MATERIALS AND METHODS

Families

Genetic analysis was performed on seven kindreds diagnosed with FIHP. The diagnosis of PHPT in the participating individuals was based on extensive and careful investigation in order to exclude the presence of MEN1 or FHH. Informed consent was obtained from all patients and the study was approved by the local ethics committees.

Family 1 is residing in France, and the members have been clinically followed since 1993 (fig 1, table 1). In this family all patients have been screened by a standardised established protocol by the French network on MEN disease (Groupe d’Etude des Tumeurs Endocrines-GTE/GENEM).²⁵ In brief, this included biochemical measurements of serum and urinary calcium, PTH, phosphate, gastrin, insulin, peptide C, glucagon, chromogranin A, pancreatic polypeptide, basic prolactin, and growth hormone, as well as abdominal ultrasound and CT (of the kidneys, adrenal glands, liver, and pancreas), thoracic CT (to search for carcinoids), cervical ultrasound, and pituitary MRI. For the last 2 years, orthopentography has been performed for all members of the family.

The grandfather (I:2) was diagnosed at 59 years of age with PHPT following a period of symptoms suggestive of kidney stones. Two benign hyperplastic parathyroid glands were identified and surgically removed, since which time he has been clinically well. His two children from the second marriage (II:3 and II:5) have both been affected and treated surgically for PHPT. The daughter II:3 was diagnosed with kidney stones at the age of 32 years, and a single parathyroid adenoma, 1.8 cm in size with multiple cystic formations and haemorrhagic foci, was identified at surgery and removed. Nine years later, at 41 years of age, she had a recurrence of PHPT and a second parathyroid adenoma was removed. The patient is presently 53 years old, and has persistent hypercalcaemia at the upper limit with elevated PTH (103 ng/L). She has also developed osteous abnormalities typical for PHPT and recurrent hypercalciuria. Her brother (II:5) was diagnosed at 21 years of age with PHPT following an atypical clinical presentation with mental affective disturbance and kidney stones. A single adenoma was identified at surgery and removed. Eleven years later recurrent PHPT was noted and one additional enlarged parathyroid gland was removed, following which the patient has been asymptomatic. The grandson (III:4) was treated at 25 years of age for PHPT with removal of one large intrathymic parathyroid adenoma (1.835 g). The younger brother (III:6) is clinically and biochemically unaffected. The dizygotic twins (III:7 and III:8) are presently 9 years old and have serum calcium levels at the upper limit of normal but are otherwise without symptoms of PHPT. The presence of MEN1 syndrome in this family was excluded by clinical screening for related tumours and by negative MEN1 mutation analyses in all affected members. No tumours were detected or suspected using ortopenthography or renal CT scanning. The observation of two renal cysts in II:3, and one cyst in each of individuals II:3 and III:4 were interpreted as normal findings because of their small size and benign appearance.

Family 2 (fig 2), which has previously been reported to exhibit FIHP linked to the HPT-JT locus,¹⁹ was clinically re-examined in this study. The family members included in the study were biochemically screened, with measurements of serum and urinary calcium, phosphate, albumin, and creatinine. The radiological screening included orthopentography and abdominal ultrasound. This revealed five individuals affected with PHPT related to cystic parathyroid adenomas, but without evidence of other endocrinopathies or neoplasias. The proband (IV:8) was diagnosed with PHPT at 22 years of age, and treated with parathyroidectomy to remove one adenoma. His grandmother (II:2) presented clinically with renal calculi in her forties. After recognition of PHPT at 58 years of age, two parathyroid adenomas were surgically removed, one of which was unusually large (4 g, 40 mm). The proband’s father (III:8) developed PHPT at 25 years of age, and a single adenoma was removed. Seventeen years later, at 42 years of age, he had recurrence of PHPT and a second parathyroid adenoma of 1.5 g was removed. He died in 2001 at the age of 51 years following a myocardial infarction. The father’s three paternal aunts (II:3, II:4, and II:7) have not shown any clinical or biochemical symptoms of PHPT. Their respective children are similarly unaffected. The proband’s aunts (III:2, III:3, and III:6) are clinically unaffected and have normal range total serum calcium levels. Two of his cousins (IV:1 and IV:3) were surgically treated for PHPT at 21 and 19 years of age respectively, with removal of one adenoma in each case. The other five cousins (IV:2 and IV:4–7) are presently unaffected, with serum calcium levels within the reference range. However, IV:5, who is presently 16 years old, has had four fractures. Neither of the proband’s brothers (IV:9 or IV:10) has clinical or biochemical signs of PHPT. As previously reported, no case of jaw or renal tumour and no evidence of MEN1 has been found in this family¹⁹. Additional screening of III:6 with orthopentography and renal ultrasonography was negative.

FIHP families 3–7 (table 2) have been clinically described and screened for MEN1 mutations as previously reported.²⁴ They were diagnosed as having PHPT on the basis of the clinical examinations, with findings of hypercalcaemia, increased parathyroid hormone (PTH), and family history of PHPT. The parathyroid lesions consisted of adenomas or hyperplasia. No other neoplastic lesions suggestive of MEN1 or HPT-JT have been found, further supporting the diagnosis of FIHP. Furthermore, previous linkage analyses in these families could not exclude the possible involvement of MEN1 and HRPT2 loci.²⁴

Mutation analyses of the HRPT2 and CASR genes

Mutation screening of the HRPT2 gene was performed by sequencing the entire coding region in two or three affected family members representing each of the seven families. The primers used for amplification and sequencing as well as the annealing temperatures for the amplification are presented in table 3.

The 17 coding exons of HRPT2 were amplified as 15 different fragments with primers derived from the flanking intronic or 3′/5′ UTR regions, to allow detection of mutations in coding regions or affecting the splicing. The DNA (50 ng) was amplified in a mix of 1× Taq Gold buffer, 1.5 mmol/l MgCl₂, 1.25 mmol/l of each dNTP and 10 mmol/l oligonucleotide primer pair, 1.5 U DNA polymerase (AmpliTaq Gold™) in a final volume of 20 μl. The thermocycling conditions consisted of an initial denaturation of 95°C, followed by 35 cycles of denaturation at 95°C for 40 seconds, annealing at 58–62°C for 30 seconds, and elongation at 72°C for 45 seconds, and a final extension at 72°C for 7 minutes.

Sequencing reactions were performed using the BigDye Terminator cycle sequencing kit (Perkin Elmer) with the following thermocycling conditions: denaturing at 98°C for 5 minutes, followed by 25 cycles of 96°C for 10 seconds, 50°C for 5 seconds, and 60°C for 4 minutes. The sequencing products were then purified with Sephadex G-50 (Amersham Pharmacia Biotech, Sweden) on MultiScreen-HV Clear Plates (Millipore Corp, Bedford, MA, USA) and thereafter run in either a capillary sequencer ABI Prism 3700 DNA analyser or in an automated sequencer (ABI 377; PE Applied Biosystems, Foster City, CA, USA).

Sequencing of the mutated exon/intron was performed for all available members of families 1 and 2, as well as for the parathyroid adenoma from II:2 in family 2, using forward and reverse sequencing. The mutation in family 1 was confirmed by restriction cleavage assay. Each PCR product of exon 2 was cleaved with DdeI (Invitrogen) in a total volume of 100 μl, and separated on 3% agarose gels. The presence of a mutation results in the introduction of a DdeI restriction site, leading to the addition of two shorter restriction fragments detected by gel electrophoresis. In addition, the mutation negative status of members II:1, II:7, and III:6 in family 2 was confirmed by sequencing of exon 2 using a different set of flanking primers. Sequencing of HRPT2 exon 2 and the flanking intronic region was performed on 215 normal unrelated control individuals, comprising 150 controls described by Carpten et al³ and the 65 controls reported by Howell et al.²⁶

In addition, sequencing of the six coding exons (2–7) of the CASR gene was carried out using the same primers and PCR conditions as described elsewhere²⁷ in families 3–7 without identified HRPT2 mutations.

Genotyping of microsatellite markers

Genomic DNA was extracted from peripheral leukocytes using standard methods and genotyped with polymorphic microsatellite markers. Seven markers from the HRPT2 region in 1q25–q32 were selected, including cen-D1S222-D1S428-D1S422-D1S461-D1S542-HRPT2-D1S412-D1S413-tel. The localisation and order of the markers are based on the information presented in Carpten et al³ and in the Genome Database (www.gdb.org). For all markers, the forward primer was labelled for fluorescent detection after electrophoresis on a laser fluorescent sequencing machine (ABI 377, PE Applied Biosystems). Polymerase chain reaction was performed as described for the mutation screening of HRPT2. The data were analysed using Genescan softwere (PE Applied Biosystems).

RESULTS

The seven kindreds studied here all had 2–5 members affected with PHPT related to benign parathyroid tumours (table 2). At the initial diagnosis there was no evidence of other endocrinopathies or tumours found in the families. During follow up, these patients have not presented with any signs suggestive of other familial syndromes such as MEN1 or HPT-JT, thus further supporting the diagnosis of FIHP.

HRPT2 mutations and mosaicism in FIHP

The two HRPT2 germline mutations identified in the two families constitute of a missense mutation and a splice mutation (fig 3). Neither of these mutations was identified by sequencing of 215 normal reference individuals. Two polymorphisms (IVS2 +28 C→T and IVS7+33 (GA)₈) in intronic sequences were detected in two of the families and in normal reference individuals.

In family 1, the base substitution (G→C) detected at the boundary of exon 2 and intron 2 (IVS2+1G→C; fig 3) is predicted to cause defective splicing. The mutation was found in all four affected family members (I:2, II:3, II:5, and III:4) and in both children (III:7 and III:8) of II:5, who are presently 9 years old and unaffected (fig 1). Furthermore, genotyping of markers from the HRPT2 region showed that all mutation carriers shared the same disease associated haplotype (fig 1, table 1). Here, the intronic HRPT2 polymorphism (T/C) revealed at position +28 in intron 2 served as an intragenic marker of the affected haplotype.

In family 2, the missense mutation in exon 2 consisted of a base substitution (CTT→CCT) causing an amino acid change from leucine to proline at position 64 (L64P, fig 3). The mutation was present in four affected family members (III:8, IV:1, IV:3, and IV:8) and in three presently unaffected members (III:2, III:3, and IV:5) (fig 2). One of these mutation carriers is an obligate carrier (III:2) and one has a history of multiple fractures (IV:5). All mutation carriers were also shown to share the same disease associated haplotype of the HRPT2 region. The polymorphism detected in intron 7 (IVS7+33 (GA)₈) was also present on the affected haplotype, supporting the linkage results. The mutation was not detected either in the oldest affected family member (II:2) or in her three sisters (II:3, II:4, and II:7) and one daughter (III:6). However, two of these individuals (II:7 and III:6) carried the affected haplotype for the entire interval, including the intron 7 polymorphism (fig 2). Interestingly, the mutation (L64P) was also found in the parathyroid adenoma of II:2 but not in her leukocyte DNA. Repeated genotyping and mutation analyses on a new set of DNA samples from this family confirmed the findings in all cases. Taken together, the findings in family 2 suggest that the L64P mutation occurred de novo on the disease associated haplotype in individual II:2. This haplotype was then inherited with the mutation by three of her children (III:2, III:3, and III:8) but without the mutation by her daughter, III:6 (fig 2).

FIHP families negative for HRPT2, CASR and MEN1

In five families (3–7), no HRPT2 mutation were detected. Previously, these families had been screened for MEN1 mutations with negative results.²⁴ In an attempt to investigate the third locus involved in FIHP, we addressed the CASR gene by screening for mutations in the six coding exons. However, no mutations could be demonstrated.

DISCUSSION

The seven families reported here were all shown to have FIHP, which in two of them was found with constitutional mutations of the HRPT2 gene. In family 1, the HRPT2 mutation destroys the splice site at the 5′ end of intron 2. In family 2, the missense mutation (L64P) in exon 2 alters an evolutionarily conserved amino acid that is present among orthologous sequences such as mouse, Drosophila melanogaster, and Caenorhabditis elegans (fig 3). It can be speculated that the effect of the single residue alteration may change the local structure or interfere with the ability for parafibromin to interact with other proteins. In both families the inheritance of microsatellite alleles is in agreement with linkage of PHPT to the HRPT2 region, as also determined from analyses of intragenic polymorphisms. The mutations are present on the respective affected haplotypes, but were not detected in 215 normal controls. Taken together, these findings suggest the possible pathogenetic importance of the two mutations, thus demonstrating that HRPT2 is a disease gene for 1q linked FIHP. In further support of the association between FIHP and HRPT2, it can be noted that the parathyroid tumours from both families tended to be large and have cystic features, similar to that seen in HPT-JT. Furthermore, loss of heterozygosity of the wild type 1q alleles in the parathyroid tumours of family members II:2, IV:1, and IV:8 in family 2 has previously been reported,¹⁹ in agreement with the likely tumour suppressor function of HRPT2.³

No mutations were identified in the remaining five FIHP families (3–7) in any of the three genes investigated (table 2). However, mutations in regulatory or intronic parts of HRPT2, MEN1, and CASR cannot be excluded by this study. There is also a possibility that other chromosomal regions may contribute to the development of parathyroid tumours in FIHP.

In dissecting the genetic aetiology of FIHP, the diagnostic aspects are of great importance. Familial PHPT is recognised as a genetically heterogeneous disease that can occur either as a separate entity or as part of an inherited cancer syndrome such as MEN1 and HPT-JT.¹ Therefore, some families with PHPT can initially present as isolated PHPT, but subsequently develop the full blown syndrome of HPT-JT or MEN1,^28–30 while in others the diagnosis of FIHP remains during follow up.^19–21 In the present study, the FIHP families with HRPT2 mutations have been followed up for more than 10 years, during which only benign PHPT has been diagnosed, and the possibility of MEN1 and HPT-JT has been carefully evaluated and excluded. Furthermore, affected members from three generations were investigated in both families, including several adult mutation carriers well above the expected age of onset for symptoms detectable at screening.

The missense and splice mutations detected in the study are not representative of the mutation spectra of families with HPT-JT syndrome (fig 3). In the initial publication of the HRPT2 cloning, 12 of the 13 different germline mutations were predicted to give impaired protein function due to truncation or premature stops.³ The only missense mutation detected was located at the first ATG, leading to loss of the initiation codon.³ All of these families were clinically characterised with PHPT, and in addition jaw tumours, renal lesions, or parathyroid carcinoma.^3,31 Although based on a limited number of observations, it is striking that the HRPT2 mutation spectra in FIHP v HPT-JT is similar to that observed for the MEN1 gene. With regard to the MEN1 gene, FIHP families predominantly have missense mutations,^20–24 whereas in MEN1 families truncating or nonsense mutations are more frequent.^13,14

In one of the FIHP families, the HRPT2 mutation (L64P) was not found in the initial mutation screening as it had occurred de novo. Extensive analysis of additional family members led to the identification of the mutation, which was reported in Carpten et al³ and is further described in the present study. As illustrated in fig 2, three of the four children of II:2 inherited the affected haplotype with the disease mutation from their mother (II:2) and in turn passed the same mutated haplotype on to their respective children. However, the same affected haplotype was also inherited to her fourth daughter (III:6) without the mutation. This daughter has recently been thoroughly re-examined clinically and found to be clearly unaffected after examination of the parathyroids, the jaws, and the kidneys. In addition, the same affected haplotype was present without the mutation in one of the mother’s sisters (II:7), who is also clinically unaffected. The mother has been surgically treated for PHPT, and genetically she carries the affected haplotype without the mutation in constitutional DNA isolated from her peripheral leukocytes. However, the mutation was present in the tumour DNA, which would suggest that the mother is somatic and germline mosaic for the initiating mutation. In this particular family, the mutation would then have occurred de novo during embryogenesis of II:2 on the chromosome with the affected haplotype. Owing to the mother’s mosaicism, this haplotype was then passed three times with the mutation and once without it on to her four children. Similarly, mosaicism in II:2 would explain why the mutation is not present on the affected haplotype in leukocyte DNA although she is clinically affected with PHPT. Mosaicism has previously been observed in other hereditary diseases, for example retinoblastoma and neurofibromatosis, and has then been associated with similar discrepancies between the inheritance of marker alleles and disease mutation.^32,33

In summary, we have demonstrated that constitutional mutations in the HRPT2 gene are likely to be associated with parathyroid tumourigenesis. Specifically, the involvement of HRPT2 is demonstrated to be a disease gene for 1q linked FIHP. The finding that a subset of FIHP occurs as an allelic variant of HPT-JT has implications for clinical diagnostic purposes and for improved understanding of parathyroid tumour development. It is expected that further functional studies of mutations in the HRPT2 gene will provide relevant information to better explain the tissue specific involvements of HPT-JT and FIHP.