Novel PRUNE2 Germline Mutations in Aggressive and Benign Parathyroid Neoplasms

Simple Summary Recently, PRUNE2 mutations were indicated in the pathogenesis of aggressive parathyroid neoplasms. Here, we report novel, rare PRUNE2 mutations located in exons 3, 8, 9, and 12 in patients with parathyroid tumors in the genetically homogenous Finnish population. We identified PRUNE2 mutations in patients with parathyroid carcinoma, atypical parathyroid tumors, and adenomas. While further research is needed, mutations of the PRUNE2 gene could play a role in the pathogenesis of parathyroid tumors. Abstract Parathyroid tumors are mostly sporadic but can also occur in familial forms, including different kinds of genetic syndromes with varying phenotypes and penetrance. Recently, somatic mutations of the tumor suppressor gene PRUNE2 were found to be frequent in parathyroid cancer (PC). The germline mutation status of PRUNE2 was investigated in a large cohort of patients with parathyroid tumors from the genetically homogenous Finnish population, 15 of which had PC, 16 atypical parathyroid tumors (APT), and 6 benign parathyroid adenomas (PA). Mutations in previously established hyperparathyroidism-related genes were screened with a targeted gene panel analysis. Nine PRUNE2 germline mutations with a minor allele frequency (MAF) of <0.05 were found in our cohort. Five of these were predicted to be potentially damaging and were identified in two patients with PC, two with APT, and three with PA. The mutational status was not associated with the tumor group nor related to the clinical picture or severity of the disease. Still, the frequent finding of rare germline mutations of PRUNE2 may point to the gene playing a role in the pathogenesis of parathyroid neoplasms.


Introduction
Parathyroid carcinoma (PC) is a rare cause of primary hyperparathyroidism (PHPT). In Finland, around 2-3 new cases are diagnosed every year [1]. In contrast, PHPT due to benign adenomas is very common with an incidence of 1-2/1000 people [2,3]. The clinical picture of PC is similar to that of PHPT caused by benign parathyroid adenomas (PA). However, PC tumors are often larger than benign adenomas, and the clinical picture is usually more severe. PHPT due to benign adenomas is most common in postmenopausal women (gender ratio 3:1), while for PC, there is no gender difference, and incidence peaks at 40-50 years [4,5]. The diagnosis of PC is set based on unequivocal invasive characteristics on histopathological examination. Surgery is the only possible curative treatment. Of note, so called en bloc surgery, with the removal of the ipsilateral thyroid lobe and parathyroid gland gives the patient a better prognosis [6,7]. Atypical parathyroid tumors (APT), also called atypical parathyroid adenomas, are tumors which lack invasive histological characteristics The whole exome sequencing (WES) was performed from normal DNA. After WES analysis, the variant calls were filtered to have MAF < 0.05. Mutation consequences were assessed using pathogenicity predictor software and the ClinVar database. Five potentially deleterious PRUNE2 mutations were identified, distributed among seven PHPT patients. PC = parathyroid cancer; APT = atypical parathyroid tumor; PA = parathyroid adenoma; MAF = minor allele frequency.
The patient characteristics are shown in Table 1. The patient characteristics are shown in Table 1.

Hyperparathyroidism Gene Panel Findings
Gene panel findings in the cohort are listed in Table 2. The BpG custom hyperparathyroidism gene panel (www.blueprintgenetics.com, accessed on 3 January 2023, Helsinki, Finland did not reveal any additional germline CDC73 mutations in our cohort. One PC patient (152164) is the mother of one of the APT patients (152177). Otherwise, the patients are not related to each other and do not have any known cases of hyperparathyroidism in their respective families. Three PC patients are known to have an ethnicity other than Finnish. All patients were alive at follow-up. More than one surgery was performed on nine PC patients; five patients had local or distant PC recurrence. Additional en bloc surgery was performed as a preventative measure on four PC patients, as en bloc was not performed as the primary surgery. Three APT patients had more than one surgery; in all these cases, additional surgery was performed due to persistent hypercalcemia after the initial surgery, or the patient presented with additional parathyroid pathology (adenoma or hyperplasia). None of these patients were carriers of any potentially damaging germline mutations found in this study.
One PC patient (patient ID 152172) was found to have a likely pathogenic heterozygous germline MEN-1 mutation (c.1280G > T, p.Ser427Ile, ENST00000312049.6), although displayed no other clinical features of the MEN1 syndrome ( Figure 1, Table 2) [37,38]. One APT patient had a family history of MEN1, but no pathogenic MEN-1 variants were detected in this patient. The custom WES panel revealed monosomy X (Turner syndrome) in one APT patient (patient ID 152142) as an incidental or secondary finding (ISF) with clinical significance ( Figure 1). Moreover, a heterozygous germline APC missense mutation (c.2222A > G, p.Asn741Ser, rs150209825) was found in PA patient 152129. The variant is considered of unknown significance (VUS) [39][40][41]. This patient does not have any history of other neoplasia, and the possible family history of cancer is unknown. A 5 UTR variant of the APC gene (c.-128G > A, rs543098847) was found in PC patient 152174. This multiallelic single nucleotide variant is considered likely benign VUS (https://www.ncbi.nlm.nih.gov/clinvar (accessed on 3 January 2023)). A RET mutation (c.604G > A, p.Val202Met, rs751572082) of unknown significance was identified in PC patient 152165. An amino acid deletion of BRCA2 with unknown significance (c.3900_3902del, p.Met1300_Thr1301delinsIle, rs397507697) was found in APT patient 152175, and APT patient 152146 had a VUS missense AIP mutation (c.940C > T, p.Arg314Trp, rs375740557). In PA patient 152125, a multiallelic 5 UTR SDHA variant (c.-11C > T, rs1396057630) was identified.

PRUNE2 Mutations
Altogether, 25 non-synonymous PRUNE2 germline variants were found in the patient cohort. Out of those, nine variants had a minor allele frequency (MAF) <0.05 in a control population (https://gnomad.broadinstitute.org (accessed on 21 December 2022)) ( Figure 1). The details on these nine PRUNE2 variants are listed in Table 3. The nine variants were distributed among twelve patients, of which three had PC, six APT and three PA. None of the patients with rare PRUNE2 variants had mutations of previously established PHPT-related genes ( Table 2). All discovered mutations were heterozygous missense changes. Altogether, seven patients harbored a PRUNE2 variant in silico predicted to be likely deleterious (Figure 1, Table 3). Interestingly, the germline variant p.Ser595Tyr found in APT patient 152142 was previously also identified in prostate cancer patients [42]. Otherwise, the rare PRUNE2 germline variants found in our cohort have not been associated with any kind of pathology (https://www.ncbi.nlm.nih.gov/clinvar/ (accessed on 3 January 2023)). APT patient 152142 with Turner syndrome, who was 59 years old at the time of her diagnosis, had three different germline PRUNE2 variants; however, only (c.1784G > T, p.Ser595Tyr) was predicted to be damaging (Table 3). More than one PRUNE2 mutation was also found in patients 152127 (PA, 38 years old at diagnosis) and 152171 (PC, 66 years at diagnosis). Similarly, in these cases, only one mutation per patient was predicted to be pathogenic. In addition, patients 152155 (PC, 49 years at diagnosis), 152121 (PA, 44 years at diagnosis), 152123 (PA, 54 years at diagnosis), and 152131 (APT, 63 years at diagnosis) each harbored one PRUNE2 germline mutation that was predicted to be damaging ( Table 3). The location of the PRUNE2 mutations found in our cohort in relation to the PRUNE2 gene and previously found PRUNE2 mutations in PC (somatic and germline) are visualized in Figure 2. (https://www.ncbi.nlm.nih.gov/clinvar/(accessed on 3 January 2023)). APT patient 152142 with Turner syndrome, who was 59 years old at the time of her diagnosis, had three different germline PRUNE2 variants; however, only (c.1784G > T, p.Ser595Tyr) was predicted to be damaging (Table 3). More than one PRUNE2 mutation was also found in patients 152127 (PA, 38 years old at diagnosis) and 152171 (PC, 66 years at diagnosis). Similarly, in these cases, only one mutation per patient was predicted to be pathogenic. In addition, patients 152155 (PC, 49 years at diagnosis), 152121 (PA, 44 years at diagnosis), 152123 (PA, 54 years at diagnosis), and 152131 (APT, 63 years at diagnosis) each harbored one PRUNE2 germline mutation that was predicted to be damaging ( Table 3). The location of the PRUNE2 mutations found in our cohort in relation to the PRUNE2 gene and previously found PRUNE2 mutations in PC (somatic and germline) are visualized in Figure 2.  Most mutations were located in exon 8, the longest exon of the gene, but mutations were also observed in exons 3, 9, and 12. The cohort was screened for gene variants of ADCK1, CCD1, FAT3, and THRAP3 that have previously been associated with PC as described in the literature but that were not included in the BpG hyperparathyroidism gene panel assessment [4,16,25,27]. No previously described variants of these genes were found in our cohort.
PRUNE2 mutation status did not correlate with clinical parameters such as severity of disease (Ca-ion or PTH levels at diagnosis), tumor size, or parafibromin staining on immunohistochemistry, neither by looking at all rare prune mutations in our cohort nor by separately analyzing the variants considered damaging. Neither were the PRUNE2 mutations associated with a lower age at diagnosis.

Discussion
The majority of PC cases are associated with somatic alterations, although hereditary predisposition also plays a role in the genesis of this rare malignancy [14,43]. Somatic recurrent PRUNE2 mutations have been reported in up to 18% of PC cases. A reported germline PRUNE2 mutation and inactivation of the wild-type allele by LOH in the patient's tumor indicates that PRUNE2 alterations might predispose to PC [24,25]. To clarify the contribution of inherited PRUNE2 mutations in the development of parathyroid tumors, the germline mutation status of PRUNE2 was analyzed in a cohort of 37 mostly Finnish PHPT patients. We identified nine rare PRUNE2 germline variants (MAF < 0.05) in twelve patients, of which five mutations were predicted to be deleterious by disrupting the function of the PRUNE2 protein. These five mutations were distributed among seven patients, including individuals from all tumor groups (PC, APT, and PA). Interestingly, one of the mutations (c.270C > T, p.Asp90ASn) was shared between three unrelated patients: two with PC and one with PA. The MAF of the variant was 0.0148 among the Finnish controls and 0.0092 in the global control population (Table 3). This recurrent mutation might be a Finnish founder mutation, and enrichment of the mutant allele in our patient cohort may imply a causal relationship between the c.270C > T (p.Asp90ASn) mutation and the disease phenotype. The homogenous Finnish population has a unique genetic background, and founder mutations exist at high frequencies [44]. As such, Finnish population-based cohorts are valuable for the discovery of the causative genetic mutations of monogenic diseases.
Most of the rare PRUNE2 variants found in our cohort are localized to PRUNE2 exon 8, which is, by far, the largest exon of the gene, harboring approximately half of all the amino acids in its sequence. The mutations found in our study do not seem to be targeting any specific regions or previously established domains of PRUNE2 [45].
According to the literature, germline mutations of the PRUNE2 gene have been very scarcely investigated in other forms of cancers, despite somatic mutations playing a role in the pathogenesis of prostate cancer, leiomyosarcoma, and colorectal cancer among others [26,27,[29][30][31]. Of note, a recent study found PRUNE2 germline mutations in 2.8% of patients with familial prostate cancer, proposing PRUNE2 as a new prostate cancer predisposition gene [42]. The PRUNE2 germline variant c. 1784G > T (p.Ser595Tyr) found in our APT patient was also identified in the familial prostate cancer study. This mutant allele was classified as a variant of unclear association with the prostate cancer risk [42].
The patients with rare PRUNE2 germline mutations did not carry any germline mutations in previously established PHPT-driving genes. One patient was found to have monosomy X/Turner syndrome. This is relevant in the setting of parathyroid tumors due to the X chromosome harboring loci for the FLNA and KDM5C genes. The FLNA protein participates in the regulation of the calcium-sensing receptor and has been associated with increased aggressiveness in a wide range of cancers [46][47][48]. FLNA expression has also been investigated in parathyroid tumors, with unclear conclusions [49,50]. Inactivation of the KDM5C gene encoding for the histone demethylase protein JARID1C is frequent in renal cell carcinoma, and somatic KDM5C mutations have also been found in PC [25,51]. The monosomy X patient is thus susceptible to somatic mutations in these genes. One patient (152172) was discovered to have a likely pathogenic mutation of the MEN-1 gene. Other mutations of the same codon (c.1281T > A, p.Ser427Arg, rs1114167528) were previously reported in several MEN1 patients [37,38]. The family history of the MEN-1 mutationpositive patient identified in this study is not known. The patient has no other medical history of cancer except for PC, and the age at diagnosis (67 years) was not conspicuous. Still, due to the likely pathogenicity of this gene mutation, further clinical follow-up might be needed. Similarly, despite the germline APC mutations discovered in our cohort not being previously reported to be associated with pathogenicity, these patients and their relatives might also require further investigations, as APC mutations are so pronouncedly associated with hereditary colorectal malignancy [41,52].
Patients with previously known CDC73 mutations were excluded from this study, but as germline CDC73 mutations are also quite common in patients with sporadic PC it is perhaps rather surprising that none of the patients in the present cohort were found to carry CDC73 mutations. As CDC73 mutations may underlie both benign and malignant parathyroid tumors with varying penetrance and phenotypes, one can, therefore, speculate that mutations of the PRUNE2 gene could also give rise to similarly varying parathyroid tumor phenotypes with similarly varying penetrance [6,53,54]. The incomplete penetrance manifested by PRUNE2 germline defects might be the reason for the observed lack of family history of the disease.
The shortcoming of the study was the lack of tumor material; hence, the investigation of the biallelic inactivation of PRUNE2 in tumors was not achievable. Our patient cohort is of a reasonable size considering the rarity of PC. Another strength of our material is the detailed clinical, surgical, and histopathological characterization of the patients. However, the patient number is too small for the relevant assessment of the possible relationships between gene variants and clinical or biochemical parameters. All our patients were alive at the time of the study, with a median follow-up time of 7 years. Globally, the 5-year survival of PC is considered around 85%, indicating that our patients have had a rather favorable course of disease [1,55,56]. En bloc surgery was performed in 13 of our 15 PC patients (86%), either as primary or secondary surgery. The excellent prognosis of the patients likely reflects the high awareness of PC as a cause of PHPT in our tertiary centre and the close collaboration with our endocrine surgeons, preventing diagnostic delay, as well as en bloc surgery, ensuring margin-free resection. Recently, margin-free resection was reported to predict excellent long-term outcomes in PC [7].
In this study, we report that Finnish PHPT patients with CDC73 mutation-negative parathyroid tumors frequently display rare germline mutations in the PRUNE2 tumor suppressor gene. However, further work is needed to examine whether PRUNE2 plays a causative role in the genetic predisposition of parathyroid neoplasia. Clarification of this question would require additional sample sets and more extensive molecular workup. The identification of new parathyroid tumor-predisposing genes is important to improve the risk assessment of patients, and it would enable targeted testing of family members at risk.

Conclusions
Rare germline PRUNE2 variants are frequent in Finnish patients with parathyroid neoplasms, regardless of tumor type (PC, APT, or PA). Further studies are needed to clarify the role of PRUNE2 in patients with parathyroid tumors.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of the Helsinki University Hospital (Dnro 1803/2018).

Informed Consent Statement:
Written informed consent was obtained from all study subjects.

Data Availability Statement:
The data is available on request from the corresponding author.