Introduction

Primary hyperparathyroidism (PHPT) is a common endocrine disorder with the biochemical signature of hypercalcemia and either raised or inappropriately normal serum concentration of parathyroid hormone [1].

The prevalence of hypercalcemic PHPT in the population is rising, due to advances in diagnostics and the introduction of routine accurate measurements of serum levels of calcium and PTH [2]. Indeed, this rise of prevalence has occurred mainly because of the use of serum calcium determination as part of a multichannel screening profile, probably leading to an earlier detection of patients with primary hyperparathyroidism [3]. As a result, this led to a shift in the presentation of hypercalcemic PHPT from a predominantly symptomatic to an asymptomatic disease.

In addition, a new entity called normocalcemic primary hyperparathyroidism (NPHPT) has also been recognized [4]. This latter is characterized by normocalcemia and persistent elevated serum parathyroid hormone values, after excluding conditions determining raised parathyroid hormone levels.

NPHPT has been extensively studied to evaluate if complications of the disease were similar to its hypercalcemic counterpart [5,6,7]. However, a recent consensus statement acknowledges paucity of data concerning some aspects of the disease, such as pathophysiology, natural history, organ damage and management [3].

The present study was therefore carried out to investigate whether PHPT and NPHPT patient cohorts share the same pattern of genetic variation in genes known to be involved in calcium and/or bone metabolism.

Materials and methods

We genetically typed 31 patients with hypercalcemic PHPT (2 males and 29 females) and 27 patients with NPHPT (5 males and 22 females) diagnosed at Mineral Metabolism Center of Policlinico Umberto I, "Sapienza" Rome University during the period from September 2017 to October 2021. Initially, each patient underwent a thorough medical history, physical examination, and laboratory exams to exclude the presence of comorbidities and/or medications known to influence the investigated parameters (i.e., steroid therapy or diuretics and lithium salts).

The diagnosis of PHPT was made by conventional laboratory criteria, namely the finding of hypercalcemia and elevated or inappropriately normal serum PTH levels for at least 1 year. Following the international recommendation [1], in case of an increase of PTH levels with both total and ionized calcium in the normal range, before making the diagnosis of NPHPT, we excluded secondary forms of hyperparathyroidism including vitamin D deficiency, renal failure, malabsorption, low calcium intake, hypercalciuria and medications. Biochemical parameters were assayed as previously described [8]; in particular, serum PTH and 25(OH)vitamin D were measured by chemiluminescence-immunoassay (CLIA) with the fully automated LIAISON® analyzer. Intra and inter-assay coefficients of variation were 4.1% and 5.2%, respectively. Since 7% of patients studied (i.e., four patients, two each in the normocalcemic and hyper calcemic group) had serum PTH and total alkaline phosphatase determined outside of our laboratory, data for this parameter are reported as z-scores in respect to the average of normal range.

Each patient had abdominal ultrasound performed by a skilled radiologist. Each ultrasonogram was performed with a low-to-medium frequency (3.5–5 MHz, depending on the physical characteristics of the subject) convex probe and the ultrasound scanner (Esaote MyLab 70 X Vision; Esaote). Ultrasonography was performed in the supine, right and left lateral decubitus positions. The presence, number, and position of stones were evaluated. Renal stones were detected by specific ultrasonographic signs, such as hyperechogeneity and posterior acoustic shadowing. Bone mineral density values were measured by DXA (QDR-4500 Hologic Inc., USA) at the lumbar spine (L1-L4), femoral neck (FN), total hip (TH), and distal radius in all patients.

All patients gave written, informed consent before their inclusion in the study. The investigation was approved by the Institutional Review Board of the Department of Clinical, Internal, Cardiovascular and Anesthesiologic Sciences and then approved by the Ethics Committee of “Sapienza”, University of Rome (protocol number 3040N 73/14). The Research was carried out complying with the World Medical Association Declaration of Helsinki.

Genoty**

Genomic DNA was extracted from peripheral blood using laboratory standard procedures. We analyzed the 9 SNPs listed in Table S1, more involved in the regulation of calcium homeostasis and bone density. Genoty** was performed by TaqMan allele discrimination assays (Applied Biosystem) according to the manufacturer’s instructions. Genotype was assigned by registering the fluorescence emission from each sample at the VIC and FAM dye wavelengths. The same assays were applied to all in-house controls, i.e., 54 anonymous control DNAs routinely used in the lab (hereafter in-house controls), and additional 106 subjects from Central-South Italy [9], typed only for SNPs in PTH and ESR1.

The above datasets were integrated with genotypes for the 503 subjects of European descent of the 1000 Genomes project [10]. The data slicer available at http://www.ensembl.org/index.html was used to extract genotype data at the relevant and surrounding positions from high coverage sequencing results deposited at http://ftp.sra.ebi.ac.uk/.

Genetic data encoding

In order to run sparse Principal Component (sPC) analysis genotypes at biallelic loci were encoded by paired variables with 0’s and 1’s indicating the absence or the presence of the reference and alternative allele [11].

Two SNPs at the ESR1 locus (in the order rs9340799-rs2234693), and two at the VDR locus (in the order rs731236-rs7975232) were combined into haplotypes. For the patients, in-house and additional controls, phasing was obtained with Phase2 [12]. For the 1000 Genomes dataset the phasing reported in the sliced VCF files was retained.

These procedures resulted in four observed haplotypes at ESR1 and only three at VDR, that were encoded into four and three binary variables, respectively (Table S1).

Data analysis

Allele/haplotype frequencies and the testing of Hardy-Weinberg equilibrium were obtained with Arlequin [13] for each group of subjects, separately. All remaining calculations were performed in an R environment. Multidimensional analysis was performed by sPC [

Discussion

We analyzed nine single nucleotide polymorphisms in two groups of patients with hypercalcemic and normocalcemic primary hyperparathyroidism, very well characterized from a biochemical point of view. These two groups were compared with the results obtained from in-house controls and 503 subjects of European descent.

Concerning the normocalcemic group, which constituted 46.5% of the whole patient series, we analyzed genetic data on a single locus basis, as well as considering two-loci genotype distributions and a multidimensional genotype representation. We found instances of genotype imbalances at individual loci.

Our results suggest that the rs9340799-rs2234693 genotype (GC/GC) at ESR1 is associated to NPHPT. Indeed, primary hyperparathyroidism is often diagnosed in women, in the first decade after menopause, consistent with the known skeletal action of estrogen that neutralizes the hypercalcemic effects of excess PTH in bone [15]. However, it is unclear whether this genotype simply delays the transit from the normo- to the hypercalcemic state or marks subjects who remain normo-calcemic. Ad-hoc follow-up studies will be needed to clarify this point. At any rate, ESR1 can be considered for an initial genetic characterization of PHPT patients. In this context, it should be emphasized that the natural history of NPHPT is still unclear. Eastell et al. [16] performed a 5-years retrospective evaluation of a cohort of NPHPT patients. They concluded that NPHPT may be a mild form of PHPT.

As to the higher incidence of the rs6256(T/T) genotype at PTH in normocalcemic patients, the significance is elusive. At this site, the alternative (T) variant generates a premature stop codon p.(Arg83*) that determines a truncation of the mature PTH peptide to 52 amino acids, causing impairment of translocation across the endoplasmic reticulum, cleavage of pro-PTH and secretion of PTH [17, 18]. No specific function has been attributed to the portion of the secreted PTH polypeptide downstream to pos. 83, so far. Instead, the two NPHPT patients with the T/T genotype displayed significantly higher PTH levels and calcium excretion, while lower phosphorus values as compared to the alternative genotypes (Kruskal-Wallis p = 0.028, 0.037, and 0.037, respectively). Our results are at odds with the findings in ref. [19], and point to a potentiation of PTH activity when the C-terminal portion is present [15]. Further experimental models should evaluate if this polymorphism could have implications for tissue-specific biological actions of PTH.

Also, the NPHPT cohort was peculiar for pairwise associations of genotypes, pointing towards gene effect interactions that are associated to NPHPT. These associations in the NPHPT cohort raise some hypotheses. In fact, significantly unbalanced distributions of two-loci genotypes were observed for ESR1-RANKL, and VDR-OPG. It is well established how the receptor activator of nuclear factor-κB (RANK), RANK ligand (RANKL), and its decoy receptor osteoprotegerin (OPG) play key roles in regulating bone turnover [20]. Previous data showed as baseline serum concentrations of OPG and RANKL were higher in PHPT patients than in healthy controls, whilst the OPG/RANKL-F ratio was lower [21]. In addition, the latter authors also reported the absence of changes of serum osteoprotegerin values following parathyroidectomy.

Hence, even though no single SNP typed in our cohort was able to explain a relevant proportion of the overall diversity, it may be plausible that some genotype associations at two or three loci (implicated in calcium homeostasis) may guide in working out a genetic contribution to different phenotypes in PHPT or NPHPT.

Our multivariate analysis, aimed at condensing the genotype heterogeneity of the NPHPT cohort, revealed overrepresentation of unusual multilocus genotypes. As compared to previous works in the literature [19, 22], this method can potentially capture the combined effect of multiple alleles/haplotypes even when the contribution of each of them is subtle. These genotypes may help in defining subgroups of NPHPT patients which deserve ad hoc clinical and follow-up studies to tackle the question of why “some patients develop skeletal, renal or other complications while others do not [15]”. To improve genetic identification of risk factors, further polymorphisms in genes potentially involved in clinical manifestations should be analyzed. Among them, particularly relevant could be additional SNPs in genes encoding for receptors of calcium (CASR) [23, 24], PTH (PTHPR1 and PTHPR2) and calcitonin (CalcR).

Conclusion

This is the first report exploring the genetic aspects underlying the two phenotypes of primary hyperparathyroidism, i.e., the features of the normocalcemic and hypercalcemic cohorts. Long-term longitudinal studies are needed to evaluate if specific polymorphisms could be able to target those individuals transitioning from normocalcemic to hypercalcemic state.