s00256-009-0808-5

Transcript of Familial tumoral calcinosis and hyperostosis–hyperphosphataemia syndrome are different...

Page 1: Familial tumoral calcinosis and hyperostosis–hyperphosphataemia syndrome are different manifestations of the same disease: novel missense mutations in GALNT3

SCIENTIFIC ARTICLE

Familial tumoral calcinosis and hyperostosis–hyperphosphataemiasyndrome are different manifestations of the same disease:novel missense mutations in GALNT3

Leo Joseph & Sandra N. Hing & Nadege Presneau & Paul O’Donnell & Tim Diss &

Bernadine D. Idowu & Selvanayagam Joseph & Adrienne Margaret Flanagan &

David Delaney

Received: 27 May 2009 /Revised: 27 August 2009 /Accepted: 20 September 2009 /Published online: 15 October 2009# ISS 2009

AbstractObjective To report on the biochemistry and clinical andgenetic findings of two siblings, the younger sister presentingwith recurrent bone pain of the radius and ulna, andmedullary sclerosis, and the older brother with soft tissuecalcific deposits (tumoral calcinosis) but who later developedbone pain. Both were found to be hyperphosphaturic.Materials and methods The index family comprised fourindividuals (father, mother, brother, sister). The affectedsiblings were the offspring of a non-consanguineous Indianfamily of Tamil origin. Bidirectional sequencing wasperformed on the DNA from the index family and on 160alleles from a population of 80 unrelated unaffected controlindividuals of Tamil extraction and 72 alleles fromindividuals of non-Tamil origin.Results Two symptomatic siblings were found toharbour previously unreported compound heterozygous

missense UDP-N-acetyl-D-galactosamine: polypeptide N-acetylgalactosaminyltransferase 3 (GalNAc-transferase;GALNT3) mutations in exon 4 c.842A>G and exon 5c.1097T>G. This sequence variation was not detected inthe control DNA. This is the first report of siblingsexhibiting stigmata of familial tumoral calcinosis andhyperostosis–hyperphosphataemia syndrome with docu-mented evidence of autosomal recessive missenseGALNT3 mutations.Conclusion The findings from this family add furtherevidence to the literature that familial tumoral calcinosisand hyperostosis–hyperphosphataemia syndrome are man-ifestations of the same disease and highlight the importanceof appropriate metabolic and genetic investigations.

Keywords Bone neoplasms . Genetics . Mutation .

Musculoskeletal diseases . GALNT3

Leo Joseph and Sandra N. Hing contributed equally to this paper

L. Joseph : S. JosephDepartment of Orthopaedic Surgery,Vinodhagan Memorial Hospital and Dr. Joseph’s Ortho Clinic,Thanjavur, India

S. N. Hing : B. D. Idowu :A. M. Flanagan :D. DelaneyDepartment of Histopathology,Royal National Orthopaedic Hospital NHS Trust,Stanmore, Middlesex, UK

P. O’DonnellDepartment of Radiology,Royal National Orthopaedic Hospital NHS Trust,Stanmore, Middlesex, UK

N. Presneau :A. M. FlanaganCancer Institute, University College London (UCL),London, UK

T. Diss :A. M. FlanaganDepartment of Histopathology, University College LondonHospital (UCLH) NHS Trust, Rockefeller Building,London, UK

P. O’Donnell :A. M. FlanaganInstitute of Orthopaedics and Musculoskeletal Science,University College London (UCL),Stanmore, UK

A. M. Flanagan (*)Institute of Orthopaedics and Musculoskeletal Science,Stanmore, Middlesex HA7 4LP, UKe-mail: [email protected]

P. O’DonnellThe Institute of Orthopaedics and Musculoskeletal Science,University College London (UCL),London, UK

Skeletal Radiol (2010) 39:63–68DOI 10.1007/s00256-009-0808-5

Page 2: Familial tumoral calcinosis and hyperostosis–hyperphosphataemia syndrome are different manifestations of the same disease: novel missense mutations in GALNT3

Introduction

Familial tumoral calcinosis (FTC) and hyperostosis–hyperphosphataemia syndrome (HHS) are rare autosomalrecessive metabolic disorders characterised by similarbiochemical abnormalities, but, generally, they exhibitdifferent musculoskeletal phenotypes. FTC typically occursin the first decade of life and is characterised by persistentsoft tissue peri-articular deposits of calcium, which can becomplicated by skin and bone infections, scarring anddeformity [1, 2]. The condition is also associated withdental abnormalities and retinal angiomatoid streaks. HHScan also present in the first decade, results in transient butrecurrent pain, and may be associated with swelling of longbones, typically the tibia, which lasts between weeks andmonths and responds to non-steroidal anti-inflammatorydrugs [3]. Imaging of HHS typically shows a periostealreaction and cortical hyperostosis or, in some cases,intramedullary sclerosis [4]. In the rare reported casesthat have been biopsied the histological features arenon-specific and include anastomosing trabeculae ofwoven bone lined by plump osteoblasts [4]. Thus, theclinical presentation of HHS can mimic osteomyelitis,while the radiological and histological differentials includeosteomyelitis, osteoid osteoma/osteoblastoma, and somesclerosing bone dysplasias [5].

In contrast to the different phenotypic stigmata generallyseen in FTC andHHS, themetabolic abnormalities of these twoentities are similar [1–3] and include persistently, generallysignificantly, raised serum phosphate, inappropriately normalor only slightly elevated 1,25-dihydroxy-vitamin D3, normalcalcium and alkaline phosphatase, low/low normal parathy-roid hormone levels, and low intact fibroblast growth factor 23(FGF23) protein levels. The biochemical overlap seen in FTCand HHS, and the report that the same UDP-N-acetyl-D-galactosamine: polypeptide N-acetylgalactosaminyltransferase3 (GalNAc-transferase; GALNT3) mutation (1524+1G>A)was identified in two unrelated individuals, one whodeveloped FTC 6 years after the onset of HHS and one withonly HHS [6], provides evidence that these phenotypesrepresent a spectrum of the same disease. This concept issupported by the finding that an FGF23 mutation (c.211A>G)has been detected in a patient with clinical features of bothFTC and HHS [7]. Other recessive FGF23 mutations havealso been reported in individuals with tumoral calcinosis (TC)alone [8, 9]. However, the stigmata of both FTC and HHS,occurring in the same individuals with documented evidenceof either FGF23 or GALNT3 mutations, have only beendescribed twice [6, 7]. Furthermore, the occurrence of FTC inone sibling and HHS in the other has only been reported onceand was not confirmed by mutational analysis [10]. Anoverview of the genetics of familial tumoral calcinosis can beobtained in the recent review [11].

FGF23 and GALNT3 are both key regulators ofphosphate metabolism. Gain of function mutations inFGF23 result in hypophosphataemic rickets as a result ofFGF23 not being degraded, the net result of which isincreased biological activity of the protein and hypophos-phataemia [12], whereas inactivating mutations in FGF23have been shown to result in instability of the protein ratherthan failure of protein secretion [8]. Excessive production ofthe protein by neoplasms also induces tumour-induced/oncogenic osteomalacia [13]. Biologically active FGF23requires the presence of three O-linked glycans, and thisprocess is brought about by GALNT3. Hence GALNT3inactivating mutations result in production of inactive FGF23and consequently hyperphosphataemia [14]. The interdepen-dence of FGF23 and GALNT3, and the explanation as tohow mutations in either gene give rise to similar phenotypes,is highlighted by the finding that low levels of intact serumFGF23 in HHS occur in the presence of inactivatingGALNT3 mutations [15].

Materials and methods

Patient and control individuals

A single non-consanguineous Indian family of Tamil origin,comprising four individuals (father, mother, brother andsister) including two affected individuals (brother andsister), was studied. In addition 160 alleles from apopulation of 80 unrelated unaffected control individualsof Tamil extraction and 72 alleles from individuals of non-Tamil origin were also analysed. The family providedwritten consent for publication of this article, and the studycomplied with the standards of the Central Office forResearch Ethics Committee.

Mutational analysis

Genomic DNA was extracted by conventional methodsfrom peripheral blood of the family members of theproband and from buccal scrapes of 80 individuals ofTamil ethnicity and 36 individuals of non-Tamil ethnicity.

Bidirectional sequencing of DNA from family members

Primer sequences for exons 1–10 of the GALNT3 genewere kindly provided by Anna Barbieri, Italy (Table 1).The polymerase chain reaction (PCR) mixture contained50 ng of genomic DNA, 10.0 pmol of each primer,200 μM deoxyribonucleotide triphosphates (dNTPs) and 1unit of DNA polymerase (Amplitaq Gold, Applied Bio-systems Inc., San Diego, USA) in 1x DNA polymerasereaction buffer [containing between 1.0 mM and 2.5 mM

64 Skeletal Radiol (2010) 39:63–68

Page 3: Familial tumoral calcinosis and hyperostosis–hyperphosphataemia syndrome are different manifestations of the same disease: novel missense mutations in GALNT3

magnesium chloride (MgCl2)]. A touchdown PCR proto-col was employed, with annealing temperature reduced by1°C per cycle from 65°C to 56°C followed by 35 furthercycles at 56°C. Each cycle was carried out as follows: 93°C for 45 s, annealing for 45 s, 72°C for 1 min 30 s. Thepolymerase was activated at 95°C for 15 min beforecycling, and a final extension at 72°C was performed for10 min. The PCR products were purified through apurification kit (Qiagen, Crawley, West Sussex, UK). Amaster mix was made up of 10 ng of the purified DNA,5 pmol of either forward or reverse primer, 4 μl CEQ dyeterminator cycle sequencing (DTCS), and was made up toa final volume of 10 μl with distilled water for thesequencing cycling reaction, which underwent 30 cycles at96°C for 20 s, 50°C for 45 s and 60°C for 4 min. Thecycling products were purified with ethylene diaminetetra-acetic acid (EDTA), sodium acetate and glycogenwith ethanol. The pellets were washed in 70% ethanoland air-dried at room temperature prior to beingresuspended in 40 μl sample loading buffer. Each casewas sequenced in both directions using the DTCS QuickStart kit (Beckman Coulter UK Ltd., Bucks, UK) withanalysis on a CEQ™8000 DNA genetic analysis system(Beckman Coulter). Amino acid substitutions wereanalysed for their predicted effect on protein functionusing the Sorting Intolerant From Tolerant (SIFT)software program [16, 17].

Polymerase chain reaction and restriction digest mutationalanalysis of control population

Part of exon 4 of GALNT3 was amplified using primers 5-aaattctgttccctccaggtg-3 and 5-ggacttacgacagccgtgtagtt-3,which span the mutation site. The products were digestedwith BslI (New England Biolab, Hitchin, Herts, UK) for3 h at 55°C and run on 8% polyacrylamide gels. Themutation creates a BslI site so that a band of 70 bp is seenin addition to the 91 bp fragment in individuals carryingthe mutation.

Part of exon 5 of GALNT3 was amplified using primers5-ttcagaacacccacttttgc-3 and 5- ccccagatttccatttcttca-3,which span the mutation site. The products were digestedwith HpyCH41V (New England Biolab) for 3 h at 37°Cand run on 8% polyacrylamide gels. The mutation createsan HpyCH41V site so that a band of 67 bp is seen inaddition to the 95 bp fragment in individuals carrying themutation.

Results

Clinical findings

The proband, a 13-year-old girl, presented with recurrentbilateral ulnar pain and tenderness and complained ofsimilar symptoms in both tibiae 9 months later. Imaging ofthe bones involved showed diffuse diaphyseal medullarysclerosis, a variable periosteal reaction and no significantcortical thickening (Fig. 1). Magnetic resonance imaging(MRI) of the bones showed medullary and periosteal highsignal intensity, suggesting oedema (Fig. 2); high uptakewas identified in the involved bones and in both calcanei onisotope bone scanning during the symptomatic period(Fig. 3). Although in the first instance the diagnosis ofmultiple osteoid osteomas was raised, following review ofthe imaging, this diagnosis was discounted. Curettage andbiopsy of the ulnar and tibial lesions, respectively, wereperformed, and the histopathological findings showed non-specific features that included dense woven compact/cortical and trabecular bone. Further questioning revealedthat the proband’s older brother had developed calcifiednodules in the buttock at the age of 2 years and later atother sites, but skeletal imaging had not been undertaken atthat time. At the age of 9 years he developed pain in bothtibiae, and the surgeons made several burr holes in thetibiae, reporting that the bones had a ‘thick sclerosedmedullary cavity’, but no biopsy was taken. The youngman has been asymptomatic for the past 4 years (he is nowaged 19 years). The girl has never had signs of FTC.

Both siblings had persistent hyperphosphataemia(6.9 mg/dl). The father and mother had no symptoms, but

Table 1 Primer sequences for exons 1–10 of GALNT3

Exons Primers Product size

Exon 1 5′–3′ ccatcgatcatttctgtttatagg 718 bp

Exon 1 3′–5′ tccttagctcacccctctctc

Exon 2 5′–3′ ggtgagtgatttgcttgtaaaaa 469 bp

Exon 2 3′–5′ caagctctgagatggcataca

Exon 3 5′–3′ cattttgctggaaggacaca 447 bp

Exon 3 3′–5′ ctgttacctgcttgggctgt

Exon 4 5′–3′ tctgaggaagaaagaaatctcca 587 bp

Exon 4 3′–5′ gagctcactcactgctacctctt

Exon 5 5′–3′ caatgggagaggacacgaag 385 bp

Exon 5 3′–5′ accagccgattagaacacaa

Exon 6 5’–3’ atggcaggggacagagacta 416 bp

Exon 6 3′–5′ atgaatcgacgcaaaaggac

Exon 7/8 5′–3′ ggctgttgaattgcctcttg 637 bp

Exon 7/8 3′–5′ aggcaacatctcacttgtgct

Exon 9 5′–3′ aaccacctgttgatgaaggaa 429 bp

Exon 9 3′–5′ tgttccactcattttcccaga

Exon 10 5′–3′ tcagacatggctcaccttagaa 351 bp

Exon 10 3′–5′ tttagctgcttttgcataattttc

Skeletal Radiol (2010) 39:63–68 65

Page 4: Familial tumoral calcinosis and hyperostosis–hyperphosphataemia syndrome are different manifestations of the same disease: novel missense mutations in GALNT3

the father had slightly raised phosphate levels [5.4 mg/dl(the upper limit of normal is 5.0 mg/dl)], and there was nofamily history of either soft tissue tumours or bone lesionsin the extended family.

Mutational analysis

Direct DNA sequencing of GALNT3 revealed two sequencevariations which were present in both the proband and herbrother:

(1) in exon 4, a transition of adenine to guanine atcodon 281 (c.842A>G), resulting in an amino acidchange of glutamic acid (E), a polar and negativelycharged amino acid, to glycine (G), a non-polar andneutral amino acid, and (2) in exon 5, a transversion ofthymidine to guanine at codon 366 (c.1097T>G.),resulting in an amino acid change of leucine (L), anon-polar neutral amino acid, to arginine (R), a polarpositively charged amino acid. The father was hetero-zygous for the exon 4 sequence variation and themother was heterozygous for the exon 5 sequencevariation.

Fig. 3 Technetium-99m isotope bone scan. a Planar anterior image ofthe whole body shows increased activity in the left ulna and bothcalcanei; b lateral image of the left foot shows increased activity in thecalcaneum (similar appearances were seen on the right). Normal highactivity is seen in the distal tibial physis

Fig. 2 Axial T2-weighted MR image with fat saturation of the righttibia. Hyperintensity consistent with oedema is seen in the tibialmarrow and circumferentially on the bone surface, consistent withperiosteal reaction

Fig. 1 a Lateral radiograph of the right tibia, showing diffusemedullary sclerosis; b lateral radiograph of the left elbow, showingmedullary sclerosis of the ulna and a unilamellar periosteal reaction.Lucency in the proximal ulna is secondary to previous curettage

66 Skeletal Radiol (2010) 39:63–68

Page 5: Familial tumoral calcinosis and hyperostosis–hyperphosphataemia syndrome are different manifestations of the same disease: novel missense mutations in GALNT3

All control material (160 alleles and 72 alleles fromindividuals of Tamil and non-Tamil ethnicity, respectively)was wild-type for exon 4 and exon 5 GALNT3 variations.Furthermore, the GALNT3 sequence variations in theparents and children were not found in any of the followingdatabases: http://www.sanger.ac.uk/genetics/CGP/cosmic/;http://www.ncbi.nlm.nih.gov/projects/SNP/; http://www.hgmd.cf.ac.uk/ac/index.php. Finally, analysis of the substi-tution at position 281 from E to G (exon 4) and at position366 from L to R (exon 5) using SIFT software gave a scoreof 0.00 for predicting the affect on protein function. Thesoftware predicts that substitutions with a score of less than0.05 will affect protein function [16, 17].

Discussion

This is the first report where compound heterozygousGALNT3 mutations have been identified in two siblingsof whom one exhibits stigmata of HHS and the othershows features of FTC. Furthermore, both mutationsreported herein have not been previously reported.However, there has been a previous report in whichsiblings are described as exhibiting a ‘mixed’ phenotype(one with FTC and one with HHS), but genotype analysiswas not performed in this family [10]. This is also onlythe fourth report in which an individual has presentedwith stigmata of both HHS and FTC, but in only two ofthese reports had mutations been documented previously(one GALNT3 and one FGF23) [6, 7]. Hence, the geneticand phenotypic findings in this family add to the limitedpublished evidence that FTC/peri-articular soft tissuecalcium deposits and HHS/cortical and or medullarysclerosis represent different manifestations of the samedisease. It is interesting to speculate that the stigmata ofHHS and FTC occur together more commonly than haspreviously been thought, as it is likely that sclerosingbone disease is not sought in patients with FTC becausethe symptoms are often transient and non-specific andoften are not directly temporally related to the episodes ofcalcification.

The identification of the two GALNT3 sequence varia-tions in this report, both of which result in amino acidchange, when inherited together are associated with thephenotype described. Of note is that these were never foundin the control cohort or in databases that we searched.Furthermore, both mutations were non-synonymous,occurred in the coding regions, and involved alteration ofthe polarity and the charge of the substituted amino acids.The substituted amino acid in exon 4 is located on apotentially highly flexible loop between two sub-catalyticdomains, which is a high risk predictor for disruption ofprotein conformation and, consequently, function. The

evidence that the mutations are implicated in disease isalso supported by the result generated when the alteredsequence was analysed with SIFT software, which predictsif the changes will be deleterious by studying theconservation of the amino acid implicated across theprotein family members [16, 17]. In view of the strongevidence that the exon 4 and exon 5 mutations alter proteinfunction and, therefore, are highly likely to be implicated indisease, we did not screen for FGF23 mutations. Never-theless, it would be of interest to know how these mutationsalter the function of the protein, but this would require aseries of biochemistry experiments, which we considered tobe beyond the scope of this study.

The finding that mutations in FGF23 or GALNT3resulted in similar phenotypes is easily explained, giventheir close physiological relationship with respect tophosphate metabolism [14]. It is therefore not surprisingthat mutations in Klotho have also been found to beassociated with FTC [15], as Klotho is involved inphosphate metabolism by regulating FGF23 protein expres-sion by binding to a specific FGF receptor [18]. However,the varied phenotype, including stigmata of FTC and orHHS, within and across family members, is not understood.In the absence of environmental differences that couldaccount for different phenotypes in our two affectedsiblings, it is possible that the different phenotypes mightbe related to either inherited or epigenetic modifying factorsresulting from compensation by other ppGa1NAc trans-ferases present in skin and bone [6].

FTCmust be differentiated from other forms of ectopic softtissue calcification that can occur as a result of inflammation(dystrophic) or as a result of abnormalities with serum calcium(metastatic). The latter can occur in the setting of hyper-calcaemia secondary to renal failure, hyperparathyroidism,malignancy (plasma cell neoplasms) or sarcoidosis [5]. Softtissue calcification is also seen as a result of mutations insterile alpha motif domain 9 (SAMD9) a gene that regulatesendothelial cell function and fibroblasts; however, this is notassociated with hyperphosphataemia [19].

HHS can mimic infectious or non-bacterial osteitis, asclerosing bony dysplasia, multifocal osteoid osteomas, oreven a malignant neoplasm. Once a neoplastic process hasbeen excluded, osteomyelitis is probably highest on the listof differential diagnoses, and this can lead to inappropriateantimicrobial therapy for extended periods. Multiple/multifocal osteoid osteomas are extremely rare and usuallysited in the same bone, and, therefore, this diagnosis does notcorrelate with typical features of HHS. In contrast, sclerosingbone dysplasias are a group of poorly defined radiologicalentities, some of which are characterised by hyperostosis andsclerosis, although these are not generally associated withhyperphosphataemia [5]. It is therefore interesting tospeculate that HHS may account for some of the poorly

Skeletal Radiol (2010) 39:63–68 67

Page 6: Familial tumoral calcinosis and hyperostosis–hyperphosphataemia syndrome are different manifestations of the same disease: novel missense mutations in GALNT3

classified bone dysplasias, such as Ribbing’s disease, wherethe bone biochemistry has not been studied [20].

The transient and recurrent nature of HHS means that thediagnosis is at risk of being delayed for long periods,leaving both patients and clinicians dissatisfied. Hence,awareness of the genetic, biochemical and phenotypicoverlap of HHS and FTC, and that both phenotypesrepresent different clinical stigmata of the same disease,should increase diagnostic accuracy and allow the diagnosisto be reached rapidly. The variable clinical findingsreported in the affected children in this family underscorethe need for a detailed clinical history which focuses onwhether family members have, or have experienced, peri-articular calcification, skin scarring and ulceration and,particularly, bone pain. The report also emphasises the needfor serum bone biochemistry to be analysed, particularlyphosphate. This simple clinical and blood chemistryapproach should provide or exclude a diagnosis of HHSin most cases, reduce the need for bone biopsy, and informthe primary clinician and radiologist that genetic studiesshould be considered.

Acknowledgements We are grateful to the families for participatingin this research and to all those in Dr Joseph’s Ortho clinic who wereinvolved in caring for these patients. The research was generouslyfunded by Skeletal Cancer Action Trust (SCAT), UK. The work wasalso supported by the University College London Hospital/UniversityCollege London (UCLH/UCL) Comprehensive Biomedical ResearchCentre. B.D.I. was funded by the Royal National OrthopaedicHospital (RNOH) National Health Service (NHS) Trust. UCL is apartner of the EuroBoNeT consortium, a European Commissiongranted Network of Excellence for studying the pathology andgenetics of bone tumours.

References

1. Palmer PE. Tumoural calcinosis. Br J Radiol. 1966;39:518–25.2. Prince MJ, Schaeffer PC, Goldsmith RS, Chausmer AB. Hyper-

phosphatemic tumoral calcinosis: association with elevation ofserum 1, 25-dihydroxycholecalciferol concentrations. Ann InternMed. 1982;96:586–91.

3. Mikati MA, Melhem RE, Najjar SS. The syndrome of hyperos-tosis and hyperphosphatemia. J Pediatr. 1981;99:900–4.

4. Ichikawa S, Imel EA, Kreiter ML, Yu X, Mackenzie DS, SorensonAH, et al. A homozygous missense mutation in human KLOTHOcauses severe tumoral calcinosis. J Clin Invest. 2007;117:2684–91.

5. Vigorita VJ. Orthopaedic pathology. 2nd ed. Philadelphia:Lippincott, Williams, Wilkins; 2008. pp. 73–90.

6. Frishberg Y, Topaz O, Bergman R, Behar D, Fisher D, Gordon D,et al. Identification of a recurrent mutation in GALNT3demonstrates that hyperostosis-hyperphosphatemia syndrome andfamilial tumoral calcinosis are allelic disorders. J Mol Med.2005;83:33–8.

7. Benet-Pages A, Orlik P, Strom TM, Lorenz-Depiereux B. AnFGF23 missense mutation causes familial tumoral calcinosis withhyperphosphatemia. Hum Mol Genet. 2005;14:385–90.

8. Garringer HJ, Malekpour M, Esteghamat F, Mortazavi SM, DavisSI, Farrow EG, et al. Molecular genetic and biochemical analysesof FGF23 mutations in familial tumoral calcinosis. Am J PhysiolEndocrinol Metab. 2008;295:E929–37.

9. Masi L, Gozzini A, Franchi A, Campanacci D, Amedei A,Falchetti A, et al. A novel recessive mutation of fibroblast growthfactor-23 in tumoral calcinosis. J Bone Joint Surg Am. 2009;91:1190–8.

10. Narchi H. Hyperostosis with hyperphosphatemia: evidence offamilial occurrence and association with tumoral calcinosis.Pediatrics. 1997;99:745–8.

11. Ichikawa S, Imel EA, Econs MJ. Genetics of familial tumoralcalcinosis. Am J Kidney Dis. 2009;53:563–4.

12. ADHR Consortium. Autosomal dominant hypophosphataemicrickets is associated with mutations in FGF23. Nat Genet.2000;26:345–8.

13. Shimada T, Mizutani S, Muto T, Yoneya T, Hino R, Takeda S, etal. Cloning and characterization of FGF23 as a causative factor oftumor-induced osteomalacia. Proc Natl Acad Sci U S A.2001;98:6500–5.

14. Fukumoto S, Yamashita T. FGF23 is a hormone-regulatingphosphate metabolism—unique biological characteristics ofFGF23. Bone. 2007;40:1190–5.

15. Ichikawa S, Guigonis V, Imel EA, Courouble M, Heissat S,Henley JD, et al. Novel GALNT3 mutations causing hyperostosis-hyperphosphatemia syndrome result in low intact fibroblastgrowth factor 23 concentrations. J Clin Endocrinol Metab.2007;92:1943–7.

16. Kumar P, Henikoff S, Ng PC. Predicting the effects of codingnon-synonymous variants on protein function using the SIFTalgorithm. Nat Protoc. 2009;4:1073–81.

17. Ng PC, Henikoff S. SIFT: Predicting amino acid changes thataffect protein function. Nucleic Acids Res. 2003;31:3812–4.

18. Drueke TB, Prie D. Klotho spins the thread of life—what doesKlotho do to the receptors of fibroblast growth factor-23(FGF23)? Nephrol Dial Transplant. 2007;22:1524–6.

19. Topaz O, Indelman M, Chefetz I, Geiger D, Metzker A, AltschulerY, et al. A deleterious mutation in SAMD9 causes normophos-phatemic familial tumoral calcinosis. Am J Hum Genet.2006;79:759–64.

20. Ziran N, Hill S, Wright ME, Kovacs J, Robey PG, Wientroub S, etal. Ribbing disease: radiographic and biochemical characteriza-tion, lack of response to pamidronate. Skeletal Radiol. 2002;31:714–9.

68 Skeletal Radiol (2010) 39:63–68