Giant cell tumor occurring in familial Paget's disease of bone: Report of clinical characteristics and linkage analysis of a large pedigree


  • The first two authors contributed equally to this work.


Neoplastic degeneration represents a rare but serious complication of Paget's disease of bone (PDB). Although osteosarcomas have been described in up to 1% of PDB cases, giant cell tumors are less frequent and mainly occur in patients with polyostotic disease. We recently characterized a large pedigree with 14 affected members of whom four developed giant cell tumors at pagetic sites. The high number of affected subjects across multiple generations allowed us to better characterize the clinical phenotype and look for possible susceptibility loci. Of interest, all the affected members had polyostotic PDB, but subjects developing giant cell tumors showed an increased disease severity with a reduced clinical response to bisphosphonate treatment and an increased prevalence of bone pain, deformities, and fractures. Together with an increased occurrence of common pagetic complications, affected patients of this pedigree also evidenced a fivefold higher prevalence of coronary artery disease with respect to either the unaffected family members or a comparative cohort of 150 unrelated PDB cases from the same geographical area. This association was further enhanced in the four cases with PDB and giant cell tumors, all of them developing coronary artery disease before 60 years of age. Despite the early onset and the severe phenotype, PDB patients from this pedigree were negative for the presence of SQSTM1 or TNFRSF11A mutations, previously associated with enhanced disease severity. Genome-wide linkage analysis identified six possible candidate regions on chromosomes 1, 5, 6, 8, 10, and 20. Because the chromosome 8 and 10 loci were next to the TNFRSF11B and OPTN genes, we extended the genetic screening to these two genes, but we failed to identify any causative mutation at both the genomic and transcription level, suggesting that a different genetic defect is associated with PDB and potentially giant cell tumor of bone in this pedigree. © 2013 American Society for Bone and Mineral Research.


Paget's disease of bone (PDB, OMIM 602080) is a common skeletal disorder of the elderly characterized by focal abnormalities of bone remodeling.1 The abnormal bone turnover disrupts normal architecture and structure of single or multiple bones, leading to bone pain, deformity, and the development of various complications including pathological fractures, deafness, nerve entrapment syndromes, and secondary osteoarthritis. Studies have also indicated that patients with PDB have an increased incidence of osteosarcoma in the affected bone.

The etiology of PDB has remained largely unknown for several decades and still has to be conclusively determined. Both a genetic and a viral etiology have been suggested for this disorder.1–3 Initial studies demonstrated the presence of paramyxovirus material in pagetic osteoclasts, suggesting a viral etiology for the disease.1–3 However, familial clustering has been clearly recognized in up to 40% of cases, and at least seven potential susceptibility loci for the disease have been identified by genome-wide searches and candidate locus linkage studies.2, 4 Within most families, the disease is inherited as an autosomal dominant trait with genetic heterogeneity and incomplete penetrance.

In 2000, Hughes and colleagues identified a mutation on the TNFRSF11A gene, encoding receptor activator of NF-κB (RANK) in a single family with severe, early onset, polyostotic PDB.5 This mutation was similar to those reported in familial expansile osteolysis, a rare autosomal dominant skeletal disorder related to PDB.5, 6 However, further screening in different populations excluded mutations in TNFRSF11A as a common cause of PDB.7 Of interest, since 2002 mutations in a different gene, SQSTM1, have been identified in up to 10% and 40% of sporadic and familial PDB cases, respectively.8 This gene encodes the p62/sequestosome 1 protein, which acts as a scaffold protein in the NF-κB pathway as well as an intermediate protein in the proteosomal degradation of polyubiquitinated proteins. Even though patients with SQSTM1 mutations generally show an increased disease severity than SQSTM1-negative patients,9, 10 we recently identified SQSTM1-negative patients with a severe phenotype and the presence of peculiar phenotypic characteristics, including the occurrence of giant cell tumors (GCT) originating from affected skeletal sites.9, 11, 12 This complication represents a very uncommon clinical feature of PDB (described in less than 100 cases worldwide), and mainly occurs in patients with severe polyostotic PDB.11–15 GCTs may be multifocal and aggressive, leading to increased morbidity and mortality of patients. Patients with extensive, recurrent, and/or biologically more aggressive tumors may require wide excision, and often do not respond to antiresorptive compounds commonly in use to treat PDB such as calcitonin or bisphosphonates.

Importantly, both familial clustering and the evidence that GCT occurs with higher prevalence in PDB patients from Campania (accounting for up to 50% of all GCT cases), even if they have lived for several years in other countries, strongly suggest the hypothesis of a genetic factor in the etiology of this complication.13–15 However, some studies demonstrated the presence of the typical nuclear inclusions seen in PDB in the ultrastructure of the giant cells, suggesting a viral etiology.16

We recently characterized a large Italian pedigree with 14 affected members of whom four developed GCT, which despite the early onset and the severe phenotype were negative for the presence of SQSTM1 and TNFRSF11A mutations.9, 17 The high number of affected subjects across multiple generations (clinically followed from 1977 to date) allowed us to better characterize the clinical phenotype and look for possible susceptibility loci. A mutation screening of other candidate genes encoding for components of the RANKL/OPG/RANK/NF-kB signaling pathway (namely TNFSF11A and TNFRSF11B genes) or previously associated with PDB-related syndromes (namely VCP, encoding for valosin containing protein) or classical PDB in SQSTM -negative patients (OPTN gene, encoding for optineurin, and CSF1, encoding for MCSF-1) was also performed.

Materials and Methods

Pagetic pedigree

The pedigree of the examined family is reported in Fig. 1. All family members were born in Campania, and the majority of them were born in a rural area in the surroundings of Avellino, approximately 40 miles from Naples, within the region at the highest prevalence rate for the occurrence of GCT in PDB.11, 13, 15 Only subjects IV-2, IV-3, IV-4, and their children were born in Naples, outside that region. The diagnosis of PDB was confirmed in all patients by clinical examination, 99mtechenetium methylene diphosphonate (99mTC-MDP) bone scan, and subsequent X-ray examination of areas of increased isotope uptake. Moreover, all family members underwent a biochemical screening of total alkaline phosphatase (ALP) by standard techniques. In patients III-1, III-3, and IV-3, the PDB diagnosis was performed after the admission of their affected first-degree relatives in the Department of Clinical and Experimental Medicine of the “Federico II” University. An initial clinical screening of patients V-8 and V-9 patients, performed in July 2005 using a 99mTechnetium methylene diphosphonate bone scan and the measurement of total and bone-specific alkaline phosphatase serum levels, was not indicative for PDB. However, a subsequent screening performed in 2010 for the occurrence of bone pain demonstrated the presence of polyostotic PDB in both family members. All nonaffected subjects from generations III to V underwent a 99mTC-MDP bone scan to exclude the presence of asymptomatic PDB. As shown in Fig. 1, four of the 14 PDB patients developed GCT in one or more affected bones. Clinical characteristics of these patients developing GCT were previously described15 and are summarized in Table 1 and Fig. 2. Briefly, patient IV-02 developed a single GCT lesion of the jaw at the age of 59 years and did not respond to calcitonin or intravenous clodronate treatment. Then she underwent surgical curettage and subsequent radiation (external beam cobalt-60 with a total dose of 40 Gy during 6 weeks), with marked clinical improvement and resolution of the facial swelling. Conversely, the other cases (IV-1, V-4, and V-11) developed multifocal GCT with unsatisfactory clinical response to either calcitonin or bisphosphonate treatment (intravenous clodronate or pamidronate) and variable response to radiotherapy. In particular, patient V-4 (Fig. 2A–E) developed three GCT lesions of the skull (of the parietal, temporal, and occipital regions, respectively) and did not respond to multiple radiotherapy cycles (50 Gy). He died at the age of 61 years (after 9 years from the diagnosis of GCT) resulting from cerebral ischemia. All these patients had received previous bisphosphonate treatment (clodronate intravenously) at the time of GCT diagnosis, with partial biochemical response and total ALP levels well above the normal range (Table 1).

Figure 1.

Pedigree of the PDB/GCT family.

Table 1. Clinical Characteristics and History of Paget's Disease of Bone (PDB) Cases With Giant Cell Tumor (GCT)
 Patient IV-01Patient V-02Patient V-04Patient V-11
  1. ALP = alkaline phosphatase; CLO = clodronate; NER = neridronate; PAM = pamidronate; ZOL = zoledronate.

Age at PDB diagnosis (years)57443541
Age at first GCT diagnosis (years)67585151
ALP at GCT diagnosis (IU/L)6940422432003259
Sites affected by GCTMandible (67 years)Mandible (58 years)Parietal skull (51 years)Right clavicle (51 years)
 Skull (77 years) Temporal skull (51 years)Mandible (52 years)
 Maxilla (78 years) Occipital skull (51 years)Pelvis (55 years)
    Lumbar spine (56 years)
    Maxilla (58 years)
Medical therapyCalcitonin iv (for 3 years)Calcitonin iv (for 3 years)Calcitonin iv (for 3 years)CLN iv (for 10 years)
 CLN iv (for 6 years)CLN iv (for 16 years)CLN iv (for 4 years)PAM iv (for 1 years)
 PAM iv (for 1 years) PAM iv (for 5 years)NER iv (for 4 years)
    ZOL iv (for 2 years)
Radiotherapy1 cycle (45 Gy, mandible)1 cycle (40 Gy)3 cycles (50 Gy)No radiotherapy
 1 cycle (41 Gy, skull)   
 1 cycle (45 Gy, maxilla)   
SurgerySurgical curettage (mandible)Surgical curettageNo surgerySurgical curettage (clavicle)
    Surgical curettage (mandible)
    Laminectomy (lumbar spine)
Age at death (years)80766158
Cause of deathMyocardial infarctionMyocardial infarctionCerebral ischemiaHearth failure (2 years after a myocardial infarction)
Figure 2.

Phenotype characteristics of two patients with GCT complicating PDB. Patient V-04: (A, B) pictures showing dilated scalp veins and multiple giant cell tumors of the skull; (C) CT scan and (D) radiograph of the skull showing an occipital giant cell lesion originating from pagetic bone; (E) anterior and posterior bone scan images showing marked and diffuse radioisotope captation of the entire skull. Patient V-11: (F) CT scan showing the giant cell lesion originating from the vertebral body of L4; (G) Bone biopsy showing the typical aspect of the giant cell tumor with numerous osteoclast-like giant cells.

A specific questionnaire exploring the place of birth and residence, place of residence during childhood and adolescence, housing, animal contacts, dietary habits, occupation, and pharmacological history was performed in all patients. A detailed medical history including the record of common complications of PDB and other comorbidities was also performed in affected family members. When available, information from echocardiography and carotid Doppler ultrasound was recorded, particularly concerning the occurrence of signs of high cardiac output and the presence of valvular or carotid artery calcifications. When more than one analysis (performed at different ages) was available from a single patient, the closest one to the date of PDB diagnosis was selected. Cardiac index (a vasodynamic parameter that relates the cardiac output to body surface area) was calculated as a major index for high cardiac output.18 The study was approved by the local ethical committee, and all subjects had given informed consent to being included.

The general and clinical characteristics of patients from this pedigree were compared with those observed in 150 consecutive and unrelated PDB patients living in the same geographical region who were referred to the Department of Clinical and Experimental Medicine of “Federico II” Medical School of Naples.

Candidate gene analysis

After excluding SQSTM1(p62) and TNFRSF11A (RANK) mutations, we conducted mutation screening of all exons of other candidate genes (TNFRSF11B [OPG], TNFSF11A [RANKL], and VCP [p97]). Genomic DNA of four affected patients was specifically amplified using several sets of primers designed to amplify each of the coding exons of all five genes and was analyzed by direct sequencing. Given the results from linkage analysis in SQSTM1-negative PDB familes19 and of two genome-wide association studies20, 21 suggesting a possible susceptibility locus in 10p13 within a region encoding for optineurin (OPTN) and an additional locus next to the CSF1 gene, we also searched for possible OPTN or CSF1 mutations in affected members of our pedigree.

Linkage analysis

Genomic DNA from seven affected and three unaffected members were hybridized on the Affymetrix (Santa Clara, CA, USA) Genome-Wide Human SNP Array 6.0 according to the manufacturer's protocol. This array contains more than 1.8 million genetic markers, including more than 906,600 single nucleotide polymorphisms (SNPs) and more than 946,000 probes for the detection of copy number variation (CNV). Genotype of each SNP was generated with Birdseed v2. Quantile normalization was performed at probe level on the whole data set (sample + 240 references). For each single marker (SNP or CNV), the ratio in log 2 scale between the sample and reference set was then calculated.


General and clinical characteristics of affected members with and without GCT

All family members reported animal contacts for at least 10 years before the clinical onset of PDB (without any difference between PDB cases with or without GCT). These included pet ownership (cats and dogs) and contacts with pigs and rabbits. All examined subjects also referred the recurrent use of unpasteurized milk and of fresh, homemade meat products without sanitary controls, without any difference between affected and nonaffected family members. All cases affected by PDB showed a polyostotic disease (sites affected mean 5.7 ± 2.6, median 5, range 2 to 12) with a preferential localization in the skeletal axial bones (skull 12/14, 85.7%; pelvis 12/14, 85.7%; and vertebrae 13/14, 92.8%). As is evident in Table 2, there was no decrease in disease severity (as expressed by total ALP levels and the number of affected skeletal sites) across generations. On the contrary, there was an apparent decrease in age at diagnosis from generation III to generation V, likely owing to the extension of detailed clinical analysis in all family members of generation V, at a younger age than in generations III and IV.

Table 2. Clinical Characteristics and Complications of Family Members Affected by Paget's Disease of Bone (PDB) and Giant Cell Tumor (GCT) or PDB Alone
  1. Total alkaline phosphatase was expressed as percent increase from the upper normal limit.

Age at diagnosis (years)5744354170747538466059414125
Total alkaline phosphatase (%)15362523120719504573865412313712763217241241457
Affected skeletal sites (n)951264452545559
Bone pain++++++++++
Bone deformity+++++++
Hearing loss+++++
Neurological complications+++
Coronary artery disease++++++++
Myocardial infarction+++

Pagetic patients who developed GCT had an increased number of affected skeletal sites than patients without GCT (8.7 ± 2.9 versus 4.8 ± 1.7, p < 0.0001), without significant differences in age at diagnosis (44.2 ± 9.3 versus 52.9 ± 17.1, p = 0.36). ALP levels were also significantly higher in patients with GCT than in patients without GCT (1592.7 ± 646.8 versus 500.5 ± 296.1 IU/L, p < 0.001). Moreover, as shown in Table 2, there was an increased prevalence of bone pain (100% versus 60%), bone deformity (100% versus 30%), and fractures at pagetic sites (50% versus 10%) in PDB patients with GCT than in those without GCT. These differences reached statistical significance concerning bone deformities (p < 0.05 Fisher's exact test). Of interest, the prevalence of coronary heart disease was also higher in patients with GCT than in PDB patients without GCT or nonaffected family members (100%, 40%, and 6%, p < 0.001 in PDB/GCT, PDB, and nonaffected members, respectively). This difference was also evident in generation V (100%, 57%, and 0%, p < 0.001 in PDB/GCT, PDB, and nonaffected members, respectively), despite the relatively younger age of subjects. The onset of coronary heart disease was lower in patients with PDB complicated by GCT than in patients with PDB without GCT (55.0 ± 8.4 versus 61.2 ± 11.1 years, respectively, p = 0.38) with an overall age of onset of 58.4 ± 9.9 years. Moreover, all four PDB cases with GCT had myocardial infarction or died from cardiovascular complications. Subjects affected by PDB (with or without GCT) died at a younger age than nonaffected family members, with a difference approaching statistical significance (73.0 ± 10.1 versus 86.5 ± 4.8, respectively, p = 0.09).

Fig. 3 summarizes the variation of ALP levels in PDB cases with or without GCT from the diagnosis to the last follow-up analysis. Importantly, GCT occurred in those patients who did not respond to repeated treatment courses with calcitonin and/or bisphosphonates and in a condition of persistent active disease. Conversely, patients who responded to treatment with a marked reduction in ALP activity did not develop GCT. Moreover, the occurrence of GCT was associated with a consistent increase in ALP levels, as previously described in the case of osteosarcomas complicating PDB.22

Figure 3.

Temporal occurrence of GCT in relation to variation in ALP (expressed as percent increase from the upper normal limit). Time 0 indicates the time of PDB diagnosis. The black circles indicate the temporal occurrence of GCT in the four members of the pedigree (IV-01, IV-11, V-02, and V-04). Dotted lines indicate the variation in ALP in PDB members from the pedigree who did not develop GCT.

Clinical characteristics of PDB patients from this pedigree were compared with those observed in a cohort of 150 unrelated familial and sporadic PDB cases from the same geographic region (Table 3). Overall, the affected members of the pedigree showed an increased number of affected skeletal sites and an increased prevalence of polyostotic disease than either sporadic or familiar PDB cases from the same region. Moreover, the mean number of PDB complications per patient was significantly elevated in affected members of this pedigree than in the unrelated group of PDB patients (2.09 ± 0.9 versus 1.96 ± 0.8 versus 2.64 ± 1.3, in sporadic PDB versus familial PDB versus PDB/GCT pedigree, respectively, p < 0.05 ANOVA). Even though some PDB complications such as bone pain, fractures, nephrolitiasis, and neurological syndromes were more represented in the PDB/GCT pedigree than in the group of sporadic or familial unrelated patients, these differences did not reach the threshold for statistical significance. Conversely, a statistically significant increased prevalence of coronary artery disease was observed in patients from the PDB GCT pedigree (57.1%) than in the reference group of sporadic (11.7%) or familial (13.3%) PDB cases (p for trend <0.01).

Table 3. Comparison of Clinical Characteristics of Cases From the Paget's Disease of Bone (PDB)/Giant Cell Tumor (GCT) Pedigree and PDB Patients From the Same Geographical Area
 PDB from Campania
SporadicFamilialPDB/GCT family
  • *

    p < 0.05;

  • **

    p < 0.01; and

  • ***

    p < 0.001 vs. PDB/GCT family.

No. of cases1203014
Males/females (n)67/5316/145/9
Age at diagnosis (years)57.6 ± 10.654.7 ± 9.750.4 ± 15.4
Affected family members (n)2.93 ± 1.2***14
Affected sites (n)2.80 ± 1.9***4.00 ± 2.4*5.93 ± 2.7
Bone pain54 (45.0%)*20 (66.7%)11 (78.6%)
Bone deformity74 (61.7%)19 (63.3%)7 (50.0%)
Fractures9 (7.5%)3 (10.0%)3 (21.4%)
Osteoarthritis49 (40.8%)12 (40.0%)6 (42.8%)
Hearing loss34 (28.3%)6 (20.0%)4 (28.6%)
Nephrolitiasis10 (8.3%)4 (13.3%)3 (21.4%)
Neurological complications21 (17.5%)5 (16.7%)3 (21.4%)
Diabetes14 (11.7%)4 (13.3%)2 (14.3%)
Hypertension39 (32.5%)9 (30.0%)6 (42.8%)
Coronary artery disease14 (11.7%)***4 (13.3%)**8 (57.1%)
Myocardial infarction3 (2.5%)*2 (6.7%)3 (21.4%)

To provide further insight into the relationship between PDB and the occurrence of cardiovascular complications, information concerning the presence of carotid artery calcifications, valvular calcifications, and echocardiography parameters was obtained from clinical records of patients from the pedigree. Echocardiography data were available in 12/14 cases from the PDB/GCT pedigree (12/12 in generations IV and V) and were compared with those obtained from 20 unrelated polyostotic PDB patients from Campania. All four cases with GCT underwent the echocardiography screening 3 to 5 years before the occurrence of cardiovascular complications. Of interest, the cardiac index was higher in PDB cases from the pedigree than in the other PDB patients (3.77 ± 0.72 l/min/m2 versus 2.92 ± 0.82 l/min/m2, p < 0.01), and particularly in those family members who developed GCT (4.44 l/min/m2). Cases from the pedigree also presented an increased prevalence of left ventricular hypertrophy (75% versus 25%, p < 0.01) and valvular calcifications (42% versus 20%, p = 0.09) than the group of unrelated PDB patients. Moreover, a significant and positive correlation between cardiac index and ALP levels at the time of echocardiography (r = 0.43, p < 0.05) or the number of affected skeletal sites (r = 0.51, p < 0.01) was observed in the overall population of the 12 PDB/GCT and 20 unrelated PDB cases. The correlation between cardiac index and ALP was increased when analysis was restricted to cases from the PDB/GCT pedigree (r = 0.58, p < 0.05), whereas it decreased and became not statistically significant in the 20 unrelated PDB patients (r = 0.23, p = 0.3). Results of carotid Doppler ultrasounds were available from all 14 affected members of the pedigree and from the group of 20 unrelated PDB patients. Consistent with the echocardiography findings, affected family members of the pedigree had an increased prevalence of atherosclerotic plaques than the other PDB cases (78% versus 35%, p < 0.05). A trend for an increase in vascular calcifications was also observed (50% versus 25%, p < 0.10).

Genetic analysis

As is evident in Fig. 1, both PDB and GCT were inherited as autosomal dominant traits. In previous analysis, all the affected members of the pedigree were screened for mutations in the entire coding regions of SQSTM1 and TNFRSF11A genes, failing to detect any genetic alteration.9, 17 In the current candidate gene analysis, we now excluded the presence of mutations in five additional genes: OPTN encoding for optineurin, TNFSF11A encoding for RANKL (the ligand of RANK), CSF1 encoding for M-CSF, TNFRSF11B encoding for OPG (the decoy receptor of RANKL), and VCP encoding for valosin-containing protein. Mutations in the latter two genes were previously associated with juvenile PDB (also known as idiopathic hyperphosphatasia) and the syndrome of hereditary inclusion body myopathy-PDB-frontotemporal dementia (IBMPFD), respectively.23–25

As shown in Table 4, genome-wide screening allowed us to identify five possible candidate regions containing putative genes predisposing to PDB on chromosomes 8 (39 Mb between rs278559 and rs2280871), 5 (50 Mb between rs12514992 and rs17597145; 15 Mb between rs353287 and rs10462946), 6 (56 Mb from 6pter), 20 (2 Mb near the SNP rs11905231), and 1 (47 Mb between rs12142090 and rs6702754). Assuming a dominant model and a 100% penetrance, parametric analysis resulted in LOD score of 2.06 on chromosome 8 (94 to 133 Mb). Conversely, nonparametric analyses revealed most suggestive linkage on chromosomes 10 (Zmax = 6,2 at 20 to 30 Mbp from 10pter) and 8 (Zmax = 6,2 at 120 Mb). Given the consistency of the former association with some previous studies in SQSTM1-negative patients,19–21 we performed a genetic screening of the OPTN and CSF1 gene loci (see above), but we failed to identify any causative mutation at both the genomic and transcription levels.

Table 4. Summary Results From Parametric (P) and Nonparametric (NP) Linkage Analysis of the Pedigree
LocusRegion (cM)SizeLinkageInterval
chr 894324446-13379185139 MbPrs278559-2280871
chr 575554502-12588895550 MbPrs12514992-rs17597145
 148770478-16409562515 MbPrs353287-rs10462946
chr 6110391-5677432657 MbPrs4959515-rs9382665
chr 205127693-71280562 MbPrs1292244-rs6085920
chr 1104184672-15111704947 MbPrs12142090-rs6702754
chr 1024006978-2406368757 KbNPrs4491127-rs11013680
 30784089-3087063586 KbNPrs650788-rs3124176
 100584411-100919720335 KbNPrs10883215-rs2796755
chr 876976762-77100149124 KbNPrs17321328-rs830437
 10424956-12058691216 KbNPrs2447179-rs17794271


Epidemiological and clinical evidence clearly indicated a higher prevalence of neoplastic degeneration in PDB, with up to 1% increased risk of developing osteosarcomas at pagetic skeletal sites.22, 26 GCT is also a very rare but well-recognized neoplastic complication of PDB, accounting for only a small proportion of all neoplasms arising from pagetic bone.16 All GCTs complicating PDB occur exclusively in bone affected by the disease and generally differ in the age of onset and the skeletal localization with respect to GCTs occurring in nonpagetic subjects.16, 27, 28 In fact, GCT usually occurs in PDB patients older than 50 years and involves the craniofacial bone, the humerus, the femur, the pelvis, or the vertebrae. Conversely, nonpagetic GCT generally affects patients from 20 to 40 years of age and involves predominantly the distal femur, the proximal tibia, or the distal radius.16, 27, 28 Of interest, the skeletal sites more frequently affected by GCT in PDB patients were those preferentially affected by PDB in subjects with familial disease from Campania.12 Moreover, multicentric giant cell tumors arising in PDB patients have been described in 11 cases, including those presented in this study.15, 16, 27–30 In the international literature, up to 50% of cases of CGT complicating PDB have been reported in patients from Campania or with ancestry in this geographical area.11–16, 27

Despite the description of different GCT cases occurring in PDB, the clinical phenotype of patients with this complication has not been fully investigated. Moreover, at this stage, the evidence for a familial clustering on the occurrence of neoplastic PDB complications is virtually limited to a few patients from Campania. In fact, other than our patients and those described by Jacobs and colleagues,13 only Wu and colleagues described neoplastic degeneration in 2 of 3 family members of a PDB family from Castellammare di Stabia (about 30 miles from Avellino), with long-standing polyostotic disease. However, in this case, the two subjects developed osteogenic sarcoma, not GCT.31

In this study, we extend the knowledge about this rare complication. In fact, we clearly evidenced an increased disease severity in our pedigree of PDB/GCT patients and particularly in the four patients who developed GCT. Moreover, in contrast to the recent evidences from other PDB cohorts,32–34 we did not find a decrease in both clinical extension and severity of the disorder in the last generation with respect to the previous generations. In a previous study, we demonstrated that the region of Campania is associated with an enhanced clinical severity of PDB with respect to the other regions,8 and thus we preferred to compare the clinical characteristics of patients from this PDB/GCT pedigree to a cohort of unrelated patients from the same region. This also reduced the potential bias related to a different genetic background or lifestyle between the PDB/GCT pedigree and the reference cohort of PDB cases. Of interest, in this pedigree, in addition to an increased occurrence of common pagetic complications, we also evidenced a fivefold higher prevalence of coronary artery disease with respect to the cohort of unrelated PDB cases. This association has not been previously reported in the literature and might be owing to either disease activity (because of enhanced skeletal extension) or to a mutation in a new gene simultaneously affecting the skeleton and the cardiovascular system. Indeed, total ALP levels at diagnosis were significantly higher in subjects developing coronary artery disease than in the other PDB cases from the same pedigree (1111 ± 631 versus 414 ± 91; p < 0.05, respectively), partly supporting the former hypothesis. Moreover, an increase in the cardiac index was observed in cases from the pedigree rather than in the group of unrelated PDB patients, with levels above the threshold for high output cardiac state (>3.9 l/min/m2) in most members. Notably, cardiac index was also positively correlated with the extension of the disorder (as assessed by the number of affected skeletal sites) and disease activity (as reflected by ALP levels at the time of echocardiography). A similar behavior was observed concerning GCT because at the time of the occurrence of this complication, all four cases showed a limited response to treatment and a persistent active disease, with ALP levels well above the normal range. Taken together, these data suggest that the persistence of a state of active PDB might increase the risk of GCT and cardiovascular complications, at least in genetically predisposed subjects such as those from this GCT/PDB pedigree.

Of interest, current and previous9, 17 candidate gene analysis in this pedigree excluded the presence of mutations in all the major genes associated with PDB or PDB-related syndromes, suggesting that a different genetic defect is associated with PDB and potentially GCT. In keeping with our observation, a recent study in a North-American cohort of PDB patients excluded the presence of SQSTM1 mutations in three cases of Italian origin who developed GCT.35 Genome-wide linkage analysis identified different genomic regions potentially associated with the occurrence of the disorder in our pedigree. In particular, two regions on chromosomes 10 and 8 were significantly linked to PDB occurrence in this family. It is interesting to note that the region of chromosome 10p is very close to a recently identified locus in either familial and sporadic PDB cases negative for SQSTM1 mutation. In a first study of families of British ancestry without SQSTM1 mutations, multipoint parametric linkage analysis showed strong evidence of linkage to a single locus on chromosome 10p13 close to the marker D10S1653.19 More recently, a genome-wide association study in 750 sporadic PDB cases without SQSTM1 mutations and 1002 controls identified three candidate disease loci, replicated in an independent set of 500 cases and 535 controls.20 One of these loci was located on chromosome 10p13. Three SNPs (rs1561570, rs825411, and rs2095388), all located within a 30-kb region, were analyzed in both stages of the study, and the strongest signal was observed for SNP rs1561570.20 These findings were replicated in a larger cohort of PDB patients from different countries.21 OPTN, a candidate gene located in this region, negatively regulates TNFα-induced NF-κB activation, and a putative NF-κB binding site has been identified in the OPTN promoter.20 We identified a second interesting locus on chromosome 8, near TNFRSF11B gene (encoding OPG), the gene associated with juvenile PDB (JPD; OMIM 239000).23, 24 This association was suggested by both parametric and nonparametric analysis. However, we did not identify any mutation in both TNFRSF11B and OPTN genes, indicating that genetic variation in other genes located in these regions might confer susceptibility to PDB and possibly to GCT. Alternatively, mutations in a yet unidentified gene within the other genomic regions of chromosomes 1, 5, 6, or 20 might be responsible for the susceptibility to PDB in this pedigree.

Given the peculiar familial and geographic clustering of GCT in patients with PDB, it is also likely that only particular variations within this region may be associated with the occurrence of GCT, alone or in combination with other triggers (either environmental or genetics). Although further work will be required to identify the functional variant as well as the pathogenetic mechanism, the current study has provided new insights into the clinical phenotype and the genetic cause of GCT in patients with PDB.


All authors state that they have no conflicts of interest.


This work was supported in part by the 2009 John G Haddad Research Award from the Paget Foundation to LG. Moreover, the financial support of Telethon-Italy (grant no. 11119A) is gratefully acknowledged.

Authors' roles: Study design: FG, DR, and LG. Study conduct: FG, DR, DM, TE, MA, DF, RM, MV, and LG. Data collection and analysis: FG, DR, DM, TE, MA, DF, RM, GDF, PS, RN, MV, and LG. Drafting manuscript: FG, DR, DM, GDF, and LG. Revising manuscript content: FG, DR, and LG.